Identification of paleochannels in and around “Chandraketugarh”, Ganges Delta through remote sensing techniques using fuzzy inference system

Abstract

A method for identification of most probable paleochannels in and around “Chandraketugarh” (an archeological site in the western part of Ganges Delta) was developed through remote sensing techniques using fuzzy inference system (FIS). The method was adopted considering the uniqueness of the geomorphology in terms of terrain homogeneity of the area. In order to categorize different geomorphological features, NDVI, NDWI, and elevation datasets were generated using multispectral images (LISS III) and CARTOSAT I datasets. Based on the unique spectral signatures of the different geomorphological units, the said datasets were critically examined to be utilized as the input parameters of the employed FIS-based image segmentation method. The result of subsequent analysis lead to development of a land use/land cover (LULC) map which portrayed inferred paleochannels. The developed method was validated with reference to a finer resolution image and ground truth datasets. It is concluded that the newly developed FIS-based method is quite efficient for identifying probable paleochannels in a terrain where slope variation is minimal and geomorphology does not seem to have been strongly affected by structural elements.