Effects of Deforestation Analysis Using Satellite Imagery

Abstract

Deforestation has become one of the major environmental challenges affecting biodiversity, climate regulation, and ecological balance. Continuous monