Invited Paper

SAAB IRST, the system and flight trials

Ingmar Andersson*, Leif Haglund**
Saab Bofors Dynamics AB, Development and Technology, Sensors, Linköping, Sweden

ABSTRACT

Saab Bofors Dynamics has developed an IRST-system (Infra Red Search and Track) named IR-OTIS (Optical Tracking
and Identification System) and flight trials have been carried out with the system mounted on a Saab JA37 Viggen
fighter aircraft. This paper consists of three major parts. First an overview of Saab’s IRST-programs. The second part
describes the system ( IR-OTIS(Viggen) ) that made flight trials during 1998 and 1999 and finally a report from the
flight trials.

IR-OTIS has mainly three operating modes:

1) IRST-mode where the system covers several different FOS (Field Of Search).
2) FLIR-mode (Forward Looking IR) where the systems LOS (Line Of Sight) is directed from the aircraft.
3) Track-mode where the built-in-tracker controls the LOS.
It is also possible to switch from IRST-mode to track-mode automatically.

Fig. 1: Artists view of IR-OTIS operating modes.

*
[email protected]; phone +46 13 186536; fax +46 13 186107; http://www.saab.se; RTSLI-IA, SE-581 88
Linköping, Sweden. ** [email protected]; phone +46 13 186253; fax +46 13 186107; http://www.saab.se; RTSL,
SE-581 88 Linköping, Sweden.

Infrared Technology and Applications XXVIII, Bjørn F. Andresen, Gabor F. Fulop, Marija Strojnik, 69
Editors, Proceedings of SPIE Vol. 4820 (2003) © 2003 SPIE · 0277/786X/03/$15.00

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 Physically the IR-OTIS(Viggen) consists of the SU (Sensor Unit) and the SPU (Signal Processing Unit).
The SU is operating in the longwave IR-band with a 288*4 detector. In all modes the Sensor Unit generates images in
25 Hz and it is also possible to choose one of three FOV.
The SPU consists of a Saab designed image processing hardware and several DSPs. Functions in the SPU includes a
scene-based NUC (Non Uniformity Correction), anti-Narcissus, a point-target detector including estimation of SNR and
a clutter classifier for CFAR, target association, a correlation target tracker and an AGC for image presentation.

We carried out over 50 flight trials during 1998 and 1999 in three different rounds. The functionality of the system has
increased during the rounds and at the end of the trials all major goals were achieved.

Keywords: Saab, IRST, IR, image processing, flight trials, IR-OTIS

1. INTRODUCTION AND BACKGROUND

Saab Bofors Dynamics has a very long experience in optronic systems and image processing. During the last 40 years
several products lines has been developed and delivered. Typical systems are target trackers using TV and/or IR
sensors, some systems includes a laser range finder. Other products are IR-seekers for missiles.
Below are the most relevant IRST related projects briefly described.

1.1 FUMO

1987 – 1993. A predecessor of the IRST-programs is the FUMO. FUMO is a FUnctional MOdel of a missile IR-seeker.
The program was running at Saab in the end of 1980´s and beginning of 1990´s. The detector was a 1-d detector with 64
elements working in 3-5 um, scanning in both directions.
The image processing system had a search mode for automatic location of targets in the FOV and a track mode for
tracking targets and controlling the platform from the tracker.
A missile IR-target-seeker and an IRST have much functionality in common, regarding both the sensor-unit and the
signal processing system. During the flight trials the FUMO-sensor was installed in the aircraft just in front of the pilot.
This installation made the FUMO-trials look very much like an IRST-installation. A lot of flight trials were made with
this installation and with a fighter as target.

1.2 TV-OTIS

1990 - 1993. Another prerunner of the IRST programs was TV-OTIS. The system was built around a TV-CCD-sensor
including a 10 times optical zoom and a correlation target-tracker. The system had no automatic lock-on function. The
pilot controlled the platform by centering the target in the image and then manually lock-on to the target. The image
processing system consisted of a correlation tracker and a video output with symbols to the pilot. Flight trials were
made with TV-OTIS installed in a Saab 37 Viggen fighter aircraft in 1992-93. The results were very good which
encouraged us to develop a real IRST with IR-sensor and scanning mechanism for large search-areas and a new image
processing system.

