NASA CR-'

SKYLAB PROGRAM

EARTH RESOURCES
EXPERIMENT PACKAGE

SENSOR PERFORMANCE EVALUATION

.FINAL REPORT ,
VOLUME III ($192)
{IASA-CB-144564) SKYLAB PROGDAIEARTH
N76-13563

RESOURCES EXPERIZ ENT PACKAGE SENSOR

PERFORAANCE EVALUATIOV, VOLUMER.3, (S192)

ina Beport (Martin 8arietta Corp.). 106 p Unclas

___5 5 _________--__ _ CSCL05B G3/43 03899

?!AY 5, 1975
CONTRACT NAS8-24000
AMENDMENT JSC-14S

NationalAIeronaic and Space Administration

LYNDON B. JOHNSON SPACE CENTER

Houston, Texas

 JSC-05546

EARTH RESOURCES EXPERIMENT PACKAGE

SENSOR PERFORMANCE REPORT

VOL. III (S192)

FINAL REPORT

MAY 5, 1975

Subditted by.:

Skylab B EP Sensor Performance

Evaluation Manager, JSC/HC

Technical

Review by:
Y

William Hensley

S192 Project Scientist, JSC/TF

Approved:
_______________________

Manager, Systems Analysis and

Integration Office, JSC/HC

Contract NAS8-24000

Amendment JSC-14S

Skylab Program

Lyndon B. Johnson Space Center

 MSC-05546 Volume III

FOREWORD

This volume is Section III of six sections of document NSC-05546,

submitted by Martin Marietta Corporation, in accordance with the require­
ments of Annex I to Exhibit A, Statement of Work, Part I, Data Requirements
List, of Contract NAS8-24000, Amendment JSC-lhS, Line Item 295; and was
prepared under WBS 02216.

III-iii
 MSd-05546

CONTENTS

Section Page

l.0 INTRODUCTION . . . . . . .ii

.

1.1 Purpose i . .. . .... . . . I-I

1.2 Scope . . . . I.. lll-i

1.3 Usage duide . . ..... .. I-i

1.4 Abstract ......... . 111-2

2.0 APPLICABLE DOCUMENTS . . .. .. .. 111-3

3.0 SUMMARY oP SENSOR PERFORMANCE EVALUATION INTERIM

REPORT ... .. . ................ . 111-4

3.1 Vunction/Limit Verification .. ... ... 111-4

3.2 Datd tnteteretce.. . ... ...... ....... 111-7

3M3 Response versus Scan-Angle Determination . . .

. ..... 111-8

3.4 Absolute Radiometric Accuracy and Stability . ...... 111-17

345 System Frequency Response ... ...... . . ....... 111-25.

3.6 Geometric Distortion. ......... . . . ......... 111-32

3.7 NoiseAnalyses .....

......
..

. 111-33

3.8 Pointing Accuracy and Field-of-View Determination . . . 111-38

3.9 Photographic Image Adequacy for Site Location....... 111-40

3.10 Geometrical Band-to-Band R

egisttation Error . .....
. 111-41

4.0 SUPPLEMENTARY ANALYSES ..... ....... .......I.I....111-43

4.1 -Supplementary Analysis of Scan-Related System

Response. . ... ......... . . . ....... 111-43

4.2 Supplementary Analysis of System Noise ......

...... 111-48

4.3 sIg0A, 9191 and 8192 Radiomettic Comparis6n............ 111-49

5,0 CONCLUSIONS ....

... ........ .. ...... 111-57

6.0 RECOMMENDATIONS ......... .............. . .I..111-57

I1I-it

 MSC-05546

Section Page

7.0 NOTES ............ ....................... 111-58

7.1 Acknowledgements ... ..... ............... . III58

.3­

7.2 Abbreviations ............. .................. III-59

APPENDIX A -TECHNIQUES ADDENDUM ......... ........ III-A-I

III-v

-j

MSC-05546

TABLES

Table Page

3.1.1-1 CONTROLLED HOUSEKEEPING FUNCTION SUMMARY ........ 111-5

3.1.1-2 UNCONTROLLED HOUSEKEEPING FUNCTION SUM14ARY ....... .. 111-6

3.1.2-1 CALIBRATION DATA STATUS .... ................ 111-6

3.7.1-1 STANDARD DEVIATIONS, ATtENUATORS OUT .. ........ .. .111-33

3.7.1-2 STANDARD DEVIATIONS, ATTENUATOltS IN . . .......... .. 111-34

3.7.1-3 NOISE-EQUIVAIENT SPECTRA RADIANCES (NESR),

ATTENUATORS IN ........ ................ .. 111-37

3.7.1-4 NOISE-EQUIVAttNT SPZCTMM. RADIANCtS (NESR)5

ATTENUATORS OUT ... ..................... 111-38

3.7.2-1 THRMA--ANt NtSM . , , .. . . ........ ti1-19

3,8.1-1 MAXIMUM LATITUDE AND toNITUD, nIVtUNf C f1VNPTtN s1 2

DATA AN POINTING ESTIMATES DETMMINED PFOM SKYLAt

ATTITUDE TELEMETRY . . . . . . Y , . 111-39

3.o-1 EXTRACT FROM t0i tDAtA (NORMALIZI) . . . ,

. . . . . . tiI-42

4.2-1 S192 MULTISpEcTRAL §AMNNil Mt4S1TiVITY (NOISE) . . . . . . 111-50

4,1-1 S19OA, Si1i i92 §tcmAt tAtb8tVOK WAntomTaitt

COMPARISON , . , , . . . t A . i. a . a 4 . . & . . 111-51

4.1,1,1 s1 1 PPdTRAt RADIANdZ VoR Wl A AND 8191 toMPARiSoN

SITES & , , a , s . . , a 4 ; , a I Ia a II . . . . 111-52

4.1.1-2 S190A, §191j AND q1tOUND-Tittii fl OMETtU UOMPARISON .. . . It-5

4.3.2-i 919OA, S1g ANtI (IROtlNt-tktft lDIOTfI dOOiAtfoN . . 111-54

4.3.-1 S11 IAtIANut at ktO diA1i

ReAT I EOtk tOk
UOMPARI0N OP S191 TO 8192 . , ,,.
* a
* -55

4.3,3-2 g1-91, 8190, AND tOUtD-TRUTt tIOfdTtltd eOMPAANtN . . 111-56

A.It-i 9A etE att tONST£tuCTkll TO tAt±ITATt dALUtLATION

O MAN tUtAit W IAtcttY BAND . . . . . I I I . . . . II-A-17

A. 11-2 ESttMATtD tUAf ADIANC NORMAL ZtD bY bANNf t . . a . . . III-A-17

 MSC-05546

FIGURES

Figure Page

3.3.2-1 Scan response ................. ........ III-10

3.3.2-2 Scan response (continued) ............ ....... III-l

3.3.2-3 Scan response (continued) ...... ....... .... 111-12

3.3.2-4 Scan response (continued) ........... .... . . M1I-13

3.3.2-5 Scan response (continued) .............. .I.I. 111-14

3.3.2-6 Scan response (concluded) ..................- 15

3.3.3-1 Residual effects (Z!) ........ ........... .. .111-16

3.3.3-2 Response versus scan angle ........ ...... . 111-17

3.4.1-1 Comparison of equivalent calibration lamp radiances . . . 111-20

3.4.1-2 Band 13-1 Y-3 CDDP thermal band calibration .. ...... . 111-21

3.4.1-3 Band 13-2 X-5 CDDP thermal band calibration ......... 111-21

3.4.2-1 Geometry of data utilized in off-axis rejection

determination .... ................ ... ...... 111-22

3.4.2-2 Off-axis rejection across track ...... .......... III23

3.4.2-3 Off-axis rejection along track ..... .......... . . -23

3.4.3-1 SL3 spectral response determination .... ......... ... 111-24

3.5.2-1 Average response versus distance from lunar edge for

band 1, SL3 .......... .............. .... .. 111-27

3.5.2-2 Average response versus distance from lunar edge for

band 2, SL3 .... ................. . ...... 111-27

3.5.2-3 Modulation transfer function (MTF) for band 1, SL3 . . 111-28

3.5.2-4 Modulation transfer function (MTF) for band 2, SLJ . . 111-28

3.5.2-5 Idealized modulation transfer function (MTF) . .... .. 111-29

3.5.3-1 Frequency response for SL4, Y3, LC4, band I ....... 111-30

II-vii
 MSC-05546

Figure Page
3.5.3-2 Frequency response for SL4, Y3, LC4, band 2 . . ....
. 111-31
3.7.1-1 Data-word histogram, band 1 ... .........I.... 111-35

3.7.1-2 Power spectral density plot, band 7 ...... .I.I....111-36

4.1-1 Scene data minus calculated real offset ..... .... 111-44

4.1-2 Comparison of SL3 scene (Sahara Desert) with prelaunch

data for band 3 ...... ...........I.I. ... 111-44
4.1-3 Comparison of S192 data product and Sl90A
densitometry data ...... ............ .I.I... 111-46
4.1-4 S190A computed scene radiance variation for Sahara
Desert .... ............... .I.I.. . 111-47

4.1-5 Comparison of S192 data for Sahara Desert corrected for

S190A scene radiance variation with prelaunch data
for band 3 ........ .......... .I....111-47
4.3.1-1 S191 spectral radiance plot for S190A and S191
comparison sites ...... ........... .I.I. .. 111-53
4.3.3-1 S191 spectral radiance of Rio Grande Reservoir, SL3 . . . 111-55

A.1-I Sample determination of ground-truth atmospheric and

target radiance for an S192 band ........ ..... III-A-5
A.II-1 Geometric albedo of the Moon .......... ....... III-A-I1

A.II-2 Correction for lunar phase angle ........ ...... III-A-15

A.II-3 Selection of lunar data ......... ........ ... III-A-19

A.II-4 A-trace of averaged scan line centered on lunar diameter

(LC 1, band 6) ........ .......... .... III-A-19
A.III-I Correction curves for clipped mean values .... ....... III-A-22

A.VIII-l Reflection from water surface at 0, 60, and 80'

incidence angles ............ ........... .. III-A-37

A.VIII-2 Reflectivity and emissivity of water verses incidence

angle .......... .................... III-A-37

A.VIII-3 Upwelling radiance calculation ...... ......... III-A-38

ItI-vii
 MSC-05546

1.0 INTRODUCTION

1.1 Purpose

This document reports the final results of the sensor performance evalu­
ation for the multispectral scanner (8192) of the Skylab Earth Resources
Experiment Package (EREP) and is based on data and evaluations reported in the
interim performance evaluation report (bSC-05528, Volume III, dated
September 6, 1974).

1.2 Scope

This document summarizes the results of S192 sensor performance evaluation

based on data presented by all contributors (Martin Marietta Corporation, the

Environmental Research Institute of Michigan, and the Science and Applications

Directorate of the tyndon B. Johnson Space Center) to the sensor performance

evaluation interim reports, provides the tesult of additional analyses of

S192 performance, and describes techniques used in sensor performance evalu­

ation (Appendix A). The summarization includes significant performance

changes identified during the Skylab missions S192 and EREF system anomalies

that affected S192 performance, and the performance achieved, in terms of

pertinent S192 parameters. The additional analyses include final performance

analyses completed after submittal of the 814 interim sensor performance

evaluation repotts, including completion of detailed analyses of basic perform­

atce parameters initiated during the interim report periods and consolidation

analyses to reduce independent mission data (SL2, SL3, and SL4) to determine

overall performance realized during all three Skylab missions.

1.3 Usage Guide

The basic task outline for the EKEP sensor performance evaluation was

specified EREP Mission Data Evaluation Requirements, JSC-05529, August 31,

1973. The results of these evaluations were subsequently reported in MSC-05528,

Earth Resources Experiment Package, Sensor Performance Report, Volumes I

through VIII as follows!

Volume I (Sl90A) Multispectral Photographic Camera

Volume It (S191) Infrared Spectrometer

Volume III (S192) Multispectral Scanner

Volume IV (S193 R/S) Radiometet/Scatterotneter

Volume V (9193 Alt) Altimeter

Volume VI (5194) L-Band Radiometer

Volume VII (SIOB) Earth Terrain Camera

ItI-i

 MSC-05546

These volumes were issued after prelaunch testing at KSC and updated after
each mission. The single exception is Volume VII (S190B), which was origi­
nally issued after SL3, with a single update after SL4.

This document is based on the data 'and'analyses in the first six volumes

of the sensor performance report, MSC-05528 (Volume VII, S190B, is not

included). The same volume designation used for MSC-05528 has been retained

for the individual sensor volumes, with the individual volumes bound in a

single cover and identified as MSC-05546. The individual volumes are designed

so they can be used independently of the 'full six-volume report, if desired.

1.4 Abstract

A prelaunch and flight performance evaluation of the iultispectral

scanner (S192) has been made and a summarization is provided. The following

aspects of S192 performance are among those discussed--absolute radiometric

calibration, noise, system frequency response, and scan-related response.

The primary performance problem was signal noise, especially in the thermal­
band when 'the Y-3 detector was in use. Final results regarding,adhieved

performance, significant performance changes, and anomalies are presented.

S192 accomplished its primary objective to simultaneously gather

quantitive radiance values suitable for image reconstruction in 13 spectral

bands.

Techniques used for sensor performance evaluation are described.

111-2

 MSC-05546

2.0 APPLICABLE DOCUMENTS

MSC-05528 Earth Resources Experiment Package, Sensor

Performance Report, Volume III (S192), Engineering
Baseline, SL2, SL3: and SL4 Evaluation; Lyndon B.
Johnson Space Center Houston, Texas, September 6,
1974.

MSC-05531 Ground Truth Data for Test Sites (SL2); Lyndon B.

Johnson Space Center, Houston, Texas, August 15,
1974.

MSC-05537 Ground Truth Data for Test Sites (SL3); Lyndon B.

Johnson Space Center, Houston, Texas, March 29,
1974.

MSC-05543 Ground Truth Data for Test Sites (SL4); Lyndon B4

Johnson Space Center, Houston, Texas, April 30,
1974.

111-3

 MSC-05546

3.0 SUMMARY OF SENSOR PERFORMANCE EVALUATION INTERIM REPORT

After preflight testing of EREP experiments at Kennedy Space Center and

after each Skylab mission, raw data from preflight tests and each mission vere

reduced to provide performance data for each EEP sensor; These data were

presented by mission in interim sensor performance evaluation reports entitled

EREP Sensor Performance Report (Engineering Baseline, SL2, SL3,. and SL4 Eval­
uation), MSC-05528, Volumes I through VII. -Preflight test data and selected

qualification test data were the engineering baseline, and flight data were

added after each Skylab mission. This section summarizes the sensor perfor­
mance report, MSC-05528, Volume III (S192), Change 3, September 6, 1974, para­
graph by paragraph. To provide traceability, applicable interim report para­
graphs in the summary are referenced.

3.1 Function/Limit Verification

Analysis of sensor performance properly began with an evaluation of those

parameters that indicated the relative "health" of the instrument. These were

instrument oriented measurements with expected values or limits. Evaluation

of the S192 multispectral scanner included verification of four such indica­

tors of sensor operation--housekeeping values, calibration data status, data

loss rate, and alignment history. The following paragraphs summarize the data

analyzed with respect to these indicators during evaluation of the performance

of the S192, during ground test, SL2, SL3, and SL4. The detailed evaluation

is in Section 3 of MSC-05528, Volume III, dated September 6, 1974.

3.1.1 Housekeeping Functions

S192 housekeeping parameters were monitored throughout the prelaunch and

orbital operations. All parameters indicated proper operation of the S192.

Maximum and minimum values observed during these operations are summarized in

Tables 3.1.1-1 and 3.1.1-2.

Table 3.1.1-1 summarized the operation of those parameters that were reg­
ulated. These data indicate that, throughout testing and the Skylab missions,
these parameters were controlled to within 0.30 X in temperature and less than
I milliampere variation itt calibration lamp current. The cold blackbody tem­
perature was the only parameter to show a larger variation, and that was only
1,50K,

Table 3.1L-2 summarizes housekeeping temperatures that were not con­

trolled. The in-flight range of values is divided into two columns. The first
dolunmn gives the range of values that odcrred during the first eight passes
bf SL2. buting this tiie, the Skylab wali heaters were hot used due to power
botstraits. The lt# tall tetipettiireb caused ieveral sensor temperature to
go below expected limits The S§econd t0dumn summarizes the range of tempera­
ture noted throghut the test ob the gt23 9L3, and 8L4 missions. During this

1I-4

 MSC-05546

TABLE 3.1.1-1.- CONTROLLED HOUSEKEEPING FUNCTION SUb*ARY

PARAMETER DATA WORDS EXPECTED RANGE TEST RANGE IN-FLIGHT RANGE

0
Hot Blackbody 109 318 - 322 K Not applicable 3210K

Temperature 126 318 - 322 K in ambient 321 0 K

conditions

Cold Blackbody it0 258 - 262 0 K Not applicable 260.0 - 260.5 0K

Temperature 127 258'- 262 OK in ambient 260.5 - 261.5 °K

conditions

1007 Earth Cal 1ll .Cal 3 HI 0- 20 m& 0 MA 0 mA

Lamp Current Cal 3 LO 180-220 mA 184 mA 184 mA

200% Earth Cal 113 Cal 3 LO 0- 20 mA "186 mA 186 mA

Cal 3 HI 180-220 mA

Detector Array 114 950 K 940 K 940 K

Temperature

time, all temperatures were Within expected limits, except for the door tem­
perature and scan-motor temperature. The door temperature dropped below ex­
pected limits only during lunar calibration passes, while-the scan-motor tem­
perature exceeded expected limits during most passes. The scan motor's nominal

operating temperature range was from 316 to 320 0K. This was hot considered an

anomaly but an error in determining the expected operating temperature based

on theoretical functioning of the heat pipe.

3.1,2 Calibration Data Status

Calibration data are the sensor outputs in counts when viewing the calibta­
tion sources inside the sensor. Table 3.1.2-1 sunmarizes the maximum and min­
imum mean values of the six calibration words from a few example data sets for
each band in eight configurations. Where a range bf valUes is riot given, only
one data set Was available. The output values given for the automatic gain
control bands (i.e., bands 4 through 13 without attenuationj and bands 6 and
'9 through 13 with attenuator installed) are mean values of the six calibration

words (which Vary itt count value due to the variation from uniformity of the

internal calibration source radiance) one or toreof which are usually clipped

at 255. The six calibration words vary from 255 because the radiance of the

internal calibration source is non-uniform and each of the six values is taken

from a different portion of the Source as the instantaneous field of viev scans

across it. The automatic gaih control Systems attempt to make the value on the

analog-to-digital converters either 254 or 2S5 after the sixth calibration

source value has been sampled. The last word is usually clipped at 255 and

the fifth word is frequently 255j but the first four words are seldom 255. The

randomness is due to the noise on the signals.

