62 Textural Classification of Single Band SIR-B Data over a Part of Brahmaputra Basin, India Tapan Jyoti Majumdar and Kamini Kanta Mohanty Marine and Water Resources Division, Remote Sensing Applications Group Space Applications Centre (ISRO), Ahmedabad - 380 053, India Abstract Texture is an important spatial feature, useful for identfying objects of regions of interest in an image. There are a number of methods for identification of textural parameters e.g. edgeness, frequency domain analysis, gray tone cooccurrence approach etc. Geologic information in radar images of heavily vegetated areas is contained mostly in the depiction of topography in image texture. Single band SIR/ERS-1 SAR data posses a problem to the analyst for classification of the various land use/geological classes and generally multidate SAR data are used due to paucity of more number of bands. However, the multidate SAR data classification is not an ideal technique. A new technique, namely, spatial frequency band pass classification technique which generates two or more bands in the Fourier domain using the single band SAR data and then classifies various features of interest using their textural properties has been described here. Result shows higher percentage of classification using Maximum-Likelihood Classifier (MXL) with two split-band data as compared to the unsupervised classification of all the bands. in an image. The gray-tone cooccurrence matrix is frequently used for such characteristics to extract textural information from digital images (Haralick, 1979). A simple technique for textural analysis of radar images had been developed by Blom and Daily (1982) which splits the spatial frequency spectrum produced by a Fourier transform into two parts and then classifies the image after inverse transform. Petroleum-bearing structures already established by geophysical means have recently been collated with Landsat- derived information (Halbouty, 1980). The utility of radar data for discrimination of lithological units has so far been poor to moderate. Variations in lithology of surficial rocks have been successfully detected in both airborne and spaceborne SAR imagery. Analysis of Seasat and SIR-A data has suggested that orbital radar imagery can be used to detect lithological boundaries between different geological materials in certain types of environments (Jaskolla et al., 1985). The Shuttle Imaging Radar B ( SIR-B ) was launched into orbit aboard the Space Shuttle Challenger on October 5, 1984. SIR-B acquired digital radar image coverage of different terrains as well as ocean surfaces at different incidence angles. This was different compared to SIR-A which had earlier collected data at a fixed incidence angle. Details of SIR-B is given in Table 1 (SIR-B Preliminary User’s Guide, 1984). ERS-1 Synthetic Aperture Radar (SAR) data as well as Shuttle Imaging Radar (SIR - A/SIR - B) have collected data in one microwave frequency. However, due to the paucity of other channels, generally two different angles are used as is Introduction Texture is one of the fundamental elements used in human interpretation of remote sensing imagery. The concept of texture is closely associated to tone as texture represents the spatial pattern of tone in an image. The textural measures of a particular surface feature of interest can lead to very important conclusions particularly in the case of microwave data. Haralick (1976) has discussed in details various methods for estimation of different textural parameters, namely, the autocorrelation function, digital and optical transforms, structural elements, edgeness, and spatial gray tone cooccurrence probabilities. Of them, the autocorrelation function, optical and digital transforms basically relate texture to spatial frequency. Rosenfeld and Thruston (1971) first showed texture in the form of number of edges per unit area. This is the simplest of all in the existing measures for studying textural property of a material. Textural parameters estimation and classification using autoregressive models and wavelet transform has been discussed elsewhere (Chang and Kao, 1993; Mao and Jain, 1992). Methods of texture analysis are usually divided into two major categories (He and Wang, 1990, 1991; Irons and Peterson, 1981). The first is the structural approach, where texture is considered as a repetition of some primitives, with a certain rule of placement. The traditional Fourier spectrum analysis is often used to determine the primitives and placement rule. The second major approach in texture analysis is the statistical method. Its aim is to characterize the stochastic properties of the spatial distribution of gray levels Geocarto International, Vol. 14, No. 2, June 1999 Published by Geocarto International Centre, G.P.O. Box 4122, Hong Kong.