The reptile optimized deep learning model for land cover classification of the uppal earth region in telangana state using satellite image fusion.

This article focuses on improving land use and land cover (LULC) classification in the Uppal region by integrating fused Landsat-8 multispectral and hyperspectral satellite imagery with an optimized deep learning model. The proposed method employs preprocessing techniques (dark subtraction, stacking, merging for Landsat-8; 9P and Savitzky-Golay smoothing for hyperspectral data), followed by Principal Component Analysis (PCA) for feature extraction and a hybrid Non-Subsampled Contourlet Transform–Fast Discrete Curvelet Transform (NSCT-FDCT) for image fusion. Classification is performed using a modified ResNeXt-50 convolutional neural network whose weights are optimized via the Reptile Search Optimization (RSA) algorithm, inspired by crocodile hunting behavior. The approach achieves high performance metrics—97% accuracy, 96% sensitivity, 99% specificity, 97% precision, and a 95% Matthew Correlation Coefficient—surpassing existing classifiers and demonstrating the advantage of fused satellite data for accurate LULC identification in complex urban environments like Uppal.