1.3 IR-OTIS FOR SAAB VIGGEN FIGHTER

1994 - 2000. IR-OTIS(Viggen) is a complete IRST-system which made flight trials 1998 and 1999. The system is
developed to fit in a Saab Viggen fighter. Section 2 describes the system and section 3 the flight trials in detail.

1.4 IR-OTIS FOR SAAB GRIPEN FIGHTER

1997 -> . IR-OTIS(Gripen) is a complete IRST-system which is developed to suite in a Saab Gripen fighter. The
detector is mainly the same as IR-OTIS(Viggen) which is a 288*4 LWIR from Sofradir. Several improvements has
been made compared to IR-OTIS(Viggen) such as:
• The system works in 60 Hz.
• The sensor-unit has the capability to rotate 360 degrees in azimuth.
• The signal-processing unit has new microprocessors and electronics with much higher processing capacity.
• The connection between the sensor-unit and signal-processing-unit consists of two optical fibers.
Today no flight trials are made but it is possible to make flight trials in Saab Gripen in the future.

70 Proc. of SPIE Vol. 4820

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 1.5 IR-OTIS FOR GROUND SYSTEM

2000 -> . IR-OTIS(Ground) is a complete IRST-system which is developed to suite in ground systems such as tanks.
This is an ongoing program and system tests are planned to start this year (2002), installation in vehicle 2003 and real
tests in 2004. Basically it is the same sensor unit and signal processing unit as IR-OTIS(Gripen). Some changes are to
be made especially in the image processing system for target detection and man-machine-interface.

1.6 OPTRONIC SYSTEM SIMULATOR FOR GRIPEN FLIGHT SIMULATOR

2000 -> . A program is running for making a simulator for optronic systems for the Gripen flight simulator. The
simulator is scheduled to be installed in 2003 and should be able to simulate IRST, LDP (Laser Designator Pod) and
MAWS (Missile Approach Warning System).

1.7 IRS-725

1990 – 1998. A land-based IRST called IRS-725 for detection of airborne targets was developed and produced at Saab
Bofors Dynamics. See ref. 2.

1.8 OTHERS

Studies are made in several different areas for IRST-like applications such as IRST for ship installation and possibilities
to develop a modular IRST for different applications.

2. IR-OTIS ( VIGGEN), THE SYSTEM

2.1 INTRODUCTION

The development of IR-OTIS started in 1994 with a preliminary definition phase. That included inquires a choice of
LWIR or MWIR , scanning/staring system and choice of detector. The only possible fighter to install the system in was
a Saab Viggen fighter aircraft, which gave a lot of conditions of the physical installation.

The system basically consists of two units, the Sensor Unit and the Signal
Processing Unit. The Sensor Unit is positioned in the fuselage in front of
the canopy as shown in the figure to the right. The Signal Processing Unit
can be positioned in a suitable place several meters from the Sensor Unit.
For presentation of images, handling of target data and controlling modes
existing displays and controls are used. The Saab IR-OTIS communicates
with the aircraft system computer by means of the 1553 data interface.

For controlling the system in the fighter it was decided to use the same solution as the earlier flight tests with TV-OTIS.
This solution was to “borrow” the radar-control-lever with its buttons. The lever is connected to the aircraft computer
and the computer to the signal processing unit in IR-OTIS(Viggen). The pilot is for example able to choose operating
mode of the system, mode-changes, field of view, search-area, etc.

The presentation system for the demonstrator was also decided to be the same as the earlier flight tests with TV-OTIS.
The solution was to use a small TV-monitor, mounted just to the right of the center in the cockpit. The monitor
presented a CCIR-video from the signal processing unit. The image on the monitor was different depending on mode of
the IRST-system. In FLIR-mode and continuous tracking it was the normal FLIR-image and in search mode it was a
compressed image of the whole search area together with 1-4 magnifications of targets. In all modes were also
numbers, text and symbols written in the image. The signal processing unit also communicated target data to the aircraft
computer digitally for further presentation in the normal fighter avionics.

Proc. of SPIE Vol. 4820 71

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 2.2 SENSOR UNIT (SU)

The Sensor Unit comprises two functions: a thermal imager and a stabilized gimbal system.