111-5
 REPRODUCIBILITY OF THE
MSC-05546 ORIGINAL PAGE IS POOR
TABLE 3.1.1-2.- UNCONTROLLED. HOUSEKEEPING FUNCTION SUMMARY

IN-FLIGHT RANGE OK
PARAMETER DATA WORDS EXPECTED RANGE TEST RANGE SL2 Passes SL2 Passes 9-I,
(OK) (*K) 1-8 SL3 & SL4

Primary Mirror 115 289 - 311 296 - 301 286.0 - 292.0 290.5 - 294.0
Temperature 117
Secondary-Mirror 116 289 - 309 298 - 301 286.5 - 292.0 289.0 - 294.5
Temperature
Aspheric-Mirror 118 288 - 312 296 - 299 286.0 - 292.0 290.0 - 293.0
Temperature

Door Temperature 119 284 - 322 297 - 300 284.0 - 289.0 279.5 - 289.5

Scan-Motor 120 289 - 316 304 - 306 302.0 - 320.0 300.0 - 324.0

Temperature

Cooler-Case 121 289 - 311 304 - 306 292.0 - 300.0 29.0 - 303.5

Temperature

InternalScanner 122 289 - 329 305 - 307 292.0 - 300.0 300.0 - 304.5
Assembly Temperature -

Monochromator 123 289 - 311 299 - 307 288.0 - 292.5 293.0 - 298.0

Temperature

Digital-Electronics 124 2S9 - 329 304 - 311 *Less than *Less than

Assembly Temperature 304.5 304.5

* Temperature was usually less than 303.161k, the 2ero digital count output.

TABLE 3.1.2-1.- CALIBRATION DATA STATUS

SYSTEMPESPONS TO HIGH CALthfATON bOUkC8 tCoudts)

Band Fee .an.h SLY S a SIA
nus atp o.tt- torx-5

1 NA 29.A 32.7 - 40.9 43.4 0.3 - 33.2 29.0 - 12.1 29.9 - 32.1 8.1 - 9.9

2 38.8 4, 112,8 - 39.5 40.8 i7, - Z8.2 2t.2 - 11.0 27.9 - 31.9 5.6- 7.3
3 59.5 b0.4 43.8 - 48.6 44.0 4L.6 - 45.s 41.1 - 4§3 42.2 - 49.6 17.8 - 20.4

4 95.6 24§.t TI.b - )g.S 149.0 9J.4 - 83.A 02.6 ioi.s 82.5 - 89.-7 246.5 - 247.8

E 102.9 13 11.2 - 90A 18J 11.9 - 1S.4 150.0 - B48.4 110.8 - 140.3 Inoperative

6 253.8 253.4 246.1 - 24.1 24.1 45.7 - 24tA. 245.A - 241.2 244.8 - 246.4 250.0 - 252.5

7 159.9 2SI.8 137.1 - 148.3 254.9 B32.0 - 156.9 1s.i - 162.6 183.5 - 204.5 82.1 - 96.3

8 184.4 833.0 itt. - i85.1 154.3 072.3 - 0i5A 171.6 - A1.3 185.8 - 201.3 252.4 - 254.9

9 253.6 251.7 23.1 - IS1.3 253.3 246.8 - 249,J 249,6 - ilU,1 247.9 - 249.7 216.6 - 218.4
10 253.1 253A 252.6 - 153.9 24.5 244.J - 245.1 244.1 - A6,4 241.2 - 246.6 Inoperative

11 254.4 t51.7 284.6 - 84,6 84y 241A - 148,5 47.3- 249. 246.6 - 250.3 253.0 - 254.2

L2 253.9 291,4 i§4,S - 44.t W.9, b19.2 - 230.5 238.1 - M0O.1 236.4 - 241.4 239.2 -247.2

1l-i NA flA 118,8 - 541.0s 241L6 24±.0 -241.7 W3L - 242.4 141.5 - 243.0 242.2 -243.9

13-2 NA R~A i53.1 - 235.0 288 .b 253 .6 - 255.0 135.8 - fAS.0 255.0 - 255.0 255.0 -255.0

11t-6
 MSC-05546

3.1.3 Data Loss

Data loss rate was evaluated by analysis of the frame synchronization pat­
tern and housekeeping data. The frame synchronization pattern was a2 fixed

digital output in which errors could be detected. Housekeeping temperature

data were evaluated for the percent of erroneous data presented. Due to the

predictable nature of thermistor networks, anomalies were easily identified and

a loss rate was determined. Since the synchronization patterns were inserted

into the data stream at the output register of the S192 electronics assembly

the data losses due to faulty synchronization patterns were almost certainly

due to the data handling systems (probably in the Miller encoding/decoding

operation or in reformatting). The faulty housekeeping temperature words could

have arisen at any point -in the data path--within S192 or without.

Loss rates were determihed for prelaunch, SL2, SL3, and SL4 data. For

prelaunch data; the average loss rates were as high as 0.35%. The Skylab data

average loss rates varied from 0.001 to 0.01 percent.

There were data losses certainly not atributable to S192. As a result of

inconsistencies in cleaning the tape recorder heads and the high temperature

environment experienced by tapes launched on SI, data were lost from bands 3

and 11. Because both of these bands were doubly sampled, the data loss was

minimized. These were most notable on SL4 passes 44, 47, 48, 49, and LC6.

3.1.4 Alignment History

The alignment of the detector array to the incoming radiation was a crew

procedure. Voice transcripts of crew comments recorded while performing the

detector alignment verifications were reviewed to determine whether the proper

-alignment was achieved. During SL2, the cooler/dewar/detector/preamplifier

assembly was improperly mounted to the optical bench, causing difficulties in

alignment. A good alignment was not achieved until before pass 4. Alignments

for the remainder of SL2 and during all of SL3 and SL4 were considered good.

3.2 Data Interference

An analysis was performed to determine any effects of interference from

external sources detectable in the S192 data. The basic approach used was to

evaluate S192 data taken while other EREP sensors and module systems were in

various modes of operation. The data analyzed for comparison were the frame­
synchronization pattern, detector output data taken of a uniform scene, and

the noise characteristics of the data.

Data from prelaunch testing and SL2 were analyzed in detail. The results

indicated that S192 data were hot susceptible to interference from potential

sources in its operating environment. A detailed analysis was hot performed

on data from S3 and SL4; however, the results of the noise analysis (See

Section 3.7.)'did not indicate the presence of intetEetence.

Details of the data interference aftalysis were presented in Section 4 of

MSC-05528, Volume III, September 6, 1974.

111-7

 NSC-05546

3.3 Response versus Scan-Angle Determination

S192 scan-related system response was determined using the techniques pre­
sented in Sections III and IV of Appendix A of this volume. Detailed results

are presented in Section 5 of MSC-05528, Volume III, September 6, 1974.

3.3.1 Scan-Related Signal Offset Bands 1 through 12

To determine S192 response as a function of scan angle, it was necessary

to first determine the response to a zero radiance level as a function of scan

angle, that is, signal offset. Using the technique described in Appendix A,

Section III, the mean value was determined from clipped digital data values on

a pixel-by-pixel basis across the scan. Deep-space data from the lunar passes

were used as an effective zero radiance-input. Results from the six"lunar

passes performed during the Skylab missions indicated that the offsets were

not constant. For five of the lunar passes, the offsets showed similar

characteristics of shape. The other lunar passes yielded offsets with the

same frequency characteristics but shifted in phase. Additionally, it was

noted that, for a given lunar calibration, all bands had the same shaped

offset function, with the major difference band-to-band being a larger

amplitude for the fixed-gain bands than for the variable-gain bands. It

was concluded that the offset was due to scan synchronous electrical pickup

where the phase of the components varied. Examples of these characteristics

are given in Section 4 of this volume.

3.3.2 Scan-Reiated System Response Bands 1 through 12

Scan-related system response was determined using the techniques described

in Section IV of Appendix A. The response was evaluated for prelaunch and each

of the three Skylab missions.

3.3.2.1 Prelaunch - Prelaunch testing was performed using the optical

functional checkout unit as the external radiance source. This provided a

'tscene" known to be'a uniform function of scan angle. The curves determined

for scan-related system tesponse were therefore the S192 response to a uniform

input as a function of scan angle. rigures 3.1,2-1 through 3.3.2-6 present the

normalized prelaunch curves Eat bands I through 12 as well as the flight re­
sults. Band 13 could not be determined before launch due to constraints on

thermal band operation in a ground envitotdnmht. Except for band 1, attenuator­

in curves are shown foEt prelatch results; however, as would be expected, in­
sertion of the Attetitiatqr Agsembly makes tio significant difference to scan­
related system response.

Iti-8

 MSO-05546

3.3.2.2 ,Scan-Related System Response Band 1-12

The prelaunch and flight tesults are given in Figures 3.3.2-1 through

3.3.2-6.

The extent to which the flight curves are in fact affected by scene de­
tails is uncertain, so some subjective judgment is required in deciding whether

significant changes have occurred. The undulations in many of the SL4/X5 re­
sults are probably due to scene details but the X5 curves do appear signifi-,

cantly changed from the Y3 results, particularly in the lower numbered bands.

For Bands 4 through 8 the flight and prelaunch data tend to cluster about a

common mean, though individual flight curves may look significantly different

from their prelaunch counterparts. In Band 9 the flight curves slope down to

the right a good deal more than do the prelaunch curves. The same effect

occurs to a lesser extent for the SL2 and SL4/Y3 curves in Band 11. In addi­
tion, the prelaunch curves for Band 9 are much flatter than for 7, 8 and 10,

and it has been suggested that the Band 9 prelaunch curves contain an error.

It is, however, difficult to see how such an error could have been introduced.

The Band 1 flight curves appear much less arched and overall have a steeper

slope up to the tight than prelaunch. This effect repeats to a reduced extent

in Band 2 and perhaps, but hardly significantly, in Band 3. The SL4/X5 data

are an.'exception in that the slope up, to the right is -continuous with only a

little abatement through Band 3. For Bands 10, il and 12 the SL2 and to a les­
ser extent the SL4/Y3 data depart increasingly ftom prelaunch, becoming more

arched and steeper downwards at the right end,

There seems little doubt that the scan angle efforts have changed in Some

details from prelaunch and throughout the missions. However, particularly for

the middle bands; the evidence of change is ott substantial and selection of

the best description of these effects id a difficult and inevitably subjective

one.

It should also be remembered that each of the flight curves is formed by

averaging 4 td 6 cUrves, each derived from individual scenes observed during

the mission in question. Thus when making a decision as tz whether the flight

results should be used in place of the prelaunch curves, the spread of these

individual curves was compared with the difference between their mean and the

prelaunch curves. Complete details will be found in the individual evaluations

fot each mission, MSC-05528,

'tzt.-9
 MSC-05546

1.2­

~. 4 X

Prelaunch Resulsi - -- -Lt

W1.0""

iS 2 SL4.Y -.

10.9

0 .526 2 378 504 630 71 2 008 1134 1280

Band 1

1.2­

,9

<it-lu

 MSC-05546"

1.2­

01.1-
SL 4Y rl u c e ut

SL BaSd 2

1.0- 0 2 5 318 54 630

PIXEI.S
756 882 1008 1134 12600

0 126 252 378 504 630

756
PLXE7S
882 1008 1134 1260

Band 4

f1gure 33.2-2.- Scan te-ponse (continued).

rti--li
 MSC-05546

1.2

01.1

PIXELS

Band B

l. 1'slaunch l c

1.2

0 126 252 878 504 630 756 882 1008 1134 1260
pinds

IN
1 .midd

ttt-I. . . . . . .
 MSC-05546

1.2

~SL 4X
. SL 3
N __relaunch Results

L9 SL 2

0 126 252 378 504 630 756 882 1008 1134 1260
PIXELS
Band 7

1.2 -

Pr launch Res t

0 126 252 378 504 630 756 882 1008 1134 1260
PIXE LS

Band 8

'Figure 3.3.2-4.- Scan respbnse (continued).

111-13

 MSC-05546

01.1- S
1'. 22L

0.9

I II I I
00 126 252 378 504 630 756 882 1008 1134 1260

Band 9

1.2

1.10-

0.8I I I
t26 252 378 504 630 756 882 14 1260
,,IKEIA

bnd io

ttgti'e IA.2.A. - 1ntojio (CnntjtiUed)i

Mt-14
 MSC-05546

L2

01.1-
P4SL 2

~ 10-- - Prelaunch Results

0.9

0 126 252 378 504 630 756 882- 1008 1134 1260
PIXELS
Band 11

1.2 -

(*2

S4L4Y".

: Prelaunch Results

0.8­

I I I II I ,
0 126 252 378 504 630 756 882 1008 1134 1260
PIXEL

Band 12

figtite 3.3.2-6.- Scan tedponse (condluded).

ITT-i5

 MSC-05546

3.3.3 Scan Related Signal Offset and System Response - Band 13

The response vs scan angle (R) and residual,effects (Zj) as defined in

Appendix A, Section IV, are given in Figures 3.3.3-1 and 3.3.3-2. Bands 13-1
and 13-2 are equivalent but have different calibration source offsets and
gains, so that 13-1 has an extended dynamic range. No prelaunch data are
available for comparison. The results for the Y3 detector for the three mis­
sions are quite similar, though as might be expected the results for the X5
detector are significantly different.

-oSL4X ------ SL­

I I I I I I I I
o 126 252 378 504 830 756 882 1008 1134 1260

PIXEIS
Band 13-1

50­

25 SL44 -Y
0. . -X- S 4 L

- I I I I I I I I
o 126 252 378 504 630- 756 882 1008 1134 1260

Band 13-2

Figure 3.3.3-1.- Residual effects (Zf).

I1I-16

 MSC-05546

1.2 "

n1.0

-~ ~ 4' .-

o 126 252 378 504 630 756 882 1008 1134 1260
PIXEI S

Band 13-2

1.2-
Iz
0.
0 I I I I I I I I 1 .

10

0.8 1 1 I 1 1 1
0 126 252 378 504 63 756 -882 10108 1134 1260
PIXE LS

Band 13-2

Figure 3.3.3-2.- Response versus scan angle.

3.4 Absolute Radiometric Accuracy and Stability

S192 radiometric accuracy and stability was determined by analysis of the

radiometric calibration of the S192 reflective bands (1 through 12) and the

thermal band (13)i The rejection of radiation reflected from sources off-axis

to the S192 was evaluated, as was the stability of band locations. Details of

these analyses were presented in Section 6 of MSC-05528, Volume III, Sep­

tember 6 1974.

111-17

 MSC-05546

3.4.1 Radiometric Calibration

Radiometric calibration of S192 was performed in two parts. The reflec­

tive bands were calibrated using prelaunch response to a test target, ground,
truth data, and lunar radiances. The thermal band was calibrated using ground
truth data from lakes used as thermal targets.

3.4.1'l Reflective Bands (1 through 12) - Data were evaluated from pre­
launch testing, five lunar calibration data passes, and the Great Salt Lake
Desert (with ground truth data) on the SL3 mission. The evaluations were per­
formed using the techniques described in Sections I, II, and III of Appendix A.
Calibration lamp radiances consistent with these data sets were determined and

are plotted in Figure 3.4.1-1.

The ground data results are not self-consistent and differ from the lunar
data results. The lunar data results are quite self-consistent, but both lunar
and ground data results differ noticeably from the prelaunch results. A crit­
ical examination of these various results does not identify specific causes for
the differences but does identify candidate sources for some error. For exam­
ple, the ground data results are based upon a single data set from each instru­
ment which would permit undetected systematic instrument error, and the trans­
fer to the spacecraft requires correction for atmospheric effects which is

still a subject for research. The lunar data results also must include correc­
tion for atmospheric effects in the results from the literature, and the

Irvine* & McCord** data are combined and extrapolated to cover the S192 range.

Although the final prelaunch calibrations were performed under far from optics

lab conditions, the external source used was a well understood test target

which was recalibrated after launch. For these reasons, the prelaunch values

for calibration source radiance were used for S192 production data processing

although the possibility of sensor change following thy last prelaunch calibra­
tion cannot be dismissed. There can be little questions that the sensor re­
sponse in Band 1 changed before the first SL4 lunar calibration pass, but with
that exception the lunar calibrations indicated good radiometric stability.

Ground truth data in Figure 3.4.1-1 were derived from two separate instru­
ments. Data that determined curve G were taken with an interference-wedge
spectral-scanning spectroradiometer, manufactured by Instrumentation Special­
ties Company, Lincoln, Nebraska (I.S.C.O.). Data that determined curve B were
taken with a Bendix Model 100 Radiant-Power Measuring Instrument (RPI).
Ground truth radiance data resulting from these two instruments differed in
magnitude by about 20% on the average with the radiances derived from the RPMI

*Adair P. Land & William M. Irvihe: "Monochromatic Phase Curves and Albedos

for the Lunar Disk," Astronomical Journal§ April 1973.

**T. B. McCord & T. V. Johnson: "Lunar Spectral Reflectivity (0.30 to 2.50

microns) and Implications for Remote Mineralogical Analysis," Science,

August 8, 1970.

iii-18
 MSC-05546

being the higher. These results are reflected in the calculated lamp radiance

values plotted. This deviation in magnitude, while spectral shape is similar,

can be considered to indicate systematic errors.

The lunar calibration data shown in Figure 3.4.1-1 were determined from
data sets taken with the S192 scanner by orienting and rotating the spacecraft

so that the moon was viewed and scanned. The spacecraft pitch rate was such

that the scan-line advance rate was only about 5% of that normally used for

terrestrial scenes, thus nearly 1000 scans of the moon were made on each lunar

calibration pass. Using lunar radiance characteristics from the literature*,

the radiance of the S192 -internal caibratio- lamp was calculated. These re­
sults demonstrate significant changes in the calibration of Band 1 over the
missions and indicate changes in the calibration of the lower numbered bands
between LC I and the later missions. Otherwise the results indicate that the
S192 was quite stable during those parts of the three missions when the Y3 de­
tector array was in use. Discrepancies of up to 50% between corresponding KSC
and flight results for Bands 2 through 11 are not too surprising in view of thd
many uncertainties involved in both sets of measurements and particularly in
interpreting the radiometric data available for the moon. These problems are
particularly severe for Band 12, but the 100% discrepancy for this band appears
too large to be accounted for in this way, so the results may indicate a real
change in performance. In fact, the changes in the radiance profile of.the
high (internal) calibration source do suggest real changes in calibration in

all bands following the prelaunch calibration.

3.4.1.2 Thermal Band Calibration - Calibration of the thermal band was

verified by using lakes as ground- truth targets, as described int Section VIII
of Appendix A. Data were evaluated for the Y-3 from SL2 and SL3 and the X-5
from SL4. The results indicated that a linear fit between the hot- and cold­
temperature blackbodies could be used to determine the radiometric temperatures
of ground targets to within approximately 1K. Figures 3.4.1-2 and 3.4.1-3 ­
show the results obtained. Figure 3.4.1-2 is a linear plot between the radi­
ances of the hot and cold blackbodies for band 13-1. The band 13-1 ouput from
the cold blackbody was set to approximately 60 counts and the hot blackbody to
240 counts. Results of ground truth measurements of radiance versus S192
counts for the Great Salt Lake on S3 are also plotted. Figure 3.4.1-3 pre­
sents similar data for the X-5 data of Laguna Reservoir on SL4. This is an
example of the results obtained for the smaller dynamic range band 13-2, where
the hot- andcold-blackbody radiances are set to 255 and 0 counts, respectively.

3.4.2 0ff-Axis Rejection

The rejection of off-axis radiation by the S192 was 'evaluated using the

technique described in Section V of Appendix A. Using the lunar calibration

pass data, adjacent scan-line data were evaluated as the 8192 instantaneous

field of view (FOV) was rotated onto the moon, and adjacent pixels were evalu­
ated as the 8192 scanned across the moon as shown in Figure 3.4.2-1. Data were

*Ibid.

