Mg.01.

1,17/01/17

Materi kuliah GD3203 Fotogrametri II, Semester II-2016/2017

Chapter 01a

Fotogrametri Dijital
Compiled & developed by
Saptomo H Mertotaroeno, Ir., M.Sc.
[email protected]

KK Inderaja & Sains Informasi Geografis

Fakultas Ilmu dan Teknologi Kebumian
Institut Teknologi Bandung
 Daftar Isi
• Pendahuluan
• Digital Images Chapter 01a
• Digital Image Processing
• Digital Image Resampling
• Digital Image Compression
• Digital Image Measurements
• Computer Vision
• Digital Image Simulation : Computer Graphics
• Problems / Exercises
• References
• Tugas No

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 2

  Fotogrametri Dijital
 Generic Digital Photogrammetry Environment
 Relationship of Digital Photogrammetry to Other Disciplines
 The Advantages of Digital Data

PENDAHULUAN

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 3

 Fotogrametri Dijital (1)
Fotogrametri Dijital = Digital Photogrammetry = Softcopy
Photogrammetry.
 Citra masukan :
• citra dijital (fotogrametri dijital);
• citra analog / diapositif (fotogrametri analitis, dan analog)
 Ditinjau dari segi geometri :
• Fotogrametri Dijital = Fotogrametri Analitis + Pengolahan Citra
Dijital.
• Geometri proyektif yang merupakan dasar fotogrametri adalah
sama untuk aplikasi citra analog atau dijital.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 4

 Fotogrametri Dijital (2)
Data dalam bentuk Digital (dari Sensor optik atau elektronik : Kamera Digital
atau Scanner, bahkan dari Video Digital)
Data diproses dan dianalisis melalui sebuah komputer. ()
Proses analisis memerlukan adaptasi dari berbagai disiplin ilmu antara lain :
 Pemrosesan Citra (Image Processing), seperti kompresi (Compression), penajaman
citra (enhancement)
 Pengenalan Pola (Pattern Recognition), seperti image understanding dan computer
vision  Artificial Intelligence. ()

Perkembangan fotogrametri dijital tergantung pada perkem-

bangan :
 perangkat keras komputer,
 sistem pencitraan dijital (kamera/sensor) dan scanner, dan
 algoritme komputer (perangkat lunak).

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 5

Generic Digital Photogrammetry Environment

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 6

 Relationship of Digital Photogrammetry
to Other Disciplines
(concerned with improving
pictorial information for hu-
man interpretation, and pro-
cessing of images for auto-
nomous machine perception)

Pattern recognition
(e.g., automatically locate
the fiducial marks)

(begins with a description of the (e.g., automatically locate

scene and produces an image, buildings, fences, or roads)
e.g., rendering in CAD systems)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 7

Penggunaan citra dalam bentuk dijital akan menawarkan beragam ke-
untungan dalam pengolahan dan otomasi :
 pengolahan citra dijital dapat digunakan untuk mengubah citra dalam cara
yang sangat presisi dan membuatnya menjadi lebih berguna dan lebih
lentur (fleksibel),
 banyak operasi manual dapat dikerjakan secara otomatis oleh komputer (+
perangkat lunak).

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 8

 • Digital Images
• Digital Numbers, Bits, and Bytes
• Discrete Sampling, and Spatial Resolution
• Level of Quantization
• Spectral Resolution
• Digital Image vs Photo Coordinate System
• Digital Images Summary

DIGITAL IMAGES

Sudah di bahas di GD3103 Fotogrametri I. Harap dipelajari ulang sendiri.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 9

 Digital Images (1)
A digital image is : (definition)
 a computer-compatible pictorial rendition in which the image is divided
into a fine grid of “picture elements”, or pixels.
 a rectangular array of pixels in which the brightness of a scene at each
discrete location has been quantified.
 a regular array of pixels, and is described in terms of its geometry and its
radiometry.

The image in fact consists of an array of integers, often referred to as

digital numbers, each quantifying the gray level, or degree of
darkness, at a particular elements.
When an output image consisting of many thousands or millions of
these pixels is viewed, the appearance is that of a continuous-tone
picture.
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 10
 Digital Images (2)
Each pixel of a digital image is a direct or indirect sample of the
continuous incident light distribution.
 A digital sensor samples incident light directly.
 Scanning a piece of film and converting it into a digital image is
indirect sampling.

