| Abstract | The high resolution and wide look angle coverage characteristics of airborne synthetic aperture radar (SAR) and multispectral scanner (MSS) sensors introduce image processing problems not encountered in
the case of present satellite (LANDSAT) data. For the past 3 years, a significant effort has been directed toward the development and implementation of suitable airborne image analysis algorithms on the CCRSA Image Analysis System (CIAS). This paper
describes the principal components of these software packages. The radiometric and geometric fidelity of both airborne MSS and SAR imagery are known to be sensitive to topographic relief. A series of programs have been written to utilize terrain
information in the form of digital terrain models (DTMS). These tasks include conversion of DTMS to CIAS image format, model resampling, and generation of slope and aspect models from digital elevation models (DEMS). DTMS have been integrated into
many areas of image preprocessing and processing. These include: 1) rectification of imagery to UTM map grid using flight modelling techniques, 2) radiometric correction of rectified imagery for slope-aspect
effects, 3) identification of problem shadow and layover areas on SAR imagery, 4) image classification in which elevation, slope and aspect are treated as additional
features, 5) ground surface area estimation taking into account terrain slope. Finally, high resolution airborne imagery exhibits significant texture. The CIAS now has the capability to generate spatial features, derived from a number
of texture measures, and to analyze these features in conjunction with spectral information. |