111-19

 LEGEND
IT.
PRELAUNCH
AVERAGE
-%" SL2 LUNAR CALIBRATION

... ". SL LUNAR CALIBRATION AVERAGE

,' m SL4 LUNARCALIBRATION AVERAGE LCA & LC5

/ /" GROUND
TRUTH S13

S/ BROADBAND

T- /

zz,
- L
14

3 4 5 6 7 8 9 10 1i 12

Figure 3.4.1-1.- Comparison of equivalent calibration lamp radiances.

 mSC-05546

2" IIHCAL

200

160
O

NORTH SIDE S M SIDE

SOUT SIDE

212G

LO CAL

40­

0.2 0.4 0.6 0.8 1.0 1.2 1.4

2
RADIANCE CmlJ/es
-rr-stor)

Figure 3.4.1-2.- Band 13-1 YV3 CDDP thermal band calibration.

I CAL

240

20D

160

12(

Sso,
+12-C

mCAL /­
0.2 0.4 0.6 0.8 1.0 1,2 1.4

HADIANICE, (UM/cm'.-vm-ster)

Figure 3.4.1-3.- Band 13-2 X-6 CDDP thermal band calibration.

111-21

 MSC-05546

evaluated for the five lunar calibration passes, using the Y-3 assembly. Fig­
ures 3.4.2-2 and 3.4.2-3 are sample plots of the data obtained. Figure 3.4.2-2

is a plot -of cross-track pixels versus counts. The distinct rise as the S192

scans across the lunar limb, with no detectable scattered radiation in the deep

space data before the moon, indicates excellent off-axis rejection. Figure

3.4.2-3 is a similar plot for the along-track data. Due to the slow rate of

Skylab rotation of the S192 field of view onto the moon, the scan lines over­
lapped extensively. The pixel scale on the abscissa is the actual S192 instan­
taneous FOV accounting for this overlap. As in Figure 3.4.2-2, the indication

is that the off-axis rejection is excellent.

Moon Direction of Scan

Across Moon

Cross.
Scan Line Track
(Overlap Not Data
Illustrated)l

(Relative
Size Not
to Scale) ./
Sky)lab
• , _ _,, _ _-_-_- Rotation
...... / of Size
M.OVOn
Along Track Data .oo
Figure 3.4.2-1.- Geometry of data utilized in off-axis rejection determination.

3.4.3 Spectral Band Locations

Spectral band locations were verified by the Environmental Research Insti­

tute of Michigan. The technique hsed was to compare the S192 response to an
agricultural scene with the "predicted response" for that scene. The S192 re­
sponse was determined by taking the average counts for each band from a block of

pixels tdken over an agricultural area. These averaged counts were made com­
parable by using the tadiomettic calibration results of the lunar calibrations

(See-paragraph 3,4.1.1). The "predicted response" was derived by multiplying

a textbook "green vegotation" reflectance spectrum by a tettestrial solar il­

lumination spectrum fot two air masseds then integrating over the S192 bands.

Figure 3.4.A-1 Es an example of the results obtained. The dips at bands

5 and 9 show the chlorophyll absorption band and the 1.13-micrometer water­
vapor absorption band, respectively. the relatively high signal at shorter

wavelengths compared to the predicted curve is due to path radiance.

111-22

 MSC-05546

8 . •..

E­ 20.

i0 20 30 40 50 60 70 8

PIXELS

Figure 3.4.2-2.- Off-axis rejection across track.

0.

60­

10 303 - 50 0 ; 0
1201

PIXELS

' 0 Figure 3.4.2-2.- off-axis rejection across track.

U , ",• -,',.o ,, . .. * I..!

U, • • !

0 1 2......... 8 10 11! 12 13
PIXELS

Figure 3.4.2-3.- Off-axis rajection along track.

111-23
 a­

6-ass 14

P4

N?

'-4­

\-redicted Response
(Neglecting Path Radiance)
'1121314 WE IIE II s 9 11 IIi ' 1
0.4 0.8 1.2 1.6 2.0

WAVELENGTH AND PRELAUNCH BAND LOCATION (ym)

Figure 3.4.3-1.- SL3 spectral response determination.

 MSC-05546

Data tiere evaluated from the three Skylab missions. The location of the

two absorption bands previously mentioned indicates that the band locations

were essentially as measured before launch.

3.4.4 Dynamic Range

During prelaunch testing the internal calibration source for the reflec­
tive bands (1 through 12) was discovered to have a radiance too low to give the

necessary dynamic range in bands 1, 4, 5, 7, and 8. Since it was too close to

scheduled launch to make changes in the basic hardware, an attenuator assembly

(installed on the connector of the signal output cable of the preamplifiers and

containing voltage dividers including adjustable trim-potentiometers) to reduce

the signals from the detectors in those bands and thereby increase the dynamic

range. The attenuation in band 1 caused adverse frequency response character­

istics (see paragraph 3.5.1), so the trim-potentiometer was adjusted to its

lowest resistance making the dynamic range too small in band 1.

Since the Attenuator Assembly was not launched in place but was stowed

elsewhere in the MDA and since the Attenuator Assembly had been exercised in

the system minimally, ERE? passes (1, 2, and 3) taking S192 data without the

attenuation were flown before the Attenuator Assembly was installed between

passes 3 and 4. As expected a large- amount of signal saturation (i.e.,

clipped at 255) on ground scenes was experienced in bands 1, 4, and 5 and les­
ser amounts in bands 7 and 8. With the Attenuator Assembly installed only band

I still saturated ori ground scenes at high sun elevation angles.

Since the increase in dynamic range was produced by a reduction in signal,

it reduced S/N (signal-to-noise ratio). In bands 4 and 5 the decrease in sig­
nal was found to be greater than required, so between passes 24 and Z5 the at­
tenuation in those bands was adjusted with an accompanying change in dynamic
range.

The Attenuator Assembly was also determined to have increased the noise

level by permitting more "pick-up" of electronic noise. A new Attenuator As­

sembly with a modified grounding scheme was sent up with the SL4 crew and in­
stalled before pass 53. This new voltage divider network changed the attenua­
tion and therefore the dynamic range in each attenuated band but not signifi­
cantly.

The X-5 C/D/P (Cooler/Dewar/Preamplifier Assembly) was installed in place

of the Y-30/D/P on January 15, 1974, between passes 83 and 84. This C/D/P had
different detectors and preamplifiers so that the Attenuator Assembly was not

installed twith it. The dynamic ranges in bands 1, 2, 3, and 7 were drastically

increased by factors of 4, 5, 2.5, and 2, respectively, with commensurate

changes in SIN. Bands 5 and 10 were inoperative.

3.5 System Frequency Response

S192 system frequency response evaluation determined the ability of the

sensor to respond as a function of ground spatial frequency. The following is
a summary of this evaluation. This task is described in detail in Section 7 of
MSC-05528, Volume III, September 6, 1974.

fI-25

 MSC-05546

35 S tem e4uency Resp6nse' -Prelaunch

A dharacteristic apparenti& inh-rent in the high performance.1{gCdTe detec­

tor material dsed inS 192 is frequency response that varies-nnon-linearlywith:
incident radiance level. - Fixed radiance levels (called opticad Aias) iere 4p­
plied to the detectors to Adjust frequency response. Since sensiilvity,Also
varies inversely with incident radiance level a trade-off was maddbetwee.n6p­
timumi.frequency resp6nse and sensitivity in setting the optickl bia§ -levels.
The preamplifiers were then trimmed to give optimum performance in each chainel
in the range of radiance levels for each spectral band expected from terres­
trial .scetes of greatest interest. At radiance levels different from the trim

levels frequency response changed from that, existing at the trim'levels.

In preliunch testing at the Honeywell Ridiation Center4 bar'targdts (wth­

uniformly spaced bright and dark bars giving different spatial frequencies)
were used to determine the frequency response characteristics (f bands I
through 12 at three illumination levels (trim level, 1.5 trim and 0.5 trim).
These results are presented in MSC-05528, Volume III (S192)* and showed fre­
quency response in several bands to be significantly different from that de­
sired but in no case to be unsuitable for use.

In prelaunch testing at Kennedy Space Center the voltage divider network

of the Attenuator Assembly when combined with the capacitive coupling of the

alignment signal circuit in bdnd 1 destroyed the frequency response of that

band. There was not sufficient time to change the Attenuator Assembly other

than to set the band 1 adjustable resistance to minimum. This left band 1 with

such poor frequency response with the Attenuator Assembly 'installed that the

data are generally unusable.

3.5.2 System Frequency Response - SL2, SL3

The technique used to obtain S192 system frequency response in orbit is

described in Section V of Appendix A. The modulation transfer function (MIF),
which is a measure of .that response, was determined from S192 output as the
sensor scanned from the moon to deep space.

Results for S2 and SL3 were very similar. For bands 3 through 13, the

system frequency response was satisfactory. As expected, band 1 was very poor
at all frequencies; band 2 was also poor but to a much lesser extent. Average
responses of band 1 and 2 as S192 scanned off the trailing lunar edge during
SL3 are shown in Figures 3.5.2-1 and 3.5.2-2. These average responses are a
pixel-by-pixel average of 20 scan lines. The corresponding MF curves are
presented in Figures 3.5.2-3 and 3.5.2-4. An idealized MTF curve is shown for
comparison in Figure 3.5.2-5i

*Skylab -Program Earth Resources Experiment Package Sensor Performance Report,

Volume III (S192), Engineering Baseline, SL2, SL3, and SL4 Evaluation,

MSC-05528, dated September 6, 1974.

111-26

 MSC-05546

100

01I -- ----- I

o 50 4 -L-
P44 --H 1. . ..
L" . Ai- - T 42. ­

0 50 100 150 200 250

DISTANCE (pixels)

Figure 3.5.2-1.- Average response versus distance from

lunar edge for band 1, SL3.

. . . . . . .. .. . . . . ..

a i [IIIiFI- -II nIIi z iI ­

100
* I 1 ! I 1 II 1 it!1 ! II I I

tt±ll i, I W 4iJ

i (I I I I * l ,_ _ -- .LJ rt
p__-- _ . ...... u- ". Z ;- -L
5zI

.I i-

O I 4-- _

0 50 100 150 200 250

DISTA[CE (pixels) -

Figure 3.5.2-2.- Average response versus distance from

lunar edge Lot band 2s SLS.

111-2 7

 MSC-05546

ri0 I ',7

Im 1ILOn 1,, , I i L1 M
__..~
'-p*
~~:I'LT -Il , ' - T: Fm ,,

0, ip . llll liil l l 1ll4 , IeI t id

0 183

FREQUENCY (kliz)

Figure 3.5.2-3.- Modulation transfer function (MTF)

for band 1, SL3.

*-. i- -: c'z'
-.. It L
_5, ......
L '.­I . . ..

o 183
FREQUENCY (kllz)

Figure 3.5.2-4.- Modulation transfer function (MTF)

far band 2, SLJ.

111-28

 MSC-05546

0 183

FREQUENCY (kRz)

Figure 3.5.2-5.- Idealized modulation transfer function (MTF).

3.5.3 System Frequency Response - SL4

An improved attenuator was installed in S192 at the start of SL4. This

considerably improved the system frequency response for band 1 but caused some

overshoot in bands 2, 3, and 4 and decreased the signal-to-noise ratios in

bands 1 and 2. Figures 3.5.3-1 and 3.5.3-2 show the average response when

scanning off the trailing lunar edge and the MTF curve for bands 1 and 2 during

SL4 before the Y-3 detector array was replaced with the X-5 detector array'.

System frequency response for the X-5 detector array caused extensive

overshoot for many bands following the trailing lunar edge so that the digital

output was completely clipped. This meant the the true shape of the average

response could not be determined. An estimated MNF was obtained by fitting a

series of exponentials to the signals at the trailing edge of the moon, using
the positive overshoot at the leading edge as a guide. The resulting ?CFs
showed that most bands have cohdlderable frequency boost in the lower half of
the frequency range. Filters have been developed to permit improvement of the
reflective region data (bands 1 thr6ugh 12) taken with the X-5 array. This
study was not completed tt time for inclusion in this report but will be pub­
lished under separate cover,

111-29

 mSC-05546

- I 1.~ 1' I

to I I
to

4J

P44

0 183

a.
FREQUENCY (kHz)

100

av
o 500­

01 . !-

50 100 150 - 200 250

DISTANCE (pikels)

Figure 3.5.3-1.- Frequency responuo for SL4, Y3, LC4, band 1.

111-30

 MSC-05546

a . F
i t tI
LI

-- ..
-I--...
_=
..­

4I o 183
a.
-i
FREQUENCY (Mz)

100­

0 L Li
I' L"
I, i
50
U, ..50 + 100, 15 - 250

I+ij + ­

. .
'H '

. .
-"--

2
:

all r,,.:;l: I- h ;F aii~ rj'

I
.
a ai2 - ' e , , *' -;i
Ia i- .'l ja ',I,L
0 T , i. I[1;t 2 2Q HF LL . lI E 1.1 A
'
0 50 100 150 200 250

DISTANCE (pixels)

b.

Figure 3.5.3-2.- Frequency response for SL4j Y3, LC4, band 2.

111-31

 NsC-05546
3.6 Geometric Distortion

The geometric fidelity of S192 data was evaluated by:

1) Mathematical modeling of the data acquisition system to determine po­

sitional correction parameters;

2) Determination of the geometric distortion in the screening film.

The evaluations are presented in detail in MSC-05528, Volume III, September 6,

1974, Section 8, and are summarized below.

3.6.1 Geometric Distortion from Mathematical Modeling

The approach used to evaluate the geometric distortion is described in

Appendix A, Section VII. For SL2, SL3, and SL4, the mathematical model was
successful in handling time-dependent systematic errors through a determination
of the positional correction parameters. Residual errors (after eliminating

the systematic errors) were random, with a standard deviation of 4 pixels. No

prelaunch evaluation was made.

3.6.2 Geometric Distortion in Screening Film

Screening films were generated on a JSd production film processor by using

an algorithm that generated conical scan lines on the film. These scan lines
were obtained on the special raster that was guided by knowledge of the length
of the scan-line arc in degrees, velocity of the spacecraft in meters per
second, altitude of the spacecraft in meters, scans per second and degrees of
drift angle due to the earth's rotation. The width of the image was constrained
to be 5 inches which fixed the number of raster positions available in the

y-direction at a number that did not permit placing the pixel dot at its exact

computed location. Screening films thus generated were evaluated for along­
and cross-track scales (Sx and Sy) and nonorthogonality between X and Y axes.
The scales Sx and Sy differed by as much as 12 percent, and were nonorthogonal
by approximately 0.055 radian. All of the screening film had been produced by
the time these geometric imperfections were identified. Since these imperfec­
tions did not prevent or even detract from the use of the screening film for

its intended purpose of previewing sensor data and correlating data acquisition

time with-an investigator's area of interest, no screening film was generated

with improved geometric quality. The scale of the screening film provided on

5-inch filmwad found to be in the neighborhood of 1:800,000.

There were many other film products to be made for-investigators so the

pixel placing program of the production film converter was corrected by insert­
ing more exact values for scan-line arc length and scan rate, adding 3.15'
(0.055 radians) to the ephemeris drift angle and not requiring the image width
to be exactly fiW. itlchO. This resulted in reducing the difference between
across-ttadk dd ktdhg-tfack scales to about 1Z and in reducing nonorthogonal­
ity to less thAi 0&005 radian,

111-32

 MSC-05546

3.7 Noise Analyses

Noise analyses were performed on S192 prelaunch and flight data. Detailed

results are presented in section 9 of ISC-05528, Volume III, September 6,

1974.

3.7.1 Noise Analyses - Prelaunch

Prelaunch raw data were evaluated for noise by five methods. The perform­
ance data derived by each method are summarized below.

3.7.1.1 Standard Deviation - Standard deviations were determined for each

band at the center of the scan line from 1000 consecutive scan lines. Repre­
sentative values are given in Table 3.7.1-1, when the attenuators were not in­
stalled and Table 3.7.1-2 when they were installed.

TABLE 3.7.1-1.- STANDARD DEVIATIONS,

ATTENUATORS OUT

STANDARD DEVIATIONS (counts)

Prelaunch SL2

1 6.9 *
2 4.1 4.7

3 3.3 3.5

4 15.6 *
5 11.4 12.7

6 4.9 6.3
j 4&4 4.8
8 4.9 4.9

9 4.2 3.7

10 6.1 6.0
11 2.6 3.2
12 4.2 3.2

* Insufficient data

3.7.1.2 Strip-Chart Recordings - Strip-chart recordings for each of three

points in the scan line as a function of scan line, or time, were examined. A
very low-amplitude, low-frequency modulation was found. Additional evaluation

indicated that this modulation was 1/f noise (flicker noise) generated by the
detector where f was the noise frequency.

3.7.1.3 Data Histograms - Data histograms for all bands were made using
the data word at the center of the scan line for 1000 consecutive scan lines.

Figure 3.7.1-1 shows a histogram for band 1 and is typical of all the S192
bands.

111-33

 JrOnJCIBIITJY OF THE
is-05546 {4PAI, PAGE 18 POOR
TABLE 3.7.1-2. - STANDARD DEVIATIONS, ATTENUATORS IN

STANDARD DEVIATIONS (counts)

SL

Before After

BAND PRELAUNCH SL2 Attenuator Attenuator SL4

Change Change

1 7.7 7.2 7.2 *

2 4.1 4.2 4.0 4.0 3.9

3 3.3 3.2 3.1 3.1 3.2

4 7.8 8.2 7.7 9.2 9.4

5 6.1 5.i 5.9 8.2 8.0

6 5.0 6.5 5.6 6.0 5.7

7 2.8 3.0 3.5 4.3 4.2

8 5.6 4.7 4.7 6.2

9 3.3 * 4.5 3.7 3.6

10 5.4 * 6.6 6.0 5.8
11 2.6 * 2.7 2.7 2.7
12 3.6 * 3.1 3.1 3.3

* Insufficient data

3.7.1.4 Power Spectral Density - Power spectral density (PSD) plots (de­
scribed in Section VI of Appendix A) exhibited low-frequency broadband noise.

Figure 3.7.1-2 shows a typical power spectral density plot for frequencies less

than 100 hertz. The 1/f noise can be seen between 1 and 10 hertz. The spike

between 20 and 30 hertz corresponds to the cooler piston frequency.

3.7.1.5 Noise-Equivalent Spectral Radiance - Noise-equivalent spectral

radiances (NESR) were determined by converting the standard deviation in counts

for each band to spectral radiance by means of the radiometric calibration.
Tables 3.7.1-3 and 3.7.1-4 give the noise-equivalent spectral radiance corre­
sponding to the standard deviations of Tables 3.7.1-1 and 3.7.1-2.

3.7.2 Noise Analyses - SL2, SM3, and SL4

deviations and noise-equivalent spectral radiances for bands
-Standard

1 through 12, for SL2, SL3, and SL4 are shown in Tables 3.7.1-1, 3.7.1-2,
3.7.1-3 and 3.7.1-4. The values listed are from the Y-3 detector array only.
The X-5 detector array noise levels averaged about 50% higher than the Y-3
array.
The thermal band, band 13, Was not given a full system test before launch­
ing, so no prelaunch standard deviatiotis and ndise-equivalent temperature dif­
ferences (NEAT) were available to dompare with the flight values summarized in
Table 3.7.2-1. There were two outputd from band 13, labeled 13-1 and 13-2,
which differed from each other ty sodla end reference level. Table 3.7.2-1
shows the noise values for both, The valued from the Y-3 detector array and
its replacement, the X- detector atray, are given under SL4.

ti±-34

 10 "

9­

7i

I U I

5
HiI

F-1­

10 III I _
0 L ... .L'iJ ,I , ,
0 15 30 45 60
DATA WORD VALUE (counts)

Figure 3.7.1-1.- Data-word histogram, band 1.