Pixels may be arranged in any regular pattern, but in practice

most sampling grids are square or rectangular.
 Hexagonal grids have some theoretical advantages but have seldom
been implemented.
 The ratio between pixel width and pixel height is known as the
aspect ratio.
• A square pixel has an aspect ratio of one.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 11

 (a) The principle of
digital images. This
image has been
divided into a pixel
grid of 72 rows and
72 columns, with
each pixel being
represented by a
value from 0 (dark
black) to 255 (bright
white).
A portion of the
image near the
mouth of the statue
is shown enlarged in
Low numbers (b), and
(dark areas), the digital numbers
and
associated with this
high numbers
(light areas). portion are listed in
(c) (c).
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 12
 Digital Numbers, Bits, and Bytes
• Since computers operate directly in the binary number system, it is
most efficient and convenient to use ranges corresponding to
powers of 2.
‐ Numbers in the range 0 to 255 can be accommodated by 1 byte, which
consist of 8 binary digits, or bits.
‐ An 8-bit value can store 28 (=256) values, which exactly matches the
range of 0 to 255 (remember to count the 0 as a value).
‐ The entire image of Figure (a) above would require a total of 72 X 72 =
5184 bytes of computer storage.
• Notes : 1 kilo bytes = 210 bytes = 1024 bytes = 1 k bytes,
1 mega bytes = 1024 k bytes = 10242 bytes = 1 M bytes,
1 giga bytes = 1024 M bytes = 10243 bytes = 1 G bytes, etc.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 13

Contoh : Bits, Bytes, dan Digital Numbers
Contoh Perhitungan dalam bit yang mempunyai basis bilangan 2
(dua) dengan 8 bits :

10101110
10101110 = 1 x 27 + 0 x 26 + 1 x 25 + 0 x 24 + 1 x 23 + 1 x 22 + 1 x 21 + 0 x 20
= 128 + 0 + 32 + 0 + 8 + 4 + 2 + 0
= 174

11111111
11111111 = 1 x 27 + 1 x 26 + 1 x 25 + 1 x 24 + 1 x 23 + 1 x 22 + 1 x 21 + 1 x 20
= 128 + 64 + 32 + 16 + 8 + 4 + 2 + 1
= 255
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 14
 Discrete Sampling & Spatial Resolution
• Digital images are produced through a process referred to as
discrete sampling.
‐ In this process, a small image areas (a pixel) is “sensed” to
determine the amount of electromagnetic energy given off by the
corresponding patch of the surface on the object.
‐ Discrete sampling of an image has two fundamental
characteristics, geometric resolution and radiometric resolution
(quantization and spectral resolution).
• Geometric (or spatial) resolution refers to the physical size of
an individual pixel, with smaller pixel sizes corresponding to
higher geometric resolution.
‐ Geometric resolution is important for feature recognition in
digital photographs.
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 15
The effect of
different geometric
resolutions on
image clarity.

(a) 72 X 72 pixel image

(b) 36 X 36 pixel image

(c) 18 X 18 pixel image

(d) 9 X 9 pixel image

Geometric resolution
is important for
feature recognition in
digital photographs.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 16

 Konsep GSD dan Pixel Size
IFOV is the angle determined by
the pixel size and the focal
length.
GSD is the projection of the pixel
size onto the ground plane.
GSD ≠ resolution.
The resolution of a digital
image is determined by both :
 the sensor geometry, and
 factors external to the sensor,
such as atmospheric conditions,
platform motion, etc.
In practice : resolution < GSD.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 17

 Level of Quantization
• Quantization refers to the conversion of the amplitude of the
original electromagnetic energy (analog signal) into a number
of discrete levels (digital signal). (analog to digital conversion)
‐ Greater levels of quantization result in more accurate digital
representations of the analog signal.
‐ 8-bit quantization means 256 gray levels, 10-bit (1024 gray
levels), 12-bit (4096 gray levels).

• Quantization level has a direct impact on the amount of

computer memory required to store the image.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 18

 Quantization of Analog Signal (1)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 19

 Quantization of Analog Signal (2)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 20

 The effect of
various levels of
quantization.

(a) 8-bit (256 levels)

(b) 3-bit (8 levels)

(c) 2-bit (4 levels)

(d) 1-bit (2 levels), binary

image.

