Land Surface Temperature Classification Using Machine Learning Algorithms for the region of Kolli Hills, Tamil Nadu, India

Land Use Land Cover change is considered an exciting topic in earth sciences and information technology. The real-time multispectral satellite images captured for finding the variations in the Earth's surface can be explored through Remote Sensing and Geographic Information Systems. In our work, the Land Surface Temperature image was extracted using the thermal bands of the Landsat-9/8 satellite, and the temperature values were computed for the Kolli Hills, India spanning the years 2021, and 2024. This work provides novel information about the Land Surface Temperature classification using the Clustering-based maximum likelihood machine learning model. The accuracy of the work was calculated with the assistance of Google Earth Engine time series images. The random sampling method was applied to validate the correctness of the acquired maps and the result shows an overall accuracy of 96.484%. The significance of this work lies in providing Land Use Land Cover information for time series data, supporting land resource organizers in preserving forest-covered areas in hilly regions.