 - ­

10 -

1
___ - - -: ,4--

---
E 1--..--

....

-+ --

7:I
° =-------

H II F-I

444­

011
0.1 1I0 100

FQUNCY (Hz)
Figure 3.7.1-2.- Power spectral density plot, band 7.

 MSC-05546

TABLE 3.7.1-3.- NOISE-EQUIVALENT SPECTRAL RADIANCES (NESR),

ATTENUATORS IN

_NESR
(mW/cmZ-1m-ster)

SL3

Befoie After

BAND PRELAUNCH SL2 Attenuator Attenuator SL4

Change Change

1 * 0.3 0.23 0.24 *

2 0.2 0.2 0.2 0.2 0.40

3 0.2 0.3 0.2 0.2 0.26

4 0.5 0.6 0.5 0.5 0.66

5 0.6 0.8 0.72 0.146 0.58

6 0.3 0.3 0.3 0.3 0.27

7 0.25 0.3 0.3 0.35 0.27

8 0.3 * 0.22 0.21 0.27

9 0.08 * 0.12 0.10 0.09

10 0.14 * 0.17 0.15 0.15'

ii 0.05 * 0.06 0.06 0.06

12 0.03 * 0.025 0,025 0.027

• Insufficient data

1-11-37

 MSC-05546

TABLE 3.7.1-4.- NOISE-EQUIVALENT SPECTRAL RADIANCES (NESR),

ATTENUATORS OUT

NESR (mW/cm 2 -jm-ster)

BAND Prelaunch SL2

1 0.3 *

2 0.2 0.2

3 0.2 0.3

4 0.4 *

5 0.4 0.6
6 0.2 0.4

7 0.2 0.4

8 0.2 0.2

9 0.1 0.1

10 0.1 0.2

11 0.05 0.07

12 0.03 0.03

* Insufficient data

The iiagery produced from flight data had two types of structured noise.
One was low frequency and appeared as banding or streaking in the direction of

scan. The other was high frequency and appeared as a "herringbone pattern" in
the image. Sections 9.2, 9.3, and 9.4 of MSC-05528, Volume III, September 6,

1974 include descriptions of these noises and how they were suppressed.

3.8 Pointing Accuracy and Field-of-View Determination

A valuable aspect of the Skylab earth resources multispectral scanner

imagery was its ability to record the left, center, and right geographic co­
ordinates of a scan line. These data were based on spacecraft attitude and po­
sition telemetry and the geometry of the S192 system. To determine the accu­
racy of these coordinates, an evaluation wag performed to determine the point­
ing direction of S192 relative to S190A and the error between the actual

coordinates observed on the imagery and those predicted from attitude data.

111-38

 MSC-05_46

TABLE 3.7.2-1.- THERMAL-BAND NOISE

STANDARD DEVIATION NOISE -EQUIVALENT

TEMEERATURE DIFFERENCE
(counts) (OK)
Band Band Band Band

13-1 13-2 13-1 13-2

SL2 13.0 18.5 4.3 4.3

SL3 7.6 10.8 2.3 2.3

SL4 (Y-3) 7.0 9.7 2.2 2.2

SL4 (X-5) 3.0 4.2 0.9 0.9

3.8.1 Pointing Accuracy

The pointing accuracy of S192 was evaluated for each of the three Skylab
missions. Using the technique described in Section VII of Appendix A, the S192
data were registered with respect to latitude and longitude determined from
identifiable geologic features in the data. These coordinates were compared
with those resulting from spacecraft.attitude telemetry. It was found that the
difference varied with. time, indicating that the source of the error originated
in the drift of the rate gyros.

For the times analyzed, Table 3.8.1-1 shows the maximum differences at the
center of the scan.

TABL 3.8.1-1. - MAXIMUM LATITUDE AND LONGITUDE DIFFERENCE

BETWEEN S192 DATA AND POINTING ESTIMATES DETERMINED

FROM SKYLAB ATTITUDE TELEMETRY

MAXIMUM DIFFERENCES
Latitude Longitude Latitude
Longitude

min s min a
SL2, Pass 7 39020N 87050'W 5 21.9 3 26.5

SL3, Pass 52 300 10'N 900 301W 2 5.7 0 50.6

SL4, Pass 90 40'50'N 890W 4 13.8 7 24.5

111-39

 MSC-05546

3.8.2 Field-of-View Determination

Determination of the actual S192 Field of view was performed by locating

geologic features taken from S192 imagery on Geologic Survey maps. The nominal
S192 swath width (i.e., exactly in Z-LV attitude and at 235 n.mi. altitude)
is 38.61 n.mi. The results of the flight data determination showed S192 swath
width to be 39.6 ± 0.5 n.mi.

3.9 Photographic Image Adequacy for Site Location

6192 screening film served to locate sites so particular digital data or

high-quality film can be specified and ordered. Evaluation of this film is de­
tailed in Section 11 of MSC-05528, Volume III. No prelaunch screening film was

made from which image adequacy could be evaluated.

3.9.1 Image Adequacy - SL2

The standard 5-inch screening film *as not made for SL2 because the

screening film equipment was not operational in time. As a substitute, 35­

millimeter SL2 screening film was made with a temporary facility. Screening
film for bands 2, 7, and 11 was compared with S192 data analysis station (DAS)
imagery and with "Printed gray maps."

The primary image-quality factor noted in the screening film was the lack

of contrast. High-contrast objects were easily detected, but objects that pro­
duced small changes in contrast were not detectable. Objects with sufficient

contrast had resolutions comparable to those obtained from an area with about

3.5 S192 pixels on a side combined into one pixel. The granularity of the
screening film was also a significant source of degradation, making scan-line
counting difficult in some areas and impossible in most.

3.9.2 Image Adequacy - SL3

S192 screening film for SL3 was 5-inch film made by the JSC production
film converter. The film was considerably superior in resolution, contrast,
time-mark location accuracy, and convenience. Screening film for bands 2, 7
and 11 were examined.

The most serious limitation of the imagery was the quantization of the
levels by reducing the number of 256 counts to 64 density levels. Low-contrast
information below a four-count difference was suppressed. This caused an
absence of fine shades of gray and an uniform appearance on large water and
desert areas.

The resolution along the scan direction was obviously less than the reso­
lution perpendicular to it. The distance for a detectable change was a single
pixel perpendicular to the scan direction and 1 to 2-pixels in the scan

direction.

111-40

 MSC-05546

When mddium- or high-contras-t targets were usedb'the-seteening 'film was

excellent for locating targets and spdcifying the seari-line number (in teims of

time) and the -pixel-number along the scan line-. ­

3.9.3 -Image Ade - SL

There was little difference between the SL3 and SL4 screening film. The-

X-S detector array was instaifed in place of the Y-3 array-on January 16 i1974

and was used for the rst of the mission. The screening film froni the X-5 de­
'tector appeared somewhat better to the eye but -quantifying this improvement

-as impractical. Z
­

- 3.10 Geoetrical Band-toBand Registiation Error ­

Band-to-band data registration"war studied by analysis of scan lines and

pixel trades crossing the edges of: the,moon. The single field stop design of

the insttument -makes It extremely unlikely that there cdild -be detectable his­
r~ggstration prpendithlar to'-the scan lines. Comparison of corresponding­
ffx6l- traces fbr the gevdraI.'bands did not- isclose any- such misregistratibn.

-Along:.the 'scan lines; howeve,, iisregistration-c -ould-occur deto lags i

the analog eledtionics or'drda displacements in th digital electronics either,
oi-board or during ground.pro-ssing. t'oip-ari on of averaged scan lines cross­
._ing the moon showed -that -any atong-lne Misregistratins- were small. However
-

analysis of -the-tabulated 'data us&d -in carrying gnt the'.lunat -chlibrations- ­

showed some minot effects. :To eqiore-thes-e more-thorou'ghly five -or sit-pixels

covering the leading and trailing edges of-the moon were selected and the

averaged dat for these pixels in each band in-the LU tabulations were normal­
ized by dividing by the peak signal recorded. An example of the results is

shown--in TabIe -3.10-1. ­ - -. ­

.
-Generally only odd SDO data were used from the double-sampled" bands. How­
ever,- due,to poor CT.data in the odd SDO, whenever evDn-SbO data'-wd -eused the
signals in the even SDO's appear to lead the odd SDO's by abbut-1/2 pik'el. :- "
This is -.to be expected, as in these high-data-rate bailds alternate samples- go
to odd and even SDO!s.-- In each LC SDO 20 appears in registrafion with the"odd
high-rate data while in LOs 2 .through 5 SD0"S 17 and 19 lag the odd SDO's by
- about 1/2 pixel and are-thug I pixettbehind the even high-rate SDO's. However,.
in LC 1 SDO's 17 and 19 appear to be in registration with the odd high-rate
SDO's while SDO's 18 and 20 lag 1/2 pixel to become about 1 pixel behind the
even high-rate SDO's. Finally, for Las-4 and 5, SDO 1 leads the odd high-rate
.SDO's by 1/2 to 1 pixel so that it is then more than 1 pixel ahead of SDO's 17

and 19.

The results given above were based on the rise at the leading edge of the

moon. As may be seen from the table, the trailing edge data is less defini­
tive due to frequency response effects which vary from band to band.

111-41

 MSC-05546
TABLE 3.10-1.- EXTRACT FROM LCl DATA (NORMALIZED)

Pixel Number
SDO Band 657 658 659 660 661 662 712 713 q14 715 716 717

1 3 0.02 0.02 0.05 0.13 0.24 - 0.46 0.35 0.18 0.05 0.01 -

2 3
3 4 0.04 0.04 0.05 0.14 0.29 - 0.45 0.34 0.16 0.02 0.01 -

4 4
5 5 0.04 0.04 0.04 0.06 0.12 - 0.47 0.37 0.19 0.07 0.04 0.05
6 5
7 6 0.02 0.02 0.04 0.12 0.23 - 0.47 0.36 0.20 0.06 0.03 0.03
a 6
9 7 0.04 0,03 0.06 0.14 0.25 - 0.49 0.39 ,0.22 0.08 0.06 -

10 7
11 11 0.02 0.02 0.05 0.14 0.25 0.34 0.45 0.33 0.16 0.05 0.02 -

12 11
13 12
14 12 0.04 0.04 0.08 0.17 0.26 0.34 0.44 0.31 0.17 0.09 0.07 -

17 10 0.03 0.03 0.05 0.13 0.25 0.35 0.47 0.36 0.19 0.05 0.03 ­

18 2 0.02 0.02 0.03 0.09 0.18 0.26 0.51 0.43 0.30 0.18 0.13 ­

19 8 0.02 0.02 0.03 0.10 0.21 0.32 0.51 0.41 0".24 0.09 0.05 ­

20 9 0.02 0.02 0.03 0.09 0.19 0.30 0.50 0.42 0.27 0.13 0.06 ­

22 1 0.03 0.03 0.04 0.07 0.11 0.16 0.87 0.83 0.75 0,68 0.64 0.62

Note: The thermal band (SDOs 15, 16 and 21) was not considered, as it cannot be directly
compared with the other bands.

The thermal band was not considered, as the lunar surface temperature
falls off towards the edge of the nearly full moon so that thermal and reflect­
ed radiances cannot be compared in this way. Band 1, SDO 22 was not considered
in Ls 1, 2,' and 3 due to its very poor frequency response. However in LCs 4
and 5, when the frequency response of this band had been improved (i.e., during
SL4), SDO 22 appeared to behave like an odd high-rate SDO when allowance was
made for the high residual effect-it contained.

Due to the large frequency responge defects in the X-5 array bands 1
through 12 data, misregisttation should be great. When the filtering technique
is applied this problem should be greatly alleviated.

Ii-42

 MSC-05546

4.0 SUPPLEMENTARY ANALYSES

This section presents the results of analyses completed after the submit­
tal of the sensor performance evaluation interim report, NSC-05528, Volume IIIj

September 6, 1974. Included in the section are an analysis of the S192 scan­
related system response and a comparison of S190A, S191, and S192 radiometric

performance.

4.1 Supplementary Analysis of Scan-Related System Response

Scan-related system response for S192 is the scanner response over a tar­
get of uniform radiance as a function of scan position. As implied, determina­
tion of this relationship required a data set representing scanner response to

a uniform scene. The response versus scan angle functions used to reduce the

S192-data were derived by averaging and normalizing the S192 response versus

scan from several homogeneous scenes since uniform single scenes of the size

required do not exist in nature. This technique is described in Section IV of

Appendix A. The following paragraphs described an alternate technique devel­

oped by HIfC for determining the S192 scan angle response from a single scene

that is the most uniform available but is not in actuality uniform.

For each scene selected for analysis of the scan function, S192 data
across the scan were tabulated. Data points representing the mean of 3000 sam­
ples (10 pixels times 300 scan lines) were selected for 12 points across the

scan. Figure 4.1-1 illustrates a sixth-order polynomial fit to these points as

the curve marked "scene data." The scan-related signal offset was determined
using 12 data points of the same sample size from deep space data taken during
the lunar calibration closest in time to the scene data. The true mean of the

offset data points was determined using the technique described in Section III

of Appendix A. The offset curve in Figure 4.1-1 illustrates the offset as a

.function of scan angle. This offset was subtracted from the scene data to ob­
tain the scene data relative to zero counts. A sixth-order polynomial was fit­
ted to the data and a normalization made to the value observed at a point 600
from the start of the scan. This point is near to the center of the S192 scan
line. The normalized curves were then compared with the prelaunch scan re­
sponse curves. An example, Figure 4.1-2, shows the -band 3 prelaunch curve

compared with the average curve for the three scenes of the Mali Desert.

The accuracy of scan response curves determined from the scene data de­
pends on the uniformity of the averaged scenes or at least on their being sta­
tistically uniform when a large sample is taken. To assess the accuracy of the

results, it was desirable to have an independent indication of how uniform the

scenes were. This was accomplished by using S190A photographs.

The three Mali Desert scenes selected for S192 scan-related system re­
sponse evaluation were located in the corresponding S190A frames. The black­
and-white film camera stations were used (station 1, black-and-white infrared,
0.7 to 0.8 micrometers; station 2, black-and-white infrared, 0.8 to 0.9 microm­
eters; station 5, black-and-white Panatomic-X, 0.6 to 0.7 micrometers; and

station6 black-and-white Panatomic-X, 0.5 to 0.6 micrometers) to approximate

a few of the spectral bands of the S192 scanner.

111-43

 MSC-05546

200

18D -. /-SCENE DATA /,CEMROF SCAN

,I,
160

140

120

o 00 SCENE DATA
MINUS OFFSET

80

60

40

20

0
0 10 20 30 40 50 60 70 80 90 100 120 120

SCANANGIE ()

Figure 4.1-1.- Scene data minus calculated real offset.

1.2­

1.0

SI ,. I , ,
10 30 50 70 90 no
bCAR MMCE (')

Figure 4,1-2.- Cqnpdrisob 6f 80 scene (Sahara Desert) with

prluth datd fli batid 3.

11-44

 MSC-05546

A template corresponding to 300 8192 scan lines was overlaid on the SI90A

frames, covering each of three selected scenes. Six film density measurements

were taken at each of nine scan locations 150 apart across the field of view of

the S192, as illustratel in Figure 4.1-3. These measurements were converted to

original flight-film values using SL3 sensitometric data for s190A. The six

corrected values for each 150 location were then averaged to obtain an average

density for each of the nine scan angles.

A spectral distribution for each type of scene was needed to provide com­
parative calibration measurements to define spectral limits and band shape.
Because ground truth was not available for these particular sites, previously
determined spectral distribution values calculated by Martin Marietta Corpora­
tion in 1973 for the Great Salt Lake Desert were substituted for the Sahara
Desert sites.

A density from the film, assumed scene spectral distribution, and S9G0A

system transmittance values based on preflight test data were then used in a

Martin Marietta Corporation computer program. This computed effective target

radiance above the atmosphere as predicted by the S190A. These data were

plotted as a function of position on the S192 scan and a sixth-order polynomial

fit of the data was made. Figure 4.1-4 shows the polynomial normalized to the
midpoint on the scan line for the Sahara Desert scene derived from Si90A sta­
tion 6.

The results from S190A indicated that the radiance of the Sahara scene was
not uniform. Radiance over the last 600 of the scene increased by approxi­
,mately 7% relative to the center.

This result was used to correct the S192 scan function by accounting for

lack of scene uniformity. Figure 4.1-5 compares the prelaunch scan response

with SL3 data from Figures 4.1-2 corrected for scene variation indicated by

Figure 4.1-4.

Use of this technique demonstrated that, even without a large uniform­

radiance scene, scene nonuniformity could be approximated and scan function

determined in those S192 bands for which a Sl90A provided corresponding spec­
tral bands. Implications as to the scene nonuniformity might then be drawn for

other S192 bands.

11-45

 MSC-05546

AVERAGE OF "
300 SCAN LINES
10 WORDS/LINE

IAFPROX
ONE
SUB-

S192 SCAN 300

SCAN
-10 WORDS ILINES

DENSITY
MEASUREMENS
S190A FRAME
AS SEEN THROUGH
S192 TEMPLATE

NOT TO SCALE

Figure 4.1-3a- Compdridft of S192 dita ptbduet and 91g0A densitometry data.

Itt-46

 MSC-05546

S1.1­

S0.5­

0-

I I I I I II II
10 30 50 70 90 110
SCAN ANLE (e)

Figure 4.1-4.- S190A computed scene radiance variation for Sahara Desert.

1.1­

.­
0.
:,i

I I I I I I
10 30 30 70 90 110
SCAN ANGLE (0)

Figure 4.1-5.- Comparison of 9192 data tot Sdhara Desert corrected for S190A

Scene radiance vatiation with prelaunch data for band 3,

Itt-47
 M1c-05546

4.2 Supplementary Analysis of System Noise

System noise requirements of the EIS (End Item Specification) for 8192

Multispectral Scanner were stated in terms of NEAe (noise equivalent reflec­

tance increment) and NEAT (noise equivalent temperature increment). These

measures of noise level tie the system performance to the viewed scene charac­
teristic of principal impdrtance and, therefore, require certain viewing condi­
tions to be defined. The scanner is always assumed to have its optical axis

aligned with the local vertical axis, and the earth's surface is assumed per­
pendicular to it, so only the illumination conditions, and reflective properties
of the scenes for bands 1 through 12 and the temperature and emissivity of the

earth's surface are required to be specified. Those conditions used in the EIS

were selected to be representative of average or "normal" conditions for an

earth scene and are given in Table 4.2-1.

Prelaunch, both at Honeywell Radiation Center during acceptance tests and

at Kennedy Space Center during integration and checkout testing, and postlaunch
before the scheduled sensor performance evaluation could be performed an eval­
uation of the S192 noise characteristics was required. To enable direct com­
parison with the EIS values these noise values were generated in the form of
NEAe and NEAT. The prelaunch data were obtained with the scanner viewing the
uniform illuminated test target while inflight values through SL4 pass 83 (Y-3
array) were obtained from data sets taken over clear, calm (no foam) ocean free
from glitter. The SL4 values after pass 83 (X-5 array) were derived from deep
space a's for bands 1 through 12 and clear ocean a's for band 13 generated by
ERIM in the course of developing calibration data. A tabulation of noise values
for each mission is shown in Table 4.2-1 for information and as a formal record
since these values were given substantial circulation -- mostly internal to
NASA.