The thermal imager is of the second generation, i.e. it is based on a 288*4 detector and it operates in long-wave
infrared. It includes detector, cooler, focusing optics, scanning optics, optics for image rotation and devices for
correction of detector non-uniformity. The thermal imager is designed and built as a self-contained module with well-
defined interfaces to adjacent systems such as the gimbal system and the signal processing.

On the Viggen aircraft used for the flight trials, the sensor unit had to be installed in a very narrow space envelope, i.e.
maximum 30 liters. To meet this requirement, the gimbal system was designed as an integrated part of the telescope.
The telescope also includes a module for field of view selection and a servo system controlling the spherical dome.

To create a compact gimbal system with a large field of regard without compromising the optical entrance aperture, a
novel technique was developed that minimized the number of optical components on the stabilized gimbal system. This
reduced the requirements for the servo components, such as the torque motors and the resolvers. The external diameter
of the spherical head is 180 mm and, as is shown in the figure below, the volume is well occupied. The figure also
shows the large ratio between the diameters of the entrance lens and the head.

Fig. 2: Spherical head of Saab Bofors Dynamics IRST.

The gimbal system had stabilization around two axes: pitch and yaw. Stabilization is achieved by feedback of calculated
angular velocities. The aircraft angular velocity is measured with a gyro in the sensor unit. The optical principle of the
gimbal system results in image rotation. This is compensated for with a Pechan prism in the thermal imager. Angular
stabilization better than 50 µrad was accomplished.

With a high performance optical system integrated in a narrow space envelope with power consuming devices such as
power modules, electronic boards and torque motors, special attention was given to the thermal situation of the sensor
unit. The total weight of the sensor unit was not allowed to exceed 30 kg. As a consequence the mechanical structures
had to be “lean”, their material carefully selected and their levels of stress and deformation thoroughly assessed in order
not to compromise the optical performance. The figure below is taken from Saab Bofors Dynamics mechanical CAD
software and gives an impression of the highly compact design.

72 Proc. of SPIE Vol. 4820

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 Fig. 3: Mechanical design (CAD) of the sensor unit.

The sensor unit was functionally controlled by the signal processing unit. The latter commands operational modes,
search patterns, fields of view, etc. The electronics and software of the sensor unit then controls the electro-mechanical
servo systems for the focusing, the gimbal system etc.

The sensor unit was supplied with electrical power and cooling air from the aircraft’s standard supplies. The cooling air
had to be filtered to protect the sensitive electronics and optics in the sensor unit. The mechanical interface to the
aircraft was the “roof” structure of the sensor unit. This structure was adjusted in shape and material to the hull of the
Viggen aircraft.

2.3 SIGNAL PROCESSING UNIT (SPU)

The signal-processing unit automatically detects and tracks targets, sends target positions and an IR-video to the
aircraft. Here the functions will be descriptions of the sub-blocks in figure 4 and the result for the total system from
flight trials will be given in section 3.

Physically the SPU had to fit in a tight volume in the Viggen aircraft and the prerequisite was that the weight should be
less than 10 kgs in a volume not exceeding 10 liters.
Advanced signal and image processing hardware and software is used including image memories for image processing
and presentation with capacity to store several images. The SPU hardware was designed in the mid-nineties and based
on several Texas InstrumentsTMS320C40 DSP:s and ASIC:s developed at Saab for fast pixelwise operations. The SPU
is also equipped with necessary Programmable Read Only Memories (PROM) and Random Access Memories (RAM)
for program storage and data handling. The PCB are connected with fast image buses. Computing intensive filters,
mainly 2D-convolutions, and algorithms are implemented in hardware. There is also provided a possibility to register
images and other data to harddisk during flight. The registered data can be used for post flight simulations.

Proc. of SPIE Vol. 4820 73

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 Registration

Command
and
Control
Scanning, LOS Image Presentation

Sensor signal processing

IR-video

CFAR detector Evaluation

Correlation Association
tracker tracker

Fig. 4: Blockscheme of the functions in the SPU.

