A Land Use/Land Cover Classification Of Irs (Liss – Iii) Multispectral Data Using Decision Tree And Svm Classification Mechanism

Abstract

Identification of the effect of human activities on our planet concerns over worldwide land use and land cover change is very complex. Land Use and Land Cover refer to the utilization of land through events like agriculture, different types of cultivation areas, residential areas, and the physical features on the earth’s surface like the sea, mangroves forest, vegetation cover, and water bodies. However, empirical techniques used for this type of classification are cost-effective and laborious. This paper is focused on remote sensing images and various supervised classifications to identify various Land Use/Land Cover. This research work aims to use images taken from IRS (LISS III) platform to perform supervised classification. The study was performed to compare the performance of Supervised classifiers Decision Tree and SVM to classify different land use land cover classes. The Decision tree classifier gives better results than SVM for the study area. The decision tree classifier achieved 89.97 %. and SVM 81.90 %. It revealed that Decision Tree did better across different levels of occupancy of Land use/Land Cover.