Geospatial analysis of terrain through optimized feature extraction and regression model with preserved convex region

In this paper, cat optimization algorithm for feature extraction in satellite image has been proposed. In cat optimization, cost function computes the pixel in the satellite image to preserve the boundary shape and avoid non-convex part of the contour of the image. However, the existing feature extraction optimization algorithm measures the distinct data framework and thematic information to insight land cover such as waterbody, urban and vegetation. The land cover is obtained from different optimized feature extraction algorithms never provide proper boundary shape and land feature. Furthermore, the proposed cat optimized algorithm distinguishes the inner, outer and extended boundary along with the land cover. The cat-optimised algorithm for low and high-resolution satellite image shows the better result of 85%, with the preserved convex region when compared with the existing feature extraction algorithm such as fuzzy and Particle Swarm Optimization (PSO).