Forecasting of Landuse and Land Cover Changes by Applying Support Vector Machine Using Landsat Satellite Imagery

Authors- Priyanka Gupta, Sharda Haryani, V.B. Gupta

Abstract-The primary goal of this study is to identify the LULC classes for the Madhya Pradesh, India, Mandsaur district. The research was conducted using multi spectrum satellite images. The paper uses support vector machine approach based on pixel by pixel supervised categorization of Landsat satellite pictures taken over 20 years between 2003 and 2023 using Arc-GIS tool. To predict general changes, different classifications of land use and land cover features such as populated regions, water bodies, agricultural land, woods, and desert terrain are taken into account. To do this, remotely sensed Landsat 5 photographs from 2003 and Landsat 8 photos from 2023 were employed for change detection. This paper used the support vector machine technique to compare the LULC classes for the Mandsaur region. The supervised classification findings validated with Support Vector Machine (SVM) provided kappa coefficients of 0.835 and 0.803 for the years 2023 and 2003, respectively. The use of Support Vector Machine (SVM) algorithms should have a significant positive impact on land cover classification.

DOI: /10.61463/ijset.vol.12.issue4.227