2.4.1 Command and Control

The software block “command and control” handles the internal states of the IR-OTIS and also serves as the interface
between IR-OTIS and the aircraft. Typical commands are change of field-of-view, modify focus, change mode, i.e.
IRST inclusive search area or FLIR, and setting of search area. The most common command is to control the line-of-
sight (LOS) of the platform, in both FLIR mode and in IRST mode. The LOS can be controlled by the aircraft velocity
vector, by the radar, manually from the pilot or in a closed loop fashion from the image processing. The guidance of the
platform, for different search areas is also included as a function in this block.

2.4.2 Sensor signal processing

The first step in the image processing chain is to compensate for some degradations in the images. This means
a) compensation of Narcissus and of image modulation, e.g. cos4-modulation.
b) scene based non-uniformity correction (NUC), in our case meaning delining.
c) image integration in FLIR mode to increase Signal-to-Noise Ratio.
Of these corrections a) is mostly done for image presentation, while b) and c) also increase detection probability

2.4.3 CFAR detector

The CFAR detector works in pyramid structure, here a Haar wavelet is used, where signal-to-noise ratios (SNR) are
estimated in each pixel. The noise is estimated using an estimate of local standard deviation combined with an estimate
of the local background correlation. The local correlation is also used to divide the image in two different parts, one for
“bluesky” and one for “cluttered” backgrounds. The signal is estimated from one of two matched, using Wiener theory,
filters for point targets. One filter is used in bluesky backgrounds and another one in areas classified as clutter. The

74 Proc. of SPIE Vol. 4820

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 SNR is the estimated using the quotient of the estimated signal (from the filter) and the estimated noise (from the
standard deviation). This procedure is done for all scales, or pyramid levels, and it can be seen as if larger targets are
shrinked into a fixed sized matched filter. For more information of the ideas as well as some further work see reference
3.
The sixty-four strongest alarms per image are sent to the evaluation block for further processing. To be an alarm the
SNR has to be a local maximum in its own scale.

2.4.4 Evaluation
The first step in the evaluation was to remove the effect of multiple alarms in scale originating from the same signal.
The idea being that one physical target should only give rise to one alarm. After this reduction of the number of alarms
the evaluation block is used to estimate some more attributes. In the flight trials only size and SNR were used.
From each image sixteen alarms are sent to the association tracker.

2.4.5 Association tracker

The main activity for the association tracker is to combine the detections from a single image in time to see if they
behave as a possible target. The implemented association tracker is a multiple hypothesis tracker working in a stabilized
coordinate system. Each pixel in the image is given a corresponding direction in space from the gyros and angle
resolvers and the tracker uses these angular data for its calculation. The estimated size from the evaluation is also used
in the association process.
This tracker can be employed in both FLIR and IRST modes, i.e. “track while scan” is implemented in the IRST mode.
The tracker has been implemented in a very general manner, being able to track in FLIR mode with 25 Hz image rate as
well as in IRST mode with maybe 0,5 Hz update rate. In the trials alarms from the ground were suppressed based on
their kinematical behavior, as we know the trajectory, velocity and height of the plane.

2.4.6 Correlation tracker

In the case of tracking in FLIR mode Saab has a long tradition using correlation tracking. It has turned out to be a very
stable tracker if the target is more than a few pixel in size and in IR-OTIS it is used when the platform is controlled by
the image processing function. The correlation is made on prefiltered images, where the edge information is coded for
stable tracking. The reference image is update to take target maneuvers into account and the reference image is
correlated with the video stream for accurate tracking. The correlation tracker can be locked from the pilot on an
arbitrary contrast but more common is that a target is tracked in IRST mode and the mode switches to tracking a
prioritized target via a command from the pilot.

2.4.7 Image presentation

One important feature of the IR-OTIS is, of course, to present an image for the pilot. The CCIR video used for this is
used in different ways in FLIR and IRST mode. In Fig. 5 an example of an image in FLIR mode is given,. The cross in
the image means that tracking is established on this particular target. An example of the image presentation in IRST
mode is given in Fig. 6. The upper half of the figure is used for presenting all images of the search area, in this case 60
images giving a total of 120ox6o, compressed to fit the size of the monitor. The lower half is used for getting an
overview of the where the platform is directed and what search area that is used. Other markers and text information
giving the internal state of the sensor unit and status of the image processing functions are also given.
Typically the digital IR-video uses 12 bits, while the monitor maximally can use 7 bits. In Fig. 7 an example of this
dynamic compression is given. The challenge is to maintain information from the background to get a good overview
and at the same time not lose small interesting contrasts.