Notice that in lower-level

quantization, large areas
appear homogeneous
and subtle tonal
variations can no longer
be detected.
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 21
 Spectral Resolution (1)
• The electromagnetic spectrum covers a continuous range of
wavelengths.
‐ In imaging, samples of electromagnetic energy are normally
sensed only within narrow bands in the spectrum.
• The natural color film samples three bands within the visible
portion of the spectrum―blue, green, and red.
‐ The spectral resolution of color film corresponds to three spectral
bands or channels.
• The black-and-white image has only one spectral band which
covers the entire visible range from violet through red.
• Other imaging devices may take samples in several (multi-
spectral), or even hundreds (hyper-spectral), of bands.
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 22
Portion of Electromagnetic Spectrum Useful
for Imaging

(film sensitivity (solid-state

sensors sensiti-
region) vity, up to 12
μm)

(UV region is
seldom used near infrared (NIR)
for P&RS)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 23

 Citra Dijital Berwarna (RGB)

Pixel (picture element)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 24

Perbedaan gray level dan berwarna

Gray level (0-255) = 1 byte

256 warna

Red (0-255) = 1 byte

Green (0-255) = 1 byte
Blue (0-255) = 1 byte

Total = 3 byte dengan

Warna 2563 = 16.777.216 warna
09/10/2024 03:26 AM 25
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 26
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 27
 Spectral Resolution (2)
• With a higher level of spectral resolution, a more accurate
representation of an object’s spectral response pattern can
be made.

• Since each spectral band must be stored individually in

computer memory, the number of bands has a direct impact
on the amount of image storage required.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 28

 Hyperspectral vs multispectral signatures

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 29

 Digital Images Exercise
• Example. A 3000-row by 3000-column satellite image has three
spectral channels. If each pixel is represented by 8 bits (1 byte) per
channel, how many bytes of computer memory are required to store
the image ? If this image is transferred over a computer network line
having a data transfer speed of 100,000 bytes per second (bytes/s),
how long will it take to transfer the image ?

• Solution.
Number of pixels per channel = 3000 X 3000 = 9,000,000 pixels
Bytes per pixel = (3 channels) X (1 byte) = 3 bytes
Image size = (9,000,000 pixels) X (3 bytes/pixel) = 27,000,000 bytes
Transfer time = (27,000,000 bytes)/(100,000 bytes/s) = 270 s, or 4.5 min

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 30

Pixel (Raster) Coordinate System (1)

A type of instrument coordinate

system.
Raster (matrix index) coordinate
system (rows, columns) or
(columns, rows).

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 31

Pixel (Raster) Coordinate System (2)

Sensor-element is, usually, a square shape.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 32

 Pixel vector coordinate (x, y) at the center of the pixel.
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 33
 Pixel vs Image Coordinate System (1)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 34

 Pixel
vs
Image
Coordinate
System
(2)

(Concept of resampling / interpolation of digital number)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 35

 Digital Image vs Photo Coordinate System (1)
(from scanner digitation)
Image coordinate sys-
tem of digital image
(rows, columns).
The shaded, rotated
square indicates a
digitized photograph
with x,y the photo-
coordinate axes.

Coordinate transfor-
mation : (image coor-
dinate system) ↔
(photo coordinate
system), plus resam-
pling.
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 36
 Digital Image vs Photo Coordinate System (2)
(from digital camera)
y0
y
y’

x’
x0 (photo coordinate system)
“right-handed coordinate system)”

x’ = y – y0
y’ = x0 – x
(translation from
Digital Image Coordinate System
x to
(digital image coordinate system) Photo Coordinate System)
“right-handed coordinate system”

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 37

 Digital Image vs Photo Coordinate System (3)
(from digital camera)
x0
x
y’

x’
y0 (photo coordinate system)
“right-handed coordinate system)”

x’ = x – x0
y’ = y0 – y
(translation from
Digital Image Coordinate System
y to
(digital image coordinate system) Photo Coordinate System)
“non-right-handed coordinate system”

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 38

 Digital Image vs Ground Coordinate System

(“non-right-handed”)