These results for the Y-3 array show that the noise levels increased
inflight over the prelaunch values. The noise performance generally improved
as the Skylab flights proceeded due to efforts to eliminate the causes of the
increased levels. Bands 4, 5, 12, and 13 exceeded the EIS noise levels, but
this condition was known to exist prelaunch for 4, 5, and 12 and values for
13 were not obtained prelaunch due to the need for performing band 13 testing
in a vacuum chamber.

The X-5 array noise values in the reflective bands are higher than those
for the Y-3 array in all bands except in band 4 in which case it still exceeds
the EIS value. The X-5 array band 13 noise value is significantly better than
for Y-3 array and is better than the EIS requirements. These results were also
anticipated (even though prelaunch system tests with the X-5 array were im­
possible) since the dewar containing the reflective bands array was opened to
replace the thermal detector. The reflective array then had to be put through
a second outgassing take-out and schedule did not permit tuning the band 1
through 12 preamps to the changed detectors.

111-48

 MSC-05546

4.3 S19OA, S191, and S192 Radiometric Comparison

•Each BREP optical sensor (SI90A, Sl91, and S192) was designed and cali­
brated to provide absolute spectroradiometric data. These sensors also covered

Scomnon wavelength regions, which facilitated a radiometric comparison. How­

ever, the spectral bands and band-widths were different and required band

averaging to accomplish the comparison,. The spectral bands for each sensor

are given in Table 4.3-1. The radiometric values output from the sensors were

converted to comionmunits. 'The S190A output data were converted to unitsof

spectral radiance (mW/cm2-

-pm-ster) by dividing the S190A total radiance output

(mw/cm2-ster) by the bandwidth of each station. The bandwidth equaled the

difference between the limits of integration used' to calculate the Sl90A radi­
*ance output from'eqihtion A.II.I2, Appendix A, Section II of Volume I.

The- radiometric comparison as planned was to have compared radiance values

when llthree sensors were observing the same target. However, due to mis.
sion scheduling difficulties only one ground truth site suitable for radio­
metric evaluation was observed simulataneously by all three sensors. This site,
was the Willcox Playa observed during SL2 when there-were sniall cumulus clouds

near the site" And whibh probably influenced the results., However, aomon

sites wire avAilabl for e6fiparing 190A to- S191, S190A to S192, and A limited

cohparison of S191 to S192.. Based oh data" from these sites the overall tom­
* parisoa- of: all -'Ehr'e. sensors 'puld 'be, made: 'v ..

*" 443.1 Comparison of S190A to Sgl1 ' -

Comparison of S190k-tb S191 was bAsed 5in the three 9 te:­

S-Misioi: Date" -. Site'

. SL2 . .,6/3173" - . Willtox Playa, Ariiona.,

S83 8 /,- Katherine Playa, New Mexico
S4 - - 2/-774 * Kat.herine ,Flaya, New Mexico ­

"The spectral +adiande of each site was first calculated from the S191 data for

•13 narrow wavelength bands-over the spectral range of the S190A (0:44 to'0.9

.1Vm),. These data were calculated using S191,responsivity derived from ground­
"based .lunar mare measurements made,with the S191 backup spectrometer. This

-risponsivity is given in Volume II, Figure 4.1-1, channel A-5. The resulting

,spectral rddiance values for the three sites are listed in Table 4.3.1-1 and

plotted in Figure 4.3.11'. These data were theti averaged over each of the

S190A spectral bands to obtain the average ,spectraliradiance comparable to each

£S190A station. The resulting average spectral radianee for both sensors and

that derived from'the groundl truth measuremeht are listed'!i Table 4;3l1-2.

'Eddiance ratios ward 'also'calculated and'iisted to prbviae a basis for inter-'

sensor ,comparison.

111-49

 TABLE 4.2-1.- S192 MULTISPECTRAL SCANNER SENSITIVITY (NOISE.)*

Band- Spectral Assumed Atmosphere Prelaunch Y-3 Array X-5 Array

Number Interval (urn) Transmission, % EIS Y-3 Array SL-2 SL-3 SL-4 SL-4
T O.41 ta 1..4a 45 2.0 N/A** 1.4 1.3 1.1 2.2
7 (0.4 te Q.57 45 2.0 0.9 1.0 1.1 0.8 2.5
a Q.Z ta O.5l 55 2.0 1.0 1.5 1.2 1.0 1.6
It (5E ta 0.61 55 2.0 2.4 3.T 2.8 2.5 2.1
5 U-6- tW .&T 60 2.0 3.0 4.2. 2.5 2.5 ---

S 0.6tao'.7m 65 2.0 1.3 1.6 1.5 1,3 2.6

7 'aaTaMt: (.SE 70 2.0 1.4- 1.8 1.7 1.5 3.0
& Cl-99 to: IrO&8 8L 2.0 1.3 1.5 1.5 1.6 3.0
9 T. xta
I-I 5; 2.0 0.9 0.9 1.2 0.9 2.7
To 1.f2a ta T.3 80 2.0 1.8 1.9 1.T 1.7 ---
T1 TSE tat.f 70 Z.0 1.5 1.6 1,8 2.0 2.6
IZ Z.It ta Z.3E 70 2.0 .3 Z.4 1.5 2.2- 2.7
0K 2.60 K 2.20 K
T ICU. ta T215 90 1.5 N/A*** 4OK O.80 K

*Bands T thra 1Z are NAp (%earth reflectance) white Band 13 is NEAT (OK), see note below.
**Band 1 attenuation removed after ground testing completed.
***Ground test of thermal band impossible outside vacuum chamber.
NOTE: NEhp's and NEAT's were computed using the atmospheric transmissions in the table,
solar irradiance = 140 mw/cm 2 (outside the atmosphere), scene reflectance = 20% (diffuse),
angle of incidence of solar radiation 450, and (for NEAT) the earth a 300 0K blackbody.
REPRODUCIBILITY OF THE
ORIGINAL PAGE IS POOR MSC-65546

TABLE 4.3-1.- S190A, S191, AND .S192 SPECTRAL BANDS FOR HADIOMETRIC COMPARISON I

S190A S192 S191*

an Band Wavelength Wavelength Wavelength
Station Wa(em)lea BandWav t Range (om) Resolution (pm)

1 0.41 to 0.45 1 0.39 to 0.73 0.0115

6 0.48 to 0.63 2 0.45 to 0.51

3 0.50 to 0.56

5 0.58 to 0.72 4 6.54 to 0.60

5 0.60 to 0.66

1 0.68 to 0.78 6 0.65 to 0.74 2 0.68 to 1.43 0.0185

2 0.75 to 0.90 7 0.77 to 0.89

8 0.93 to 1.05

9 1.03 to 1.19
10 1.15 to 1.28

11 1.55 to 1.73 3 1.34 to 2.50 0.015 X A

12 2.10 to 2.34

4 5.82 to 11.40 0.019 X A

-13 10.07 to 12.68 5 8.30 to 15.99 0.019 X A

* S191 had a continuously variable filter; definable narrow bands are given by the
wavelength resolution.

This comparison shows that the spectral radiance values derived from S191

were consistently higher that those from S190A, with an average about 18% high­
er. The large ratios for both S190A and S191 with ground truth data for

Willcox Playa indicate that local atmospheric conditions caused the S191

ground truth calculations to give erroneously 16w spectral radiance values,

particularly in the near-infrared bands. This result supports the suspicions

of that data.

The camera station operating in the visible spectral region showed better

agreement with S191 data than did the infrared-sensitive stations.

4.3.2 Radiometric Comparison of S90A to S192

The comparison of Sl90A and S192 absolute radiimetric measurements was

based on four ground sites:

Mission Date Site

SL2 6/3/73 Willcox Playa, Arizona

SL3 9/2/73 Sahara Deserts Africa

SL3 9/13/73 Great Salt Lake Desert, Utah

SL3 9/17/73 Gulf of Mexico

II1-51

 Msc-05546

TABLE 4.3.1-1.- S1g SPECTRAL RADIANCE FOR S190A AND s191 COMPARISON SITES

SPECTRAL RADIANCE VALUE (mW/cz2-pm-ster)

Wavelength

(on) Willcox Playa, 8L2 Katherine Playa, SL3 Katherine Playas, SL4

08.448 17.63 12.05 8.542

0.475 18.64 12.50 8.975

0.500 17.09 11.45 8.853

0.552 15.40 10.1 8.828

0.600 15.04 10.0 9.267

0.657 15.78 10.9 10.286

0.675 15.65 10.5 10.199

0.700 14.61 9.21 9.866

0.725 15.58 10.04 10.468

0.741 15.29 10.65 10.474

0.800 14.46 9.86 9.920

0.850 12.42 8.75 8.454

0.901 9.28 5.97 6;418

No ground truth measurements were made at the Sahara Desert and the Gulf of
Mexico sites. Unlike the S191 spectrometer, neither the S190A nor S192 had
sufficiently narrow bands to define the detailed spectral distribution of the
ground sites. Also, the spectral bands and response of these two systems were
different. However, it was possible to obtain meaningful radiometric compari­
son data by calculating the average spectral radiance for each Sl90A station
and 9192 band; then computing the average of the S192 bands covering the spec­
tral range o each Sl90A station. Specifically, S190A station 6 was comparable
to the average of S192 bands 3 and 4j statioff 5 wad comparable to the average
of bands 5 and 6; station 1 was comparable to the average of bands 6 and 7; and
station 2 was tontparable to S192 band 1. The spectral ranges of these stations
and bands are given in Table 4.J-i.

The average opesttl tadiatice valdo and radiance ratios for the Sl90A,
S192 and the gtolid truth tdAms tement et ELlen in Table 4.3.2-1. The' S192
Willcdx Playa data ere tecorded oft pda1S J of SL2 befote installation of the
attenaattts. The "off gcale" ligtdd if the table means the output signal was
above the uppet i1mit fui bands 4 atid J, Thig condition was later Corrected
by instailitili t the fttttl atb . ResOiW b* this table ahotr simlar error
tagadtude i tie Wiiitko Playa gtound kiuth data as iA the compatison of SlY0A
to 9191a The data An8d itdiedtd good &~tdtddnt between 190A and S192 radiance
values, with ho apatetlt bystamdtic errors ot bas in the cdtiparson,

itt-52

 MSC-05546

20

U ­
18~ ~ ~ \nMO UPC>P
'AA, 51.2

14

12 - ~ n I ~ l FL YA, 513

--- .- ­
- ---

a AIUEPLA, 51.4

0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90
WAVELENGTh i()

Figure 4.3.1-1.- S191 spectral radiance plot for S190A and

S191 comparison sites.

TABLE 4.3.1-2.- S190A, S191, AND GROUD-TRUTH RADIONETRIC COMPARISON

sPETRAL RADIANCE (twcm-i -ster) RADIANC RATIO WAVELENGTH

8190A/ S191/ BAND
SITE DATE S190A S191 Ground Truth S190A/S191 Ground Truth Ground Truth Um)
HISSION

Willcox 6/3/73 11.93 16.93 10.27 0.70 1.16 1.65 0.48 - 0.63

SL2 0.58 - 0.72

Playa 14.79 16.28 10.00 0.91 1.48 1.63
11.21 14.95 7.16 0.75
1.57 2.09 0.68 - 0.87

10.93 13.93 6.07 0.78 1.80 2.29

0.75 - 0.90

8/11/73 9.47 11.27 12.00 0.84 0.79 0.94 0.48 - 0.63

SL3 Katherine
10.79 10.86 11.57 0.99 0.93 0.94 - 0.58 - 0.72
Playa
6.89 10.16 8.95 0.68
0.77 1.14 0.68 - 0.87

6.40 9.60 8.20 0.67 0.78 1.17 0.75 - 0.90

Katherine 2/1/74 9.27 9.67 9.93 0.96 0.93

0.97 0.48 - 0.63

8.4
Playa 9.07 10.57 10.14
0.86 b.89 1.04 0.58 - 0.72
No Data 10.21 8.05 -- -- 1.27 0.68 - 0.87

8.00 9.53 7.33 0.84 1.09 1.30 0.75 - 0.90

.I 53 REpRODUTCILITY OF THE
,B1GwALPAGE IS pOOR
 MSC-05546

TABLE 4.32-i.- 9190A, S192, AND GROUND-TRUTH RADIOMETRIC dOMPARISON

AVERAGESPECRAL RADIAN RATIAME RATIO GowaAATt OM0ELT

-"S190A/ $1921 S19DA S192

MISSION SITM DUE S19OA S192 Ground Truth $190A/S192 Ground Truth Ground Truth (StatLo) (BA)

SL2 Willeox 613/73 11.99 off scale 10.27 -- 1.16 -- 6 3,4 average
Playa 14.79 oft scale lo.oo -- 148 -- 5 5,6 average
11.21 1i.54 7,16 0.97 1.57 1.57 1 6,7 average
10.93 11.32 6.07 0.97 1.8u 1.86 2 7

SL Sahara 9/2/73 7.75 8.10 -- 0.96 6 3,4 average

scrI 12.25 10.68 -- 1.15 5 5,6 average
- -----
No Data Avallable --- --- No Data Avalable--------- 1 6,7 average

7,92 9.16 -- 0.95 2 7

SL3 Great 9/13173 13.27 .39 i.ou 1.17 1.21 1.04 6 3,4 average

Salt Lake 13.21 11.12 10.43 1.19 1,27 1.07 5 5,6 average

Desert 9.00 9.78 7.79 0.92 1.16 1.z6 1 6,7 average

8.73 9.17 6.93 0.95 1.26 1.32 2 7

$3 Gulf of 9/17/73 4.15 3.43 -- 1.21 .. 6 3,4 average

ex Lco 2.06 1.62 -- 1.27 .... 5,6 average

4.3.3 Radiometric Comparison of S191 to S192

Only two ground sites suitable for tadiometric comparison of 5191 to S192
were available. They were the SL2 Willcox Playa site discussed in paragraphs
4.3.2 and 4.3.2 and the Rio Grande Reservoir, Colorado, site observed on 8/8/73
during the SL3 mission. The absolute spectral radiance values based on S191
data for Willcox Playa are listed in Table 4.3.1-1 and plotted in Figure
4.3.1-1. The S191 spectral radiance for the Rio Gtanie Reservoir is given in
Table 4.3.3-1 and plotted'in Figure 4.3.3-1.

The comparison of S191 to S192 was achieved by taking the average of the

S191 spectral radiance over each corresponding S192 band. Infrared wavelengths

greater than 0.901 pm were tot considered due to A lack of analysis time. The

resulting comparative data are listed in Table 4.343-2.

A review of this table for Rio Grande Reservoir shows the S191-derived

spectral radiance is higher than that for S192 by approximately 16%. The S192

results for Willcox Playa differ significantly from those of S191. No attenu­
ators had been installed in S192 and other problems were associated with these

data. The differences with Willcox Playa ground truth ate again apparent.

4.3.4 Radiometric Comparison Summary

The derived spectral radiances from various targets agree closely for

SIOA and S192. The S191 values were about 16 to 18% higher than those for

S190A and S192. The values are given in Tables 4.3.1-2, 4.3.2-1, and 4.3.3-2a

The-agreement among the three sensors was judged to be excellent considering

that a recent study* showed the variation in radlometric calibrations made at

*Franc Grum and Joseph Cameroni "Detector Intercomparison Results,"

Electro-Optical Systems Design, Vol 6, November 1974, p 82.

111-54

REpRODUOiB1LITY OF THE
ORIGINAL PAGE IS POOR
 IviSq-05546

TABLE 4.3.3-1.- S191 SPECTRAL RADIANCE OF.RiO GRANDE

RESERVOIR FOR CObWAEISON OF 6191 TO S192

Wavelength (4m) Spectrai Radiance (mW/cm2 -m-stei)

0.448 4.94
0.475 4.48
0.50 3.69
0.552 2.74
0.60 1.72
0.657 1.42
0.675 1.37
0.70 1.41
0.725 1.42
0.741 1.46
0.80 1.37
0.85 1.12
0.901 0.80

nV-

3­

2-

S ' ' I I 1 I
1 I I i I I I

0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90

WAVELENGTH (11m)

Figure 4.3.3-1.- S191 Spectral tadianct of Rio Grande reservoir, SL3.

111-55

 MSC-05546

TABLE 4.3.3-2.- S191, S192, AND GROUND-TUTH RADIOM{ETRIC COMPARISON

AVE RE SfPEAL RnAACE RADIANCE RATIO
(m~lcm-sm-ster)
91
1914
" s192/ 5192 SPECTRAL

MISSION SITE DATE 191 192 Ground Truth S192/5191 Ground Truth Groud Truth BAND

SL2 Willcox 6/3/73 17.8 9.51 10.8 0.534 1.648 0.881

2

Plays

16.1 10.13 9.4 0.629 1.713 1.078

3

15.3" off acale 9.4 --- 1.628 --- 4

15.4 off scale 9.7 --- 1.588 --- 5

15.3 11.54 8.6 0.754 1.779 1.342 6

13.0 11.32 5.6 0.870 2.321 2.021 7

SL3 Rio Grande 8/8/73 4.43 3.49 Not available 0.788 ...... 2

Reservoir
3.20 2.70 - 0.844 .-- -.- 3

2.92 1.38 0.815 --- , 4

1.51 1.60 1.060 -.--.. 5

1.36 1.05 0.772 .... 6

1.17 1.03 Not Available 0.880 . 7

various standards laboratories to be approximately +10% (total variation 20%)

from the consensus. Considering the space environment, the variations in

spectral bands, and the less-than-optimulnm calibration procedures available

before launch, the relative radiometric absolute accuracy of these three
instruments was considered excellent.

These data also showed that, bas&d on these 5190A and S192 results it
appears that the lunar radiance values calculated using the Lane and Irvine

data* were approximately 25% high.

ORIGNAL PAGe P0

- *A.- P. Lane and W. MIttingj "Mofiochrutlid Phage Curves and Albedos for

the tuilat Distk" 'he Astoomicai Journal, Voi 78, No. 30 1972.
 MSC-05546

5.0 CONCLUSIONS

The sensor functioned very well indeed. The refrigerator which cooled the

detectors and the dwar which contained the detectors were considered prelaunch

as the-most likely source of functional problems and none materialized.

The geometric performance was excellent since the data proved to be read­
ily registerable to maps and to LANDSAT A data.

In 9 of the 13 bands radiometric performance was very good. In most in­

stances where the performance did not meet the EIS requirements the deficien­
cies were known to exist prelaunch and could have been remedied had the time

and equipment been available. The principal surprise was the poor quality of

the band 13 (thermal) data from the Y-3 array. This arose from the inability

to complete the planned prelaunch testing in a vacuum chamber. That this could

be overcome was demonstrated by the excellent quality of the band 13 X-5 array

data.

Band 1 data was generally unusable for most purposes due to its poor fre­
quency response in SL2 and SL3 and its calibration changed significantly in

SL4. Bands 4 and 5 had low signal-to-noise ratio which tends to make tests of

the utility of the additional bands which include comparison with LANDSAT A

inconclusive.

These problems notwithstanding, the S192 Multispectral Scanner produced

highly useful data in more spectral bands and in narrower spectral bands than

any set of scanner data yet available for Earth Resources Survey uses.