Proc. of SPIE Vol. 4820 75

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 Fig. 5: Image presentation in FLIR mode.

All images in
search area

Target candidates

Enlarged Search area

candidates relative to
flight direction

Fig. 6: Image presentation in IRST mode.

76 Proc. of SPIE Vol. 4820

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 Fig 7: Different ways of displaying IR images. Left conventional compression of dynamical range.
Right Combination of compression of dynamical range and high/bandpass enhancement.

2.4 REGISTRATION UNIT (RU)

During the flight tests we had a registration unit connected to the SPU. Physically the registration unit consisted of a
disc system storing digital images and data and an analog video recorder (SVHS-video) recording the same image that
was presented to the pilot during the flight.
The SVHS-recorder was normally able to record the whole flight including cursors and some text/numbers in the image.
Also the conversion to/from the pilot was recorded in the same recorder.
The disc system was able to store 27 minutes of images and data with no compression. Normally we used a lossless
compression scheme, based on a Lempel-Ziv-Welsh algorithm, to increase the registration capability. Normally this
gave us between 40 and 45 minutes registration time from each trial, which was enough to capture the interesting parts.
Together with every image also data from the SU, e.g. platform angles, the SPU and the fighter were recorded. By
recording all data we ensured the possibility to do post simulation from the trials.
Between the flight tests the SVHS-cassette was changed and the RU-discs were copied to another disc and tranported to
Saab for evaluation.
The Viggen fighter also had its own registration system, which also was utilized during the evaluation of the IR-OTIS
flight tests.

3. IR-OTIS(VIGGEN), FLIGHT TRIALS

Fig. 8: IR-OTIS(Viggen) mounted in aircraft in flight trials.

Proc. of SPIE Vol. 4820 77

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 3.1 INTRODUCTION

The flight trials were made in three different rounds. The first round took place during winter 1998, the second in the
summer/fall the same year and the third and lasts in the winter and spring 1999. Between each round the system was
uninstalled and transported to Saab for renovation work during a period of 3-4 months. The work consisted of
enhancements in the Sensor Unit and in the Signal Processing Unit and finally one month for systems acceptance and
re-installation in the aircraft.

Totally IR-OTIS(Viggen) has carried out over 50 flights, which could be split in two parts. One part is the planned trials
and a second part for others. Others could be for example the first flight in each round to find out that everything
worked, educating a new pilots or when the fighter, on which IR-OTIS was mounted, was used as a target aircraft in
flight tests. The planned tests were 5 in the first round, 9 in the second round and 13 in the last round.

The target in all tests was a Saab Viggen fighter (the same type as IR-OTIS was mounted in) which is a single engine
fighter with double delta wing and about 16 000 kg weight. The size of the target is approximately 60, 30 and 6 m2
from resp. aspects above, side and front.

After each flight a quick evaluation was made to check that everything had worked as planned. At the end of each round
all different parts and functions in the system were evaluated and decisions were made of which parts and functions
should be improved to next round.

Below are some figures from evaluations in flight trial 21:

IR-OTIS

Target

T ro
58, N ynäsh

N yköp
58, i

N o rrk ö
A rk ö s
58, S ö d e rk ö
L in k ö p
30

58,
Lat
G r
Å tv id a
58,
T a rg e t

K is
57,

57, V ä s te r
V is

H u lt s f
57,
15, 15, 15, 16, 16, 16, 16, 16, 17, 17, 17, 17, 17, 18, 18, 18, 18,
Long
d
Fig. 9: Trajectories in longitude and latitude of IR-OTIS and target during trial number 21.

78 Proc. of SPIE Vol. 4820

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 13000

12000 Target
11000

10000 IR-OTIS
9000

8000

Altitu
7000
de
(m) 6000
5000

4000

3000

2000

1000

0
12:00:00 12:15:00 12:30:00 12:45:00 13:00:00 13:15:00 13:30:00

UTC-time

Fig. 10: Altitude (m).

1,1

1,0 Target

0,9

0,8

0,7
IR-OTIS
Ma0,6
ch
0,5

0,4

0,3

0,2

0,1

0,0
12:00:00 12:15:00 12:30:00 12:45:00 13:00:00 13:15:00 13:30:00

UTC-time

Fig. 11:Speed (Mach).

