Identification of Wastelands Using Contrast in Vegetation Activity in Vadodara District, India - A Remote Sensing Approach

Abstract

The high population pressure, fast urbanization, rapid industrialization, and extensive agriculture have put great stresses on land resources, resulting into the substantial reduction in agricultural area and natural resources. Increasing population worldwide is also leading to deforestation and resource degradation that has disturbed the balance of terrestrial ecosystems. Recent studies reported that many areas covered by wastelands are decreasing because parts of wastelands are being converted into arable land. It is important to identify and monitor these changes in spatial planning and management. This paper adopts a remote sensing-based identification of culturable wastelands based on seasonal vegetation changes in Vadodara district, India. Supervised classification was applied on three MODIS images of 2016-17 of 3 different seasons. Separability analysis was applied to get the best data combination for image classification. Validation was done by ground referencing and Google earth images. The composite of winter season image with NDVI and EVI performed best with an overall accuracy of 78.2% with the kappa co-efficient of 0.7580. This method opens a possibility of using digital classification for identification of culturable wastelands in the study area which are so far mapped with visual interpretations only.