6.0 RECOMMENDATIONS

As described throughout this report and MSC-05528, there were a number of

areas in which the design and use of S192 Multispectral Scanner could have been
improved upon. In most instances the deficiencies were known prelaunch and
could have been cured given the opportunity. Since S192 will never be flown
again there will be no attempt here to detail .all the problems and recommend
detailed solutions which are irrelevant. The recommendations below are those
which are applicable to the whole genre of multispectral scanners and to image
plane scanners in particular.

6.1 For high performance image plane scanners, provide a mechanism for in­
flight determination of response versus scan angle variation. This could be a

uniform test target (which can be illuminated or left dark on command) included
in an enlarged objective aperture cover (this adds complexity, size, weight,

and power consumption to the primary sensor). It could also be a separate,

small object plane scanner with the same spectral characteristics but much low­
er spatial resolution (and data rate) to define the scene variation for use as

described in paragraph 4.1. Either of these devices should be well calibrated

and well exercised (and therefore well understood) prelaunch since they are

calibration devices in themselves.

 MSC-05546

6.2 For image plane scanners used in applications requiring accurate cal­
ibration (either relative or absolute) a mechanism for inflight check of cali­
bration, including all telescope mirrors should be provided. The near-full
moon provides an excellent external source for monitoring stability of relative
calibration between bands if operational constraints permit it to be viewed at
appropriate times. The two mechanisms described in paragraph 6.1 would be
suitable for check of either absolute or relative calibration.

6.3 For any scanner used in applications requiring accurate calibration

the internal calibration sources must be uniform and stable in radiance, inset
sufficiently into the dynamic range of the sensor to prevent clipping, and
usable in all testing environments.

6.4 For any scanner± calibration data, data processing algorithms, and

data processing computer programs should be developed in parallel with the sen­
sor hardware so that data from Acceptance Tests, Qualification Tests, and Inte­
gration Testa can be processed and evaluated as the actual flight data will be.

From this it follows that such testing should be done in "as near flight con­
figuration as possible" with the minimum requirement to record the data in same

form and format that the flight data will be recorded.

7.0 NOTES

7.1 Acknowledgements

The effort covered by this report wag sponsored by the Earth Resources

Program Office of the Lyndon B. Johnson Space Center. It is based on the re­
sults of a concerted effort by numerous individuals in:

Environmental keseatch Institute of Michigan (kRIM)

Ann Arbor, Michigan

Science dud Applications Directorate

Lyndon B. Johnson Space Center

Lockheed tlacttonkcg Cbmptust~ Incoporated

Lyndoft 1l. Johnsun Space Ctitit

Mattit Hwtiettd dorpottioh

betvet) dolotadd

Pattidutat aeknowiedgemiet is duo the lhte Mt. Charles K. Williams of the

Skylab Prograi Office bf the Lyndon h. dohndon Space Center. His dedication
&nd 1eAderbllip Ws!iP 1u6tidI td thd Utt@§9kdul boiipitioi of these evaluation
htudked. Alat tsa fiudtild Afid wiiiiAt E, hlendley of dSd/S&Al John G.
Stdithtdite and Pksk talbaot bu VRtkM ati W. Jihnlsgou and Willam M. Miller
b* mCd, and kobp d& MdihotittlM i d dhiElsu d. dtipbli of I/AI have made sub­

OltrB~t~fl
tati~ tOthi o fise:
 HSC-05546

7.2 Abbreviations

Abbreviations in common usage have been used. for English units of measure.
International units (SI) have been abbreviated in accordance with E. A.
Nechtly's NASA SP-7012, The International System of Units, 2nd Revs National
Aeronautics and Space Administration, Washington, D.C., 1973-except for stera­
dian, which has been abbreviated to stet.

AVG Average

C/D/P Cooler/Dewar/Preamplifier Assembly

DRA Data reformatting assembly

EREP Earth Resources Experiment Package

ERIM Environmental Research Institute of Michigan (Formerly Willow

Run Laboratories, University of Michigan)

FOV Field of view

JSC Lyndon B. Johnson Space Center

KSC Kennedy Space Center

td Lunar calibration

tEC/ASD Lockheed Electronics Corporation, Aerospace Systems Division

IMC Martin Marietta Corporation

MSC Manned Spacecraft Center

MTF Modulation transfer function

N/A Not available

NEAT Noise-equivalent temperature difference

NESR Noise-equivalent spectral radiance

NOAA National Oceanic and Atmospheric Administration

OFCE Optical functional checkout equipment

PASS XX Pass numbers used in this document are EREP Skylab consecutive
pass numbers

PRT-5 Barnes precision thermal radiometer, Model PRT-5

PSD Power spectral 'density

111-59

 MSC-05546

RPMI Radiant-power measuring instrument, Bendix Model 100

S&AD Science and Applications Directorate, JSC

SKYBET Skylab Best Estimate of Trajectory Ephemeris Data

SL2 First Manned Skylab Flight

SL3 Second Manned Skylab Flight

SL4 Third Manned Skylab Flight

SPE Sensor performance evaluation

X-5 C/D/P installed between EREP passes 83 and 84 during SL4

Y-3 Primary C/DIP used during SL2 and SL3 and through pass 83 during
SL4
 MSC-05546

APPENDIX A

TECHNIQUES ADDENDUM

This appendix describes in detail the: techniques used to

evaluate S192 performance as presented in the Sensor Performance

Evaluation Report, MSC-05528, Volume III, dated September 6, 1974.

These descriptions of the techniques include both the theoretical

approach and the mechanics of application.

II!-A-I

 MSC-05546

I. SPECTRORADIOMETRIC CALIBRATION OF S192 BASED ON GROUND TRUTH

MEASUREMENTS

Spectroradiometric calibration of the S192 required that the

relationship between output of the instrument (in counts) and a

known input spectral radiance (Ns) be determined. One means of

obtaining this relationship was the use of ground truth measure­

ments made by Martin Marietta concurrently with EREP overpasses.

The detailed results of these ground truth measurements were

reported for each Skylab mission.*

The expression relating N5 to the quantities measured on

the ground is

NS -_ He-T sec 0

where

N = apparent spectral radiance from the target area at

the spacecraft

H = total (direct and diffuse) spectral radiance

incident on the target

p = target reflectivity (as a function of wavelength)

T = atmospheric optical depth

6 = sensor view angle with respect to the normal

N = atmospheric pdth spectral radiance

aA

The methods used to measure H, p, T, and N ate!

ot) Total solar radiance, H, (direct and diffuse) incident

on the target. A spectral scanning spectroradiometer covering

the wavelength range from 400 to 1300 tim was used to measure

the total solar radiation incident on the target.

* MSC-05531 Ground Truth Data for Test Sites (SL2),

August 15, 1973
MSC-05537 Ground Truth Data for Test Sites (SL3),
March 29, 1974
MSd-05543 Ground Truth Data for Test Sites (SL4),
April 30, 1974

III-A-2

 MSC-05546

2) Target reflectivity, p, (as a function of wavelength).

The same specttotadiometer used to measure R was used to

measure the radiance reflected from the target area. The ratio

of these two values'gives the target reflectivity.

3) Atmospheric optical depth, t. This quantity is

calculated by Using measurements of the direct solar tadiance

as functions of the solar incidence angle.

The instrument used was a pytheliometer, which is a

spectral scanning spectroradiometer equipped with a collimator.

In use, it is pointed ditectly at the sun, ahd Produces a

meter reading, M, that is ptoportional to the direct solar

radiance at the surface. Uisilig the expression

S= M e Tseco [ A.I.2]
0

where

*M = the value of M that would be observed by the

o pyrheiiometet if it were above the atmosphere

a
0
=
solar incidence angle (with respect Ea the normal)i

rewriting the equation by taking logs of both sides and

transposing,

T sec 8. = log M° - log H tA.I.3]

By measuring values of M at various values of o simultaneous

equations can be written and solved for T and M . In practice,

the solution technique used is to plot values of M1versus

T sec e (which gives the relative air path length with respect

.to a vertical path) on a semilog plot. The slope'of the line

is -T, and the extrapolated line intercept with the vertical

axis gives the valUe of M . This technique allows a convenient

least squares fit of the data to determine M .

4) Atmospheric path spectral radiance N a This is a

calculated quantity derived from an atmospheric radiative

transfer computer model**.

** W. A. Malila, et.al: Studies of Spectral Discrimination,

Report No. NAS CR-WRL 31650-22-T, Contract NAS9-9784,

May 1971.

II-A-3

 MSC-05546

Required inputs" to the computer program ate:

a) altitude of the sensor.,

b) target reflecbivityt

c) target background reflectivity;

d) solar zenith angle;

e) soiar-sehsor azimuth angle;

f) sensor'viev angle;

g) atmospheric visual range.

These values were available from field observations, Skylab Best

Estimate of Trajectory Ephemeris Data (SKYBET), mission tapes and

ephemeris data;

Radiance at the S192 aperture determined from ground truth

measurements and the path radiance program were presented as a
curve of apparent spectral radiance 1N5 (X)] versus wavelength
(X). For example, results of SW3 measurements of the Great Salt
Lake Desert grbund truth site, te presented in Figure A.I-l.

S192 response to this input tadiance was determined as

follows.

For a constant input radiance, IdA, £rIoh A to X -l-

dX, where
I is a constant spectral radiance indtpendett of A, detector
output is SX(X) dX, fto each S192 band, X 1,2j3,.,. 12, For
an input spectral radiance Ix over the frequency range Xxl to

AX2' the detpteot output is Xf S9x) dA. 8192 tesponse should

be equal to the input spedbtni ftdiatze5 t4 ThetEfdre, a

conversiti fdctor Aj, is,pplied to the detector output so that

I = A kx) d [A.I.4]

solVifig Edt A, ffid eutivetk~ol k&cLd± Htquldd IJ glVei by

4±lf ,A h
 MSC-05546

15.0

14.0 RADIANCE AT EREP

13.0

12.*0

11.*0

10.0 1.0

i 9.0 0.9

8.0 0.8

7.0 0.7

6. .6-
SPECTRAL SENSITIVITY
5.0 0.5 S,(0) (BAND 6) %s()

4.0 0,4

3.0 ; 0.3T
2.0 0.2 S (X)

1.0 0.1 -------

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 . 1.3

44 INCREMENTS

WAVELENGTH (9m)

Figure A.I-.- Sample determination of ground-truth atmospheric and target

radiance for an S192 band.

AI -X2
[A.I.5]

SsX(A) d

xi

For a ground target input spectral radiance that is a function of

X, from A to X + dX, N (2) dX, and the related S192 output for band

X, r(X) dX

r(X)dX A Sx () d[

IdA
[A.I.6]

NS (,)dA,

III-A-5

 MSC-05546

solving for r(X)d

A S (x) N S(X) dX
r (X)dA = I [A.I.71

substituting for A and integrating, where R = fr(X)dX

XfX2 Sx(X) s( ) d
RADx = x X [A.I.8]

Xf Xsx( _ dX

where

RADX = S192 average input spectral radiance for band X,

accounting for spectral sensitivity

S (A) = instrument spectral sensitivity for each band X

x (X = 1, 2, 3, 12)

NS(X) = input spectral radiance.

Equation A.I-8 was used for determining 3192 response to

ground truth spectral radiance. The integrals were evaluated by

using a curve multiplication program that performed a point-to­

point multiplication and summation:

44 ( ) N Si(M
=i SXi X

R GTX 44[A.I.9]

i=l

where

RGTX =ground truth spetEhal radiance for each band X

(X = 1, 29 3 . ., 0)) accountitig for the spectral
sensitivity of that band

III-A-6

 MSC-05546

SX(X) = S192 spectral sensitivity for each band X

S (X = 1, 2, 3, . . . 9) for each segment, i, where

a maximum of 44 segments were used over the band

so that AX = a constant

NSi(X) = spectral radiance at the S192 aperture determined

from ground truth measurements for each segment, i,

whoseovalue was determined to coincide with each

AX increment of the band X spectral sensitivity
(X = 1, 2, 3, . . . 9)

This operation is illustrated in Figure A.I-l.

S192 digital output in counts corresponded to the ground truth

radiance. The counts were obtained by selecting the instantaneous

fields of view (picture elements or pixels) corresponding to the

ground truth site and determining average counts for each band. In

most eases, the size of the ground truth site limited the sample

size to between 50 and 100 pixels.

S192 response in counts per unit of spectral radiance could

then be determined.

Average counts over ground truth site

Response for band X =
for band X (counts)
RGTX(mW/cm2 -ster-jim)

Assuming the response was linear, it is the slope of a line

for counts (0 to 255) versus radiance whose dynamic range was

determined by the slope and count limits.

However, this calibration varied with the low-frequency noise

on the output. To account for this effect, the response determined

was used to find the equivalent calibration-lamp radiance for each

band by using the output counts when viewing the calibration lamp

for each scan line of the ground truth target.

EGRX = SP1 -X [A.I.lO]

SREP
X( x)A

where

ECRX = equivalent calibration lamp spectral radiance for band

X (X = 1, 2, 3 . . . 12)

TII-A-y
 MSC-05546

RESPX = response for band X ( = 1, 2, 3, .. 12)

CLC X = average calibration lamp counts for band X

(X = 1, 2, 3, . . . 12) of the scan lines used in

obtaining RESPX

This equivalent calibration-lamp spectral radiance is then

independent of the low-frequency noise and can be used to obtain

the slope of the calibration curve [counts versus radiance

(mW/cm 2-ster-pm)] applicable to a target of interest when the

calibration lamp counts are known, i.e.,

Radiance
Interestof
x target of) EC RX
of \ltret
/Target (Target ofoun x [A.I.Il
nterest
(interest f
Interest Counts

calibration
counts x /
 MSd-05546

II. USE'OF TOTAL LUNAR IMAGE TO DETERMINE ABSOLUTE RADIOMETRIC RESPONSE

In the S192 Production Data Processing System defined in PHO-TR524* the

difference in signal output between a high -(bright) and lo7 (black) calibration

source is used as an internal calibration reference. Thus to calibrate the

several bands it is necessary to find the difference in radiance between two

external sources which give rise to the same change in output signal as in gen­
erated by this internal reference. The mean signals from the moon'and space

have been used in this way. Assuming that the output is linear with radiance,

one can obtain the calibration constant for band j:

L L Cg

where C is the mean count output on the high calibration source

CLj is the mean count output on the low calibration source

C(. is the mean count output on the moon

C is the mean count output on'deep space at the same scan angle

j indicates the jth band

In order to make these results compatible with those obtained by others

participating in the S192 performance evaluation, CH and CL were taken as the

mean value of all six recorded values available for each SO except for those

bands with active Automatic Gain Control for which (CH - CL ) was put equal to

255. The steps used to obtain values for Lj and (Co - Qj) will be described
in the following subsections.

THE SPECTRAL RADIANT INTENSITY OF TAE MOON

While there is an extensive litetature on the photometry and tadiometry of

the moon, a search revealed only two programs whose results were relevant. (The

remaining papers were too limited in spectral coverage, spadial coverage or

lacked absolute calibration, or more often suffered from two or three of these

*Philco-Ford Corporation, Earth Resources Production Processing Requirements

for EREP Sensors, NASA Report No. tP'O-TR524 Rev. A, Change 2, Johnson Space

Center, Houston, Texas, 1974.

III-A-9

 MSC-05546

defects.) The most useful results are those of W. M. Irvine and his co­
workers** (and a series of earlier papers).

This article gives spedtral magnitudes and spectral geometrical albedos for the

moon as a function of lunar phase angle. 'Unfortunately the results, only cover

the spectral region from 0.36pm to 1.063m. However the region from 0.36m

out to 2.50m is covered by the work of McCord and Johnson***. Over the spec­
tral region (0.36pm to 1.06pm) common to both ptogtams the results are incon­
sistent by up to about 10%. This, and inconsistencies between measurements made

by the Irvine group at different times, are commented on in the referenced

Irvine paper. The possibility of a variable moon is discussed but discarded,

and the authors find no firm reason for the discrepancies. Further, the McCord

and Johnson results are not given in absolute radiometrlc units and were made
on relatively small areas of the moon, though A reasonable estimate for the
entire moon can be made from the results given. Such an estimate has been used
here.

The Irvine results are given in terms of spectral magnitudes and spectral

geometric albedos together with the change in spectral magnitudes as a function

of lunar phase. The albedbs and the changes in magnitude with phase were used,

as these two concepts are well defined in the paper whereas the concept of

spectral magnitude was not. Since the albedo is a measure of reflectance rather

than radiation this course involves using a measure of the solar spectral irra­
diance at the moon derived from some other source. Suitable tesults for the

solar irradiance at the moon derived from some other source. Suitable results

for the solar irradiance at the mean earth-sun distance are given in the Air

Force Cambridge Research Laboratories Handbook of Geophysics and Space Environ­

tnents*.

The spectral geometric albedo tpsults from the Irvine paper are given in

Figure A.II-I. Superimposed are tht tegultt ftom MtCord normalized to fit the

Irvine results as well as'possible. the solid line Is a test estimate and re­
presents the ialues used. The scatter of the poiftd Suggests that the line

lies withit 10% of the true value qtid ptbbAbly withiti 5%.

A*A. -P.Lane and W& M, Irvine, "Monochtotfiatit t'hdoo tutvs and Aliedoi for the
Lunar DistdQ' Agttotionlai Jouual, VodIi M, 1No. S, AVtil 197S, V. 267.

A*AkT,, Mc~btd Mild T, V. dttidoft, 11 t1id1! 9PP-chtUi flbildtivAity (d.ja to 2.30

tNictonsfl"' Sciehtd Nu. 169, 28 Augult 1970u 9 V, 853.
*S. t, Vailey, d., tatdbook c4 dbjdiuycd dud §pave hifl a Air oteee
Camtbidge keteakh tabotdtotibd, beAard, asutehuJtt ±965; pp. i6-5.

i-A.-1d

 MSC-05546

1.0­
0.8- -i
0.9 11

0.7
0.6 9

0.5 8 10

0.4

nd

DA Vlue~s Usedu.L
0.089- - -- MVa Uslt

0.06 -
O. . . I 1 I t I I1
I. II

0.2 0.3 0.4 0. 1.0 2 3 4 5

b 7 8 9eo

WAV L9NkJtH (pm)

frigure A.tt.-i.- Got3fiottid gil of the tom

ttl-A--il
 MSC-05546

The radiant intensity of the moon is a function of the phase angle (i.e.,

the angle between the lines joining the center of the moon to the observer and

to the center of the sun). This phase function arises partly from the change

in projected area of illuminated lunar surface visible to the observer-but to

a greater extent from the large changes in bidirectional reflectance of the

lunar surface as a function of illumination and viewing angles. (The bidirec­

tional reflectance increases dramatically as the phase angle is reduced from

50 to 00. It was therefore decided during the planning of the EREP SPE to avoid
.very small phase angles during lunar calibration passes. Phase angles of about
150 were in fact used.)

As explained earlier it was decided to use the (spectral) geometric albedos

given in Table VIII of the Irvine paper. It is a straightforward procedure

to apply relative changes of spectral magnitude with lunar phase to these re­
sults. To obtain the absolute spectral radiance of the moon, solar spectral

irradiance values at the mean earth-sun distance were taken from the Air Force

Cambridge Handbook. The procedures used are outlined below.