Proc. of SPIE Vol. 4820 79

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 180

170

160

150

140

130

120

110

100

90

80

70

60

50
Distance

40

30

20

10

0
12:00:00 12:15:00 12:30:00 12:45:00 13:00:00 13:15:00
UTC-tid

Fig. 12: Distance to target (km).

3.2 FIRST ROUND

The first round took place Jan 1998 to Feb 1998 and 5 real flight trials were made.

Narrow Field Of View (NFOV) was not in place but Wide and Medium Field of View (WFOV, MFOV) worked as
planned in both FLIR-mode and search-mode as well as changes between them. Possible search-areas were in MFOV
80*15 and 120*15 degrees and in WFOV 120*30 degrees.

The flight trials were at different altitudes and with one target.

Acquisition ranges achieved in this round were 36 km in tail-on and 17 km in head-on with MFOV.

3.3 SECOND ROUND

The second round took place between June 1998 and Sept 1998 and 9 real flight trials were made.

Before the second round several improvements was made. NFOV were put in place and a closed loop tracker was
running in the Signal Processing System. The image quality was enhanced since a lot of “electronic noise” was removed
from the image. The image presentation was also improved.

All major functions worked as intended. Several different flight tests were made with different altitudes, different target
speed, different target backgrounds such as clouds, blue-sky and ground. Tests of IR-OTIS detection range were made,
mostly head-on and tail-on. Tests were also carried out to change mode from search mode to continuos tracking. Tests
were made with a rolling target and with our own fighter rolling. Tests of tracking ground-objects such as houses, ships,
islands and bridges were made.

Best acquisition ranges achieved in this round were 69 km in tail-on and 34 km in head-on with NFOV.

80 Proc. of SPIE Vol. 4820

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 3.4 THIRD ROUND

The third round took place between March 1999 and May 1999 and 13 real flight trials were made.

Before the third round further improvements were made. The image quality was even better and improvements were
made in the image processing. The possibilities to focus in different temperatures were improved. Some improvements
were made in the mechanical system.

The flight tests were about the same as round two except that we had two targets in some trials for multi-target-tracking,
sometimes with crossing paths, trials in the dark and trials for evaluating passive distance measurements.

Best acquisition ranges achieved in this round were 73 km in tail-on and 50 km in head-on with NFOV.

3.5 SUMMARY FLIGHT TRIALS, DETECTION RANGE

Target- Range
Altitude km Calculated
Trial-nr Mode speed [km]
IR-OTIS Target range [km]
Mach in trials
FLIR/
12 7,0 8,0 0,98 34 29
NFOV
IRST/
14 7,0 8,0 0,85 20 24
NFOV
IRST/
14 7,0 8,0 1,01 23 24
NFOV
1)
FLIR/ 50
21 11,2 11,5 0,90 Detection at 39
NFOV 65
1)

FLIR/
21 11,2 11,5 0,92 35 39
NFOV
FLIR/ 1)
21 11,2 11,5 0,93 27 17
MFOV
FLIR/
21 11,2 11,5 0,84 22 17
MFOV
IRST/ 2)
24 11,2 11,5 0,81 17 28
NFOV
FLIR/ 3)
27 10 10 ca 1,3 (90) 58
NFOV
Trials during 1998

Fig. 13: Detection ranges in flight trial ,target Viggen, head on.
1) COTRA influences the result.

2) Short range because target was in a “bad” part of the detector.

3) 90 km is valid during real time simulation after trials with enhanced image processing and off-line improved images.
The target is not in FOV during trials with distances between 30 and 70 km.

The table “calculated range” was made during flight trials and is made for a sensor as close the demonstrator as
possible. Today better comparison between ranges in flight trials and calculated ranges are made, see ref 1).

