Land use/land cover classification using Sentinel-1 imagery and Support Vector Machines

Abstract

Describing the existing and quantifying the extent of change in land use for spatially distributed land cover pattern of a selected study site using different machine based algorithms has been a topic of interest from a long time. The main objective of the paper presented here is appraisal of land use/land cover pattern for Varanasi study site using Sentinel-1 images using pixel based Support Vector Machines (SVMs) Polynomial algorithm. SVMs, supervised image classifiers has been particularly appealing in image classification studies with its desirable trait to successfully handle small training data and producing better accuracies than traditional methods. The results indicated an overall accuracy of 97.09% for the Sentinel-1 image classification using SVMs, which confirms the suitability of the microwave imagery like Sentinel-1 for land use/land cover studies.