According to Irvine, et al, the geometric albedo p of the full moon is

given by the relationship:

log p = 0.4(m 0 - mi)- log sin a

where a = the angular radius of the moon at the observer m's are magnitudes
defined by:

m = -2.5 log E,

EO irradiance at the moon due to sun

E irradiance at the earth due to moon

Substituting in Eq. A.II-I we obtain

log P= -log E0 + logt( 7 log sin 2

or

P %Ei ain A.II-2

 MSC-05546

Now

L
Area of Moon
2

E
(Earth-Moon Dist)2

where L is the mean radiance at the moon;

Substituting in Eq. A.II-2 we obtain

pEo
L ­
( 7T

It is convenient to consider the unprimed quantities as representing the case

of zero lunar phase (i.e., full moon) and mean sun-moon distance, and then to

write

L'= A.II-3
7T

where primes indicate values at the time of the lunar calibration period. Then,

by the inverse square law, we can put

E6 = E A.II-4

where E = irradiance due to sun at the mean sen to earth distancd

R = mean Sun to earth distance

R' = sun to moon distance at the time of the lutidi callytation

Then, in order to cotkict fut phase dtgle Vatiatioti Amt; Wut

Ant =-fi a = -.
4- id

Itt-A-IS

 MSC-05546

or

= p antiiog ['M A.II-5

Substituting Eqs. A.II-4 and A.II-5 into Eq. A.II-3 we obtain

=1 Eo( p antilog

Rearranging and putting in terms of spectral quantities,

(AA =- AXPA antilog [-m A.II-6

Appropriate values for Am for each tC were obtained by interpolation and

extrapolation from the results of Irvine's table V. Figure A.II-2 illustrates

this process for the phase angles of LC .I and 26 12044' and 15057 '
respective­
ly. (The phase angles for t s, 4 and 5 wete intermediate: 13009 '
, 14'33'

S
and 13018t " respetiUvely.)

It Wii1 be seen that taidg 1 through 8 ate well

coveted by the available tesults and that the exttapolation to Bands 9 and 10
seems unlikely to be itiserious etror. toveverf fat Bands 11 and 12 a consid­
erable exttapolatibii hags to be made so the results or these bands are suspect,

partitularly those f£W Bad ol.

Retutnittg to Eq4 Att-6, i the meat 8patai

'4
tadiance of the lunar
disc facing the bbletnvr kith the uullufilti ted ata ieltirluddd In the avetage.
All the tetilfs oh the tight can be deriovd kfai the telerenaes eited of from the
Astronomical tphtietides and the apptoptlat RP Postpass gummaty Reports. As
explained, the phAb dafhg tOtkdfofi IN £tihd bt lfetpdltibn £tofm table V of
trvihis paper, this howe'ver tequiutes a kiotwedge 6f the Phae angie of the
moon at the tite ot i&dAuktetidftti the jhddd 8f the hlbbt tetetted to thE center
Of th eafth Mi Ad 4S ttitldi dt tifd lit l4ldtfd tJiMidt hiphemetts,* The
phase ahgla tefetted Ed th SkYtAt at the time Of th itidk enlibration differed
ktom this value by le6e Nhai bit haif dgfo. A tdlatlVt eibokte tdIctla­
titm wa u8d udbltaihi the ttta Phase ati ie to the tebkd t fainute Of Atc4
though Il etrAtdebdt the ddtibmi addUialy thig gafAV t Lh& phase atigle cot­
ktttion vas hatdly hbttedaty.

*Anrn D. MAttutig; tttidt Ephteti

nnSlttdg*&Phid butdifiahed t em tAt arid
8uH fut 173 atid 15741 RNA tepoTt NO. V-1058d1 bdatiild and Techical

infoi-matio Oi 4972,

Jtt.

 MSC-05546

The spectral radiances found using Eq. A.II-6 were weighted by the spec­
tral response (R) of the several S192 bands using the formula

ft'( RidA EL (AI)RiAX A.II-7

f RdA ZR(AX j)SAXJi

Here"i' is the band n

umber and Ai is a series of adjacent wavelength inter­
vals over which L and R
had been tabulated. Generally the 6A . were the in­
tervals for which the solar irradiance was tabulated in the Air Force Cambridge

Handbook. However at the longer wavelengths these intervals-become wider than

the structure of the R. For these, therefore, suitable narrower intervals were

selected and interpolation was used to obtain the necessary values of E0 . This

particular method of weighting was used to make our results directly comparable

with terrestrial scene results presetted in Figure 3.4.1-1 and Section I of

this Appendix.

0.80

1 11Phase Atigle

zB
0
0 hase Angle=
0.74
A L
C,,

0.4 0.6 so~ i.U 1.9 1.4 1. 1.9 .0 2.2 2.4

WAMt NtM (if)

Rigute A.1t-l Ctkcetiott tot i.utt phas§ angle.

t1t-A-iA

 - - -, - tAc-5546 I

Th6"i dt E.>Alr-7 wag tabulated ss illudtahh'ed ilTable

A.II-I. The contentS of the:tabe were tainched onto -aperl tape and the
integration carried out on the nompter,- Ndw phase corrections ahd ranges were
,calculated fori each lunar caiibration,,puncied.onto a tape qeed to correct the
earlier tape, and.a new inttgration carried out; This exercise was also
carried out for LC 6 but-thd bdiplicaEions ntr6duced by'the large phase angle,
72-25', and the severe frequency dist6rtiot in most S192-bands have prevented
completion -of a calibration ,for Ehis LC. The resulting estimated lunar radi­
ances noi-malized -byband may-be seen in Table A.tT-2.

MEAN SIGNAL LEVELS FROM THE LUNAR DISK * + "-"

Having obtained the mean radiances of the moon it remains to find the cor­
responding mean S192 outputs and calibration levels.

During each LC the SKYLAB was rotated about an axis parallel to the mid­
dle of the scan line in such a way that the scan lines moved onto and across
the moon with about 95% overlap. Thus when an image of the moon is developed
in the ordinary way from the recorded data the moon appears elongated by about
20:1.

For each LC a block of data was'selected in which the first scan line was

about 20 lines in front of the moon and the last one about 20 lines after it,

as shown in Figure A.II-3. The same section was used from each scan line.

This was about 250 pixels in length, so that every section started at least 20

pixels before-the moon. As the diameter of the moon in the scan line direction

.was about 55 pixels, this left at least 175.pixels of deep. space after the
trailing edge of thd moon. 'This wasdone so that'thesignal overshoot and de­
cfy following the moon totld'"-be taken into fcbotnt. As the moon lay apprbxi­
mate1y in the middle- of-,the scan in each LU the effects of scan line curvature
wdre very smhll-and, were -igiored. The bleok of data selected was processed in
,the compuiter .o give the mean count leAiofbt each cotlutmr of 'pixels and also a
cumulative total to facilite-nauat data manipulation, The-mean count value
* for each,,woard of the"higr-:and l& callitta-6n signal was'printed out at the
same time. -"

In 'ptinciplw the meanl'totlft, valu- over the moon cat be found by multiply­
ihg the .mean colit valt Iota bl6&k ok dat eloskig the moon by the number
of pixels, in this block ot'dati divided by tht hutnbet ot pixels failing inside
the lunar ttcumfekdtic.. tlfmptadti dgos-it{ bf Utittrtainties were
encountered.

--(1) The 'td-is liifcftaihty tsgairditg thp_ extent to tshkhc the frequency re'-
Ypoise oftht sevethl battdg thatgs -ithd' t dlanE beittg reviewed, The de­
tectcrs-used ith sb
theed kinii-feadt
h b' The ktequeticy response of the
amplifiers gssthaapd §d-tht the'-tMhtlEIft 9de dtlul- ttbnibs frequency
response wad Optimumt iyphattsfemtt'il sdee leiels Wind the degtaded
tespo ide at tth l velgd ddbpbd4. thud it wg hot iutp1silng then examination
of ploth .of'dcan lifidg dtdbding hd iftdbtt ,h6wdd, in tidst binAs a onsiderable
negative overshoot bithdiateify f olidtffng ,he Xtiat ttailit edgd followed by a
d theit- 6w -d
M&&ovety A~WAbtl- t jg ghb(oh. iii 1Pigtt A.f1-4.

Itt -46-­
 MSC-05546

TABLE A.II-1.- SAMPLE OF TABLE CONSTRUCTED TO FACILITATE

CALCULATION OF MEAN LUNAR RADIANCE BY BAUD
E R5

Interval E -- 1 to
No. (QArn) (Wm -2AX ) RI R2 R3 R4 RI12 PAX 0AX

21 0.500-0.505 9.80 0,01 0.50 0.93 0.07 0 ,0.116 0.717

22 0.505-0.510 9.80 0.01 0.43 0.98 0.12 0 0.117 0.711
23 0.510-0.515 9.65 0.01 . 0.36 0.998 0.21 0 0.118 0.718
24 0.515-0.520 9.60 0.01 0.26 0.99 0.33 0 0.120 0.718
25 0.520-0.525 9.65 0.01 0.15 0.93 0.56 0 0.121 0.718

Ri is the relative spectral response of Band i.

p4 is the lunar albedo extrapolated to zero phase angle.

0 A is the fractional correction for phase angle which must be re-entered for each lunar
calibration.

TABLE A.II-2.- ESTIMATED LUNAR RADIANCE NORMALIZED BY BAND, L

2
milliwatts cm sr1 per bandpass

KSC* LCI L02 LC3 LC4 LC5

Panel 12044'
6
15057' 13009 14033' 13018'

Band Radiance 0.400 x 10 km** 0.395 km 0.384 km** 0.355 km** 0.352 km**

1 0.92 3.99 3.70 4.02 4.05 4.19

2 1.52 5.02 4.65 5.06 5.09 5.27
3 2.34 5.31 4.93 5.36 5.40 5.58
4 3.02 5.68 5.27 5.73 5.78 5.97
5 4.13 5.74 5.34 5.80 5.85 6.04
6 5.24 5.40 5.03 5.46 5.50 5.69
7 6.54 4.30 4.02 4.35 4.40 4.54
8 6.46 3.35 3.16 3.38 3.42 3.53
9 . 6.09 3.07 2.88 3.10 3.15 3.25
10 5.57 2.78 2.62 2.81 2.86 2.94
11 3.39 1.62 1.53 1.64 1.67 1.72
12 1.51 0.74 0.70 0.75 0.76 0.78

*Values derived from Optronic Laboratories Letter Report to Mr. Richard Juay/TF3, NASA,
Johnson Space Center, Houston, Texas, 17 January 1974.
**Phase angles and sun-moon ranges given are those which existed at
the time of each
LC.

II-A-17

 MsC-05546

lJited invegtigation dt frequency response effects based on the SL2 data btig­
nested that fkequacy tespotise1 except for hand 1, was much better fo t omal
sceti- tadihtce than for the darkness of deep space. ttider these tirEmstances
the iAgihg digtiai M e an Artifat of the hysteat nohttlieatitieS and Should be.
excluded fto the meari signal calculated for the moon.

To fbllow the Way in which this was done refer to Figure AMIt-3. The
block of data from scan lines N to (1 + M - 1) and frot pixeiS A to At Was
Useds A and A' were the pixel traceg judged to overlap the edge o the Moon.
The total of the outputs fOt each pixel in this rectangle can easily be forimed
from the printout referred to above. From this total must be subtracted the
ktsiddai signals linareas a and b and the residual plus lagging signals in
area8 t. Prom areas a the contribution of the raw residual signals is removed.
Fdt Ih idti-k area b tesidual Signals corrected for clipping are removed. For
ateas C corrections fot the lagging lunar signals must be made. these cottec­
tions tor the lagging lunar signals must be made. These ctrrections t tera sti­
mated by iniectioti of the printout values for columns of pixels immediately
Eoilot ng the moon (the region indicated by B). WThile this last is not a very
exact procedure, the bortections for area c are typically about 1% of the total
doUtit so eveli Al unlikely 30% error in its estimation would not be vaty
gigtiificalt;

To Investigate the size of errord which might be introduced by these

bgsUmptlot on Ere4denCy tespoflse, the mean signals from the tioof were calcu­
lated trao the saae computer printouts in an alternative way. Then the average

bf Ali pixels froM A to (n + m - 1) (Figure A.It-3) WtaS fobmed corrected tor

tesidual effects by subttacting the mean of all pixels from lq to (A-i) k The
esult wad theti multiplied by the area of the rectangle bounded by pixels A and
(t + - i) divided by that of the lunar image to give the mean inatr sigatal.
This Was eqfuivalent to Assuming that each band has constant ftequecy response
Mnd thetaetre the Coditts due to the lagging signals follwig the moon wete
iftriuded It the tolnt t4owever, as will be seenj the differetceg between the
tesUitd givtd by the two methods are not too significant.

(2) the area of the lunar image was found by inspection from ttaces of
tAfn iitied and pixels crogsing orthogonal lunar diameters, Small cottedtiohf
(tl%) Were made fot the areas of the mooh in shadow at the time of observatLoti.
In the scan litie direction this could be done with fair accuracy and thedletd by
bofpatiso with an estimate based on the scanning parameters of the 81925 and
ephemotides ahd Skylab data from which the size of the mobil and the tooti-to­
bkIYLAY distance cart be detetiined. The results are believed accutate to 0.25
pixels or 41/2%, As the lunar image was always elode to the center of A gcAni
line sld occupied less than 65 pixels, effects of scan line curvature were
small amid hdiat been ignored.

Vor the long (pixel-trace) diameter the situatibtf was awkward due to the
sdan lifie overlap of about 95%. if the moon appeared Uniformly bright a pixel
tkace aiong the long diameter would appear as a regular trapezoid add It wotild
be a stralgtLfortwazd matter to estimate the length of the lutiar image, which
ig the distance between the half-down points of the sloping entds of the ttpe­
zoid. AS the btightness of the moon is quite irtegular however 3 the shouldets

I-A-l8
 MSC-05546

N n A A' n4m-1
a

an

--F--------------------------------------­
al
N+M-1

B
The printout gives the mean count for all pixels in columns n,
n + 1, n + 2, .... n + m - 1 consecutively. Thus the sum of
these values multiplied by M gives the total of all counts in the
rectangle of size m XM pixels.

Figure A.II-3.- Selection of lunar data.

250 2
High Cal
Signal

200

150

50

Low Cal Signal

/ I
126
I
252 378
I
504
I
630 756 882 1008 1134 1260 q
PIXELS

Figure A.II-4.- A-trace of averaged scan line centered on lunar

diameter (LC 1, bat 6).

III-A-19

 MSC-05546

of the trapezoid are so poorly defined that it is not possible to determine the
half-down points with any accuracy. Consideration was given to solving the
problem geometrically, but some formidable coordinate transforms would have
been required and it appeared that the SKYLAB ephemeris data was too infrequent
(20-second intervals) to allow unambiguous determination of the SKYLAB rotation
rate for the 12 to 14 seconds of the observations. However it can be shown
that the overall length of the trapezoid LT is related to the true length L by

the expression

LT

1 + oN

where 0 = angular width of scan line

0 = angular diameter of the moon

This method was used because as e is known, 0 can be calculated from

ephemerides data and LT can be estimated from appropriate pixel traces to bet­
ter than ±0.5%. Again, corrections were made for the shadowed area of the
moon.

izl-A-20

 NSC-05546

III. DETERMINATION OF TRUE MEAN VALUES FROM CLIPPED MEAN VALUES

One of the S192 performance characteristics, scan related

system offset, was its response as a function of scan angle to

zero radiance input. Deep space was used as a scene with

essentially zero radiance. The digital output of S192 ranged

from 0 to:255 counts. For low input radiance values, many

digital output values were zero. These zero digital values

corresponded to any input radiance values that were equal to ox

less than the input threshold radiance (0.5 equivalent counts)

corresponding to the digital output of zero. A mean of the

output values, some of which are zero, was designated a clipped

mean, while an unclipped, or true, mean was one taken over values

that were not constrained to be zero or greater. The relation­

ship between a clipped mean and its corresponding true mean was

obtained in the following manner.

It was assumed that a signal X not constrained to be zero

or greater has a true mean and standard deviation of p and u,

respectively, and a Gaussian noise distribution. The relation­

ship between pc3 the clipped mean, and 1, cr, and X is given by

(255

JO&5 xe~a

The units are the equivalent counts for the analog signal. The

integral's lower limit is 0.5 because the action of the analog­

to-digital convertor was to set all valueg of the signal X equal

to or less than 0.5 equal to zero, This, expressed for pc in

terms of V and a, results in the family of curves 9howi in

Figure A.III-1,

by using the figure, given a value of the standard deviation,

as it was possible to convert rlipped mean values, pc§ to true

mean values3 p) and thereby detetttdtie the S192 response to a zero­

radiance scene.

III-A-21

 - i .-
. .. t.....
4 V ",
,-.'-

r--- I--
- '. .-
- "-- .-
t-- .-- f-'-.
- -r -- i -j .Im
94

a.a --­

a 6
L­

'7 7
ha­

t7o­

-,- L~. J4-i L -- +1

D.r -5 -4 -3 -2z -1 0 1 12 3 4 5 6 i 7

UNCLIPPED MAN VALUE (counts)

Figure A.III-I.- Correction curves for clipped mean values.

 ':MSC-05546

iV. DETERMINATION OF RESPONSE VERSUS SCAN ANGLE CORRECTIONS METHOD FOR BANDS 1

THROUGH 12

To correct raw S192 outputs for scan angle effects an additive and a mul­
tiplicative effect are required. These are defined in PHO-TR524* by an equa­
tion of the form

Cij = Rii - Z. -CE] A.IV-1

where i is the band number

j is the pixel number

C i are raw data in counts

C'j are the corrected data in counts

GL is the average count value for the Low Calibration Source

Rii is referred to as the Response vs Scan Angle function and Zii is re­
ferred to as the Residual Effect. In an ideal system they would be identically

unity and zero respectively. In practice they are not and must be found

experimentally.

Z can be found easily by viewing a black scene for which C is zero so

that

Zij = Ci - CLi A.IV-2

Before launch a test panel in total darkness was used. In orbit, deep

space viewed during the lunar calibration (LC) periods provided an ideal scene

for this purpose. Tapes of raw data were obtained for each IC. Usually one

sample before and one after the moon were used, both being spaced as far from

the moon as possible, using available data.

However a major problem arose in using this data, in that the S192 digital

system translates negative numbers as zeros. Thus the noise and negative val­
ues occurring in the residual signals cause many negative output values to be

recorded as zeros. As a result each pixel value in an averaged scan line tends

to be offset by this clipping of the negative values. The method described in

Section III of this Appendix was used to "unclip" the clipped values.

*Philco-Ford Corporation, Earth Resources Production Processing Require­

ments for EREP Sensors, NASA Report No; PHO-TR524 Rev. A, Change 2, Johnson

Space Center, Houston, Texas, 1974.

III-A-23

 4SC-05546

A computer program was written which implemented the "unclipping" scheme

on a pixel by pixel basis. All the restlts given here have been corrected us­
ing this "unclipping" program.

Having found Z using Eq. A.II-2, Itcan then be found using data from a

uniform scene, for which C is constant.

