Monitoring Desertification in Adrar Algeria Using Deep Learning and Spectral Indexes with Landsat-8 Images

Abstract— Desertification is a critical environmental issue, exacerbated by climate change and human activities. This study focuses on monitoring desertification in the Adrar region of Algeria using deep learning techniques and spectral indexes derived from Landsat-8 images. By using convolutional neural networks (CNNs) combined with indices such as NDVI, TGSI, MSAVI, and NDBI, we aim to enhance the accuracy of land classification and improve the detection of desertification patterns. The proposed method involves a temporal analysis of Landsat images from 2015, 2018, 2021, and 2024. Our results demonstrate a significant increase in desertified areas, highlighting the urgent need for intervention and sustainable land management practices.