(“right-handed”)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 39

 Digital Images Summary
• Pixels : raster system, rows, columns, digital numbers/gray
level.
• 1 byte = 8 bits = range 0 (black) – 255 (white).
• Geometric/spatial resolution (the physical size of an
individual pixel) : result in geometric accuracy, important for
feature recognition.
• Radiometric resolution : level of quantization, and spectral
resolution.
• Level of quantization (8-, 10-, or 12-bit) : result in digital
representation accuracy of the analog signal.
• Spectral resolution (1-band/B&W, 3-bands/RGB, several-
bands/multispectral, hundreds of bands/hyperspectral) :
result in representation accuracy of an object’s spectral
response pattern.
• Problems : image storage, and image transfer time.
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 40
 DIGITAL IMAGE PROCESSING

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 41

 Digital Image Processing (1)
Definitions :
 The use of computers for manipulating digital images in order to
improve their quality and/or modify their appearance.
 To perform digital processing on the image, in order to modify or
emphasize certain aspects of the image.
• E.g., we may brighten the image to see more detail in shadows, or change
the frequency content to emphasize edges.

The process :
 The digital number of each pixel in an original image is input to a
computer, with its inherent row and column location.
 The computer operates on the digital number according to some
preselected mathematical function or functions.
 Stores the results in another array which represents the new or
modified image.
09/10/2024 03:26 AM Fotogrametri Dijital, SHM 42
 Digital Image Processing (2)
DIP is strictly concerned with the image values of each pixel.
 There is no consideration of the content of the image.

Radiometric DIP, in which individual pixel values are modified

based on the pixel value distribution across the image or within a
local neighborhood.
Spatial DIP, in which the values of certain neighboring pixels or
the pixel within the image is use in the procedure.
Spatial DIP ≠ geometric processing (the image geometry is
changed)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 43

 Tasks of Digital Image Processing
Compression, whereby the original image is converted to
another form needing less storage.
Enhancement and restoration to improve the quality of a noisy
image.
Segmentation and description, which entail
 conversion to parts or primitives,
 measurement of their properties, and
 description of the image in terms of primitives and their properties.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 44

 Pengolahanan Citra Digital

Visualization Acquisition
(Scanner & Digital Camera)

Enhancement &
Image Storage
Restoration Processing

Segmentation Compression
09/10/2024 03:26 AM 45
 Image Enhancement
(Radiometric Processing)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 46

 Image
Enhancement
Examples

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 47

 Image Enhancement Examples for Edge Detection
(Spatial Processing)

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 48

 PROBLEMS / EXERCISES

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 49

 Problems
2-30. A panchromatic (one-band) digital image has 16,000 rows and 16,000
columns. Assuming each pixel takes 1 byte, how many bytes are
required to store the image ?
2.31 A 6200-row by 6200-column digital image has six spectral channels.
Assuming each pixel requires 1 byte per channel, how many bytes are
required to store the image ?
6.1 A digital frame sensor has pixel dimensions of 0.010 mm by 0.010 mm
and a principal distance of 300 mm. What is the instantaneous field of
view (IFOV) in microradians ? What is the ground sample distance (GSD)
when imaging vertically at 3000 meters above terrain ?

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 50

 References
 McGlone, J.C., ed., 2004. Manual of Photogrammetry, 5th ed., American
Society of Photogrammetry and Remote Sensing, Maryland 20814, USA,
1151 p.
 Mikhail, E.M., J.S. Bethel, and J.C. McGlone, 2001. Introduction to Modern
Photogrammetry, John Wiley & Sons, Inc., USA, 479 p.
 Schenk, T., 1999. Digital Photogrammetry, vol. I, Background, Fundamentals,
Automatic Orientation Procedures, TerraScience, OH 43135, USA, 428 p.
 Wolf, P.R., and B.A. Dewitt, 2000. Elements of Photogrammetry : with Appli-
cation in GIS, 3rd ed., McGraw-Hill, New York, 608p.
 Internet browsing. (for illustration only).

Aknowledgements
Some data, information, illustrations, and diagrams in this presentation are quoted
from .ppt files developed by Dr. Ir. Bobby S Dipokusumo, M.Sc, and Dr. Ir. Andri
Hernandi, MT.

09/10/2024 03:26 AM Fotogrametri Dijital, SHM 51

Materi kuliah GD3203 Fotogrametri II, Semester II-2016/2017

Bersambung ke Chapter 01b

Fotogrametri Dijital
Compiled & developed by
Saptomo H Mertotaroeno, Ir., M.Sc.
[email protected]

KK Inderaja & Sains Informasi Geografis

Fakultas Ilmu dan Teknologi Kebumian
Institut Teknologi Bandung