To obtain results for flight conditions, S192 screening films were search­
ed for uniform or at least homogeneous scenes. These proved-more difficult to

find than had been expected. Most of the data used was of agricultural areas

which appeared homogeneous, though inspection of averaged scan lines formed

from this data led us to reject some samples. In the rejected samples the

averaged scan lines appeared much less smooth than those generated from pre­
launch data. A few ocean scenes were used for Bands 1 through 5. Unfortu­
nately these scenes, though often appearing highly uniform, give outputs which

are too low to be of any value for Bands 6 through 12. One desert and one

cloud scene were used but otherwise such scenes were never uniform on the scale

(40 miles) required. In some scenes an apparent scan angle effect may occur

due to changes in the bidirectional reflectance of the scene as the azimuth of

the viewing anglo changes thtoughout the scene. To reduce the effects of

gradients introduced itt this way or by other scene anomalies, the results for

several scenes wetre used.

In the use of Eq, A,1V-l i this Vag the seale of 0t is defitied,by putting
the avetage value of 1/ki equal to unity, Avetaged ciutvd kot it and z were
formed tot each Iibish o mission sectio whaeti a chtidg in the system config­
utation of the 1lfttUiUitt hfda th{ d a§1bld. Phd§d Avtraged ttitVes Wete then
fitted by ledst bqUAt s techti4tfeg, by sixth etdae poioynomaials, or for some
of Ehe Z'g4 by two such poloyomil pieced together to give a smooth Junction
between the two artt Of the cu-e, ttive fitting In this way proved a satis­
factory way of kfoothint the tioise did small stere artifacts left after the
averaging ptocetsed desetibed above.

METHOb - 13AtD 13
At explaied vreViosily deep spabe det etat be used ii a §tmaightforwatd
way to obtaiti residual etftets ±eduits (Zi) tot the tisible tear-ik, Bands i
through 12.

Untotoingtely tot BLid IS thig tdtinot be done a the low end of the te­
corded dynamic range tots otds hi a tadihdd high efhogh that the deep §pace
data Is completely clipped,
Zapiaihed teldtid the altetiatlve inethod whkh ban bFeti Used with band 13
tlight data t8 obtai bsitiates of k11 6tedh respbttsd LUtibtioti afid tesldial
effects cortedtih.

'tt-A-14
 wSC-05546

As explained previously the scale of R is fixed by normalizing Rij so that

for each Band (i), (I/Rj) is unity. Thus by averaging Eq A.IV-1 over j we can
write for a uniform scene
C, =C - Zij - CiL A.IV-3

Substituting back into Eq. A.IV-l gives

- Z - C = R. [C. - C - Z.] A.IV-4

J j

where we have dropped the subscript i as we are only considering Band 13.

Then if we have data from two uniform scenes with significantly different
C.
2 as denoted by subscripts 1 and 2, we can tewrite Eq. A.IV-4 as a pair of

simultaneous equations.

ol -z.j-c C Cl -cTL- A.IV-5

-2j - = R j -
2- 2L-

Subtracting and rearranging we obtain

R = j - r2 j - rIL +
_2L

A.IV-7
Clj - - '. + 2L

which gives R. in terms of quantities whrh ca readily be derived from the

3
data. As R is tow known we can rearrange Eq. A.IV-5 to giveZ thus

IY

Zj =C 1 C1L [-[CL
j A.IV-8

The terms in Eq, AIV-8 are known etept lot z. Inspection of Sections

5.3.9.1 and 5.3.9.2 bf PHO-TR524 thbws that the Utigin of the Ct stal fo
Band i must be selected so that scene elemdntg tot which C' = 0 have the same
radiance as the low talibtatio soUttce.

Thus we must have

C' = kl(Lsj -L ) tr1 -3 AV-9

 4SC-05546

where kI = a constant which determines the scale of C'

L = radiance of scene element j

Si

LL. "= radiance of the low calibration source

combining this with Eq. A.TV-l we obtain

L 4 A.rz-Lo
R j L)
- r
For uniform scenes we can average over j obtaining

L Sj _-L +' A. IV-Il

k
kIc

Thus we see that if data from uniform scenes of known radiance are used to find
Sit will not be possible to separate it from the constant term in
Eq. A.IV 1l, viz:

+ LL

k1

Thus if Band 13 is to be calibrated in this way we are free to assign any

value we may choose to Z.. We have chosen to assign the value zero to part-
J
ly for convenience and partly because in the abscence of a valid calibration
this is the most probable value for this constant, as in an ideal sannei Z.
would be zero for all J.

In practice we used the scene (or an average of two scenes) which appeared
to be most uniform in Band 13 as scene I and derived R using this together
with each of the remaining scenes. The average for each j of the R.
J
found in

this way was then formed and plotted as the best available values of R and

was also used with the scene data to give Z with the assumption Z = 0.

We can however demonstrate that Eq. A.IV-3 provides an accurate descrip­

tion of the way a thermal scanner actually operates and thereby find a value

for Z in terms of system parameters.

For such a scanner a non-uniform scan angle response implies that the op­
tical transmission is a function of scan angle. However, when the transmission

is less than unity the optics must have emissivity and/or reflectivity in

complement:

+ r + e = I A.IV-12

III-A-26

 MSC-05546

where t = transmissivity

r = reflectivity

e = emissivity

Thus assuming that the S192 optics and its surroundings have a uniform tempera­
ture T we can write the apparent radiance of the scene as modified by the op­
tical train

L = tL S + (r + e)P(T)

where L S is the scene spectral radiance integrated over Band 13 and F(T) is the

corresponding radiance of a black body with a temperature T. If the system is

not at uniform temperature we can find an effective Pj for each scan angle pro­
vided the temperature distribution does not change with time.

Adding the subscript j to those terms which may be dependent on scan

angle, adding a term E. for electro-magnetic pick up, and using Eq. A.IV-12 we

obtain

L.=tL + (1 t) + E A.IV-13
J isi iii j
L is thus the radiance presented to the detector when the scanner views
the scene. If 1.2, is the radiance presented to the detector when the scanner
views the cold black body used as the internal AOC reference and assuming that
the detector is linear we have.

C -CL = k (L A.IV-14

where k 2 is a constant to be determined experimentally. Substituting from

Eq. A.IV-13 we obtain

+
C -CL = k2[t Lsj (i - t )pj + E -T4] A.IV-15

Substituting in Eq. A.IV-9 gives

o-cL C'

-jr- CL = t' 'i + (I t) +E+ L

kI j 1J l-jr E - +tL

Rearranging, we obtain

C j - L- - )pl + - + t 1] A.IV-16

III-A-27

 MSC-05546

This equation is identical in form to Eq. A.IV-1 indicating that Eq.

A.IV-l is valid for band 13 provided the temperature distribution and optical

constants of the scanner optics and their surroundings do not change. By com­
.paring coefficients we see that

k 1

Rj t A. IV-17

and

Z= k 2 [(l - t.)P + E - L + t LI A.IV-18

While these equations give physical meanings-to R and Z they do not pro­
vide means for determining these parameters as several of the terms are unknown

in practice.

Thus we are unable to determine the true value of Z.. As demonstrated

this is of no moment if an external calibration can be carried.out. Putting
estimated values into the righthand side of Eq. A.IV-18 we obtained Z = 2.6
counts. -However, in doing this it became apparent that this effect was caused

by the finite emittance of the primary and secondary minors. In fact, the PDP
system makes corrections for this effect when the raw data is converted to
radiometric units. Thus, in the absence of an external calibration, Z. = 0 is
a
the best estimate we can use for such data. However gradients in transmis­
sivity and emissivity across the optics may produce errors of a few counts.

III-A-28

 MSC-05546

V. DETERMINATION OF FREQUENCY RESPONSE AND OFF-AXIS REJECTION

Data gathered by S192 when scanning the moon and surrounding

deep space were used in determining two sensor performance

cha-acteristfcs. The sharp radiance cutoff at the edge of the

moon was used to determine system spatial frequency response. The

relatively small solid angle subtended by the moon approximated

the ideal point source required to check off-axis rejection, which

is a measure of system sensitivity to undesired light scattering

or reflecting from S192 surfaces.

A. Frequency Response

A measure of a system's frequency response is-its modulation

transfer function (MTF), which is the amplitude of the output sine

wave divided by the amplitude of the corresponding input sine wave.

The MTF is normalized so that its numerical value is one at zero

frequency.

One method used to determine a system's MTF is based on its

response to an input step function, which instantaneously changes

from one constant value to another. Such step.function inputs for

S192 were approximated during each of the six lunar passes. First,

as the S192 scanned from deep space past the leading edge of the

moon, and again, in the opposite sense, as the scan went off the

trailing edge of the moon and returned to deep space. Because the

surface of the moon is not uniform in radiance, its trailing edge

provided a more suitable step function'for determining MTF than

the leading edge.

The image of the moon produced by S192 was a very elongated

ellipse because the along-track motion of the S192 field of view

was too slow to provide an accurate reconstruction of the lunar

image. Because of this characteristic, an average scan line could

be obtained from 100 consecutive scan lines chosen so that the

moon's trailing edge occurred at essentially the same point in

each of them. The scanner output averages from deep space were

corrected for clipping as described in Section III of this

appendix.

The MTF was obtained from the average scan line by first

taking its derivative with respect to scan dngle, then a fast

Fourier transform of the negative of this derivative, and finally

normalizing the transform to one at zero frequency.

III-A-29

 MSC-05546

B. Off-Axis Rejection

Off-axis rejection is the rejection of undesired light

scattering or reflecting from S192 surfaces. It was checked by

looking for the presence of a ghost or halo near the edge of the

moon.

.Two edge geometries were examined. The first was the leading

edge-of the moon, as discussed in subsection V.A. One hundred

consecutive scan lines were averaged, pixel by pixel, along the

flight path to determine an average scan line. A plot of these

averages as a function of scan angle revealed the presence or

absence of a ghost or halo near this edge of the moon. The other

lunar edge considered was that encountered when the moon first

appeared in the scan, or, in other words, at 900 to the first edge.

In this case, 10 consecutive output values centered on the moon

were averaged. These averages were determined for consecutive scan

lines, starting when S192 scanned only deep space and ending when

the scan was well into the moon. A plot of these averages as a

function of scan line revealed the presence or absence of a ghost

or halo near this edge of the moon.

III-A-30

 MSC-05546

VI. USE OF POWER SPECTRAL DENSITY PLOTS FOR NOISE EVALUATION

One noise characteristic of concern in S192 data was the noise

frequency distribution as represented by power spectral density

plots. Let X(t) represent the 8192 output as a function of time

when scanning a uniform scene. If X(t) is thought of as a voltage

applied to a unit resistor, the average noise power P during the

time interval from t = 0 to T would be

rT
2
= 1 [X(t)- j3 dt [A.VI.1]

where X = fT X(t) dt

While P would have provided a measure of the noise in the S192

output, it was more useful to determine the power spectral density

S(f), where f denotes frequency. S(f) was found by squaring the

absolute value of the fth Fourier component of S(t), i.e.,

S(f) f ej 27f t
[X(t) - XJ dt 2 [A.VI.2]

where

j= T
e = natural log base

= where n = 0, ±1, ±2, . .

S(f) = power associated with the component of noise whose

frequency is f.

P and S(f) are related by

n ­

III-A-31

 MSC-05546

The usefulness of the power spectral density, S(f), stemmed

from this relationship in that it provided the frequency

distribution of the total noise powers P.

It should be noted that false peaks were introduced into the

power spectral density plots by the periodic interruptions in the

values for X(t) caused by the instrument scanning characteristics.

X(t) was only known for that third of the time when the scan was

actually occurring. These bogus peaks must be discounted when

evaluating the power spectral density plots.

ttt-A-i
 MSCr05546

VII. DETERMINATION OF GEOMETRIC DISTORTION

Analysis of S192 geometric distortion used a three-dimensional

mathematical model, based on the geometry of the S192 dynamic data

acquisition system. The model performed absolute spatial

registration of the S192 data. -The theory and the mathematical

details of the data reduction techniques are given in LEC/ASD

Technical Report "Absolute Spatial Registration of the Skylab S192

Conical Scanner Imagery by Means of Dynamic Modeling", August 1973

(LEC #0801). An "S192" computer program in FORTRAN V source

language was developed and used for data reduction in accordance

with the theory given in LEG'#0801 for the S192 sensor geometric

performance evaluation. The procedure is given below.

A set of topographic detail poirits were selected on topographic

maps (1:250,000 or larger scale) and also'located on the S192

digital image display of the scene being evaluated. The map

provided the object space coordinates (geographic, geocentric,

geodetic, or local vertical) of these points; whereas, the digital

display provided the scan line and sample number counts to, locate

the same points. The line and sample counts were converted to

elapsed time relative to a predefined reference time and were used

to obtain the scan direction in the S192 image coordinate system.

For each topographic detail point, two observational equations,

one each for X and Y coordinates, were then established by the

mathematical model, which was based on collinearity between the

scan direction and the vector defined by the scanner location and

ground point being scanned.

Additional observational equations were written for each of

the parameters being solved for by using previous knowledge of the

.parameters. A weighted least squares solution resulted in the best

fit between the set of points in the imagery and on the ground,

which resulted .in residuals (AX, AY) in the X and Y coordinates of

the points used in the solution. Residuals were also computed at

several check points over the entire scene.

Salient features of the mathematical model and the data

reduction technique included:

1, The earth was considered an ellipsoid of revolution.

(Fishers': Major Axis = 6,378,166.00 meters,

Minor axis = 6,356,784.28 meters)

2) The rotation of the earth was accounted for.

IiI-A-a3

 MSC-05546

3) Solution for the parameters for absolute registration of

data was based on a rigorous least squares method. There

were 18 parameters solved for:

a) Position vector (X, Y, Z) of the sensor at the

reference time for a "frame of t"data-take";

b) Velocity vector (K, Y, Z) of the sensor at that time;

c) Acceleration vector (XYZ) of the sensor at that

time;

d) Attitude vector (W, *, K) of the sensor at that time;

e) Rotational velocity vector (W, P, K) of the sensor at

that time;

f) Rotational acceleration vector ,4,i) of the sensor

at that time.

4) The SKYBET (Skylab Best-Estimate Trajectory) information

was used to approximate initial values of the parameters

in the solution, and in the formation of observational

equations from previous knowledge of the parameters, while

solving for these by the method of weighted least squares.

However, the solution did not depend on SKYBET information,

but used it only for an initial approximation of the

parameters.

5) The mathematical model considered the elevations of the

control points, which provided a means to consider terrain

topography in resampling to generate a registered data

tape. As an extension of the sensor performance

evaluation, the technique can be used to generate a

registered data tape of a scene.

6) Assuming no abrupt discontinuity (due to thrust) in Skylab

Cor sensor) motion defined by the parameters, the limit­

atiti Of the mathemahleal model was the length of the time

blice for which data could be registered. However,

modifications could be made in the mathematical model to

consider arty desired Lime slice of data.

The following ptoduct:s were needed to ptepare input data for

the computer t r to solve the parameters of the mathematical model
for tegisgtatioft of b192 scatiie data:

It-A-34
 MSC-05546

1) A computer-compatible digital tape to display a selected

band (band 7 or 8) that gives the best topographic

details in a scene;

2) A hard copy of the scene from the Data Analysis Station

(DAS) of NASA, Houston, Texas, with an overlay grid of

100 lines X 100 samples over the scene. This produ&t

was used to obtain scan line and sample number counts

to locate a topographic detail in the scene;

3) Topographic maps at 1:250,000 or large scale for the

scene. Geographic coordinates of the selected

topographic detail points were obtained;

4) Relevant SKYBET information needed in the solution.

II-A-35

 MSC-05546

VIII. USE OF LAKES AS THERMAL CALIBRATION TARGETS

The use of lakes as infrared calibration targets for the

thermal band was especially useful because for measurements made

at normal incidence their emissivity is greater than 0.98 over

most of the spectral region. Figures A.VIII-1 and A.VIII-2 show

that the emissivity remains high and the reflectivity low at

incident angles up to 60'

.

For lakes yiewed by Skylab, the radiometric surface

temperatures were measured using a PRT-5 radiation thermometer.

When viewing any target less than 4000 meters above sea

level (99% of all targets) corrections for atmospheric absorption

and re-emission must be made at wavelengths longer than 0.82

micrometers. Therefore, a correction for atmospheric constituent

absorption was made.

Atmospheric transmission was calculated using a computer

program developed by R. F. Calfee, NOAA, Boulder, Colorado. This

program calculated atmospheric transmission by dividing the

atmosphere into layers. The require& inputs are average

temperaturej atmospheric pressure, and total water content'for

each layer. These inputs for most S192 targets were derived from

temperature/humidity profiles measured with radiosondes that were

balloon launched or dropped from a helicopter at the time of

Skylab overflight. These radiosonde data were supplemented by

adjacent NOAA network radiosonde data. The NOAA data, as well as

the ground-truth team's radiosofide data, were then plotted on

skew T log P diagrams. The ground-truth radiosonde data were

most heavily teighted in the meteorological andlysik performed to

arrive at a "composite" watet-vapot temperature profile for a

particular Skylab obsetvatioti, but weather features such as cold

fronts were also taken Into account in developing the "composite"
curves. The atmosphere wag then divided into convenient layers
based on the "Compbosite" ttofile, Watet amounts and representa­
tive temperatures were thet seiected for each layer.

The layer ta-ftperttej water amoaut, ahd lbyet thicknesses

were used aa input patamieterg to the Calfue atmospheric
ttatrtsiunft ptogmm 1 which irn the "'doptssed line" data* but

R. Mcclatehy tt.ai, iA1'dktAtnttigheititi Aligiltsripe tite Parameter

k
du ilatio, Atc lu ivittti e tttl Research Laboratory Publitation
Nu21bei 43 ed idi Muaehusgttg.

it,
tL aiteg b V.Schweiotn NIu AVerge Zttdkd Atwitptidn
tueflitiehts tE Watt Vapoti NOAA Techniral Report #ERL 274-WPL24,
K Bodet, Cootado.

fli-A-3g

 MSC-05546

40­

30­

1,
20,

80cnidneanls

0,
20­
6 8 90
2 10

WAVELENGCIMCE(°)

at 0, 60, and

Figure A.VIII-l.- Reflection from water surface

800 incidence angles.

6o0

100

r4 8
H

Hf 0

40

;->

U'' 20

0r

40 60
80 90

0 20

ANGLE-.OF INCIDENCE(0

of water versus

Figure A.VInI-2.- Reflectivity and emissivity

incidence angle.

III-A-37

 MSC-05546

did not take into account the magnitude of radiation from the

target nor the upwelling radiance from each layer. The Calfee

program was therefore modified according to the following

equation, as illustrated in Figure A.VIII-3:

R = (TI - 1) BI + {T2 - Ti) B2 T +

+ (Tn - Tn_1) B'
n Tn_ + B t Tn

where

R = radiance above atmosphere at a particular wavelength, X

Bn blackbody radiance of nth layer (starting at the top

atmosphere layer)

B= blackbody radiance of target

T=
n transmission from space through nth layer.

NO ABSORPTION

T2

IF
• Tn I

Figure A*VIII-3,- Upwelling radiance calculation.

The radiance at the spacecraft was then adjusted for the

thermal-band §pectal sensitivity by the technique described for

bands 1 through 12 i Section I of this appendix.

The radiances of the two temperature calibration sources were

also adjusted for thp thertial-bhrid spectral sensitivity. These two

calibration source radiances plotted versus their respective S192

output counts defined a gtraight line that was the thermaL-band

calibration Rune. The deviatiotn from this calibration line of the

point resulting from the radiance computed above and the correspond­
ing observed 9152 output counts was used as a measure of the
accuracy of the calibrationi.

NASA-JSC

III-A-38