Enhanced land use and land cover classification using modified CNN in Uppal Earth Region

This study presents a sustainable and effective approach for obtaining land uses and land covers (LULC) details utilizing remote sensing images. The modified Convolutional Neural Network (CNN), Inception-Resnet V2, is developed for accurate classification of various land cover classes in the Uppal region using Landsat-8 satellite data. The input data undergoes thorough preprocessing, including radiometric calibration, layer stacking, and resolution merge techniques, to enhance image quality and accuracy. The proposed classifier achieves a remarkable precision of 95% and kappa values of 87%, effectively predicting water bodies, bare lands, vegetation, and development land in this area. The integration of Landsat-8’s multispectral and panchromatic sensors, along with thermal imaging, enhances the classification process. This research offers a valuable and up-to-date data source for sustainable land use analysis and management.