Land Use and Land Cover Classification using Fuzzy Logic for better Accuracy: A Case Study of Ranchi

Abstract

The aim of present study is to classify the LISS III (Linear Imaging Self Scanning Sensor) image of Ranchi area of February, 2015 using standard maximum likelihood (ML) and fuzzy logic for different land use and land covers (LULC). Fuzzy logic is relatively a new concept. Now, fuzzy logic is widely used in the classification of remotely sensed images for various land use and land cover classes of mixed pixels where as standard ML classification method is unable to classify mixed pixels. Classification of images mainly includes five LULC classes viz. standing water bodies, natural vegetation and agricultural lands, dense built-up and low-density builtup area. Dense built-up area is mainly related to urban area and low built-up area is of rural areas. Image classification performed first using ML supervised and then in fuzzy logic approach. Producer's accuracy, user's accuracy, total accuracy, and kappa coefficients were calculated and tested for standard and fuzzy supervised classifications. Standard classification procedures have an overall accuracy of 86.12 percent, while fuzzy classification approaches have an accuracy of 91.56 percent. A kappa coefficient for standard method of classification is 0.84 where in fuzzy approach of classification, the kappa coefficient is 0.89. So on the basis of overall accuracy and kappa coefficients; it has been observed that the fuzzy classification technique provides better accuracies than the standard ML supervised classification approach.