DEVELOPMENT OF SOFTWARE FOR LAND COVER CLASSIFICATION OF REMOTELY SENSED DATA

Abstract

To keep pace with the explosive increase in the amount of inforMation that needs to be processed,,rapid proliferation of computers with newer and newer computing methods and technology is required. Pattern recognition is a major area of activity that encompasses the processing of pictorial information obtained from interaction between science and society. Digital image processing concept became 'popular since the middle of 1960's when third generation computer began to offer the speed and storage capabilities required for practical implementation of image processing algorithms (Oppenheium, 1978). The concept was geared up at the time when ERTS program stimulated a significant amount of digital image processing activity because of digital nature of MSS and advancement in computer aided information extraction technology (Swan, 1988). But now digital processing of image data acquired from remote sensors has passed the research and development stage. The volume of remote sensing imagery for managing a provincial resource is such that one must use digital image analysis system. By combining the remote sensing analysis and GIS resource manager can have timely and accurate" knowledge of renewable and peilshable resource (Solomonson, 1979). 2. METHODS OF LAND COVER CLASSIFICATION AND ADVANTAGES OF DIGITAL • TECHNIQUES Classification can be done by following methods i) Conventional Surveying methods- ii) Photogrammetri.c Methods iii) Digital'image processing Systems 2.1. CONVENTIONAL SURVEYING METHODS This is an outdated. method of classification. In. this method field survey is done and almost all the work is completed on the spot. This method is very laboriousycostly and time consuming. Accuracy depends on how minutely details have been collected. 2.2. PHOTOGRAMMETRIC METHODS For land cover classifidation AERIAL PHOTOGRAMMETRY haS been successfully used in the past but with advancement in information extraction technology this method is least desirable now a days. As having a flight is very costly affair, frequent field trips are difficult to manage. This technique can be successfully applied for ground verification required in 'DIGITAL IMAGE PROCESSING. 2.3. DIGITAL IMAGE PROCESSING METHODS Digital image processing has got wider application and popularity because of more and more availability of finer and finer different type of digital data from various satellites. Once data from remote sensors on board space borne platforms are available further processing becomes easy with the help of available speedy and versatile digital hardware and efficient processing algorithms. The digital techniques of image processing and analysis has an edge over the visual techniques. This is 'because of the fact that various techniques can be used over the same data for enumurous direct themes like change detection, edge detection, classification, vegitation cover/ land use etc. and -indirect themes like vegetation index, linaments and fracture. etc. Digital image processing has got the following advantages over other methods: Other methods take long time to classify the area while for defined data. base through digital techniques it is possible in