Proc. of SPIE Vol. 4820 81

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 Target- Range
Altitude km Calculated
Trial-nr Mode speed [km]
IR-OTIS Target range [km]
Mach in trials
FLIR/
06 5,0 7,0 0,8-0,9 68 64
NFOV
FLIR/
06 10,0 12,0 0,8-0,9 60 66
NFOV
FLIR/ 3)
12 7,0 8,0 0,93 69 (114) 70
NFOV
IRST/
14 7,0 8,0 1,05 56 59
NFOV
IRST/
14 7,0 8,0 0,87 61 59
NFOV
FLIR/ 1)
20 3,0 2,0 0,8-0,9 56 74
NFOV
FLIR/
20 3,0 2,0 0,8 -0,9 43 43
MFOV
2)
FLIR/ 72/73
25 8,0 9,0 0,8-0,9 109
NFOV (Two targets)
Trials during 1998.

Fig. 14: Detection ranges in flight trials,target Viggen, tail on.

1) Bad focus and difficult background.

2) Clouds between target and IR-OTIS.

3) 114 km is valid during real time simulation after trials with enhanced image processing and off-line improved
images.

4. IR-OTIS(VIGGEN), POST FLIGHT TESTS

The system is currently running at Saab Bofors Dynamics and has been used, and is still used, for several different tests
and post flight simulations.

4.1 POST FLIGHT, REAL TIME SIMULATIONS

It is possible to do post flight simulations in real time of the IR-OTIS. Instead of the Sensor Unit and an aircraft we can
use the recorded digital images and data on the RU disc system. It is possible to reverse the image and data-flow in the
disc system and send the images/data to the Signal Processing Unit. An identical SPU as the one used in the flight trials
receives images/data in the same way the images/data is sent from the Sensor Unit and aircraft during flight trials.
These simulations are made in real time. It is possible to remove noise from images off-line and put them back into the
Registration Unit to simulate a better sensor. Tests for improving the image processing and get better results, e.g longer
acquisition ranges, were also carried out.

For example in trial number 12 above, the detection range was 69 km (tail-on). With reduction of noise in the images
off-line we improved the detection range to 88 km. With improvements in the Signal Processing Unit, in this case the
temporal integration made most of the improvement, we achieved 114 km detection range and we had target
indications in the image as far as 127 km.

The possibility to do real time simulations is a very useful feature in our system and makes it possible to use every trial
a lot of times with different parameters.

82 Proc. of SPIE Vol. 4820

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms

 4.2 TRIALS FROM STATIONARY PLATFORM

Since the last flight trial in May 1999, IR-OTIS is running on a roof on one of our buildings in Linköping. Two major
improvements have been made:
• It is possible to do a temporal integration of images. The integration time can be automatically controlled from the
rotation of LOS. The integration time is set to different values or would be turned off if LOS rotates a lot.
• It is possible to move LOS 1/4, 1/5, 1/8 or whatever part of Field Of View between every image in search-mode.
This makes it possible to get 8 indications in a row but longer time to next (8) indication(s) if the gimbals system
mows LOS 1/8 of FOV.

It is also possible to record images/data with the RU from the roof and use them in ongoing projects.

5. FUTURE WORK
The work at Saab Bofors Dynamics in the nearest future that is directly connected to IRST is the ongoing projects
described in Chapter 2.5 and 2.6.

The IR-OTIS(Viggen) and IR-OTIS(Ground) is planned to run trials/recordings from our roof-laboratory during 2002
and filed tests with IR-OTIS(Ground) is planned to 2004.

ACKNOWLEDGEMENTS
For their support and valuable assistance during flight trials the following are acknowledged:
Mr. Bo Ohlsson at Swedish Defence Materiel Administration, Avionics dept.
Mj. Jerry Pousette at Swedish Defence Materiel Administration, Flight trials dept.

REFERENCES
1. Åsa Engvall, “ Airborne IRST target range – modeling and flight trials ”, Proceedings of SPIE, vol. 4820-09, 2002,
to be published.
2. Hans Samuelsson, “Infrared surveillance system for surface – to – air defence”. Proceedings of SPIE, Infrared
Technology XVIII, vol 1762-17 , p. 146 (1992).
3. Ulf Gustafsson, Leif Haglund and Ingmar Andersson, “Target detection and background suppression using
quadrature pyramid filters”, Proceddings of The 12th Scandinavian Conference on Image Analysis, pp 553-560,
June 2001, Norway.

Proc. of SPIE Vol. 4820 83

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/17/2013 Terms of Use: http://spiedl.org/terms