4.0 IMAGE PROCESSING FUNCTIONS, TECHNIQUES USED AND SYSTEM
CONFIGURATIONS
4.1 Image Processing Functions 4.
Data acquisition, ground truth measurements, data processing
and data management are some of the important aspects for effective
utilization of remote sensing technology. The image processing
functions can be broadly categorised as:
- data format manipulations
- radiometric restorations
- image enhancement
- image compression (de-compression)
- pattern recognition
- image annotation
- image computations
4.2 Techniques
To meet the image processing functions described just now, a
number of techniques have been developed. Until recently, these
techniques used either complex analog circuits or off-line digital
computers. However, it is now practical to do the processing in real
time using video-rate image processing system. Some of these tech-
niques are described below:
(i) Shading correction for opto-electronics devices and sensors.
(ii) Image manipulation such as addition, subtraction, multiplica -
tion, inversion, rotation, cycling, zooming etc.
(iii) Contrast modifications in linear, logarithmic,
exponential, and non-linear form. 4,1
(iv) Histogram generation, equalization, level slicing, colour
encoding, pseudo-colour generation.
(v) Transform coding techniques such as Walsh Hadamard slant and
others,
(vi) Convolutions Kernels, spatial filtering, Gaussian and Lapalacian,
logarithmic filtering, image smoothing, image band width
compression and de-compression.
(vii) Correlation and auto-correlation functions.
(vii) Graphic generation, gridding, labeling, outlining, generating
arbitrary shaped images, image mixing, splits, overlay and
annotation.
(ix) Image parameter computations and measurement.
4.3 System Configurations
A number of image processing systems are now commercially
available which can carpy out a combination of these techniques on real
time basis. A typical video-rate image processing system which carries
out a variety of image processing functions will consist of the following
110
EN EX A — 172
md RR S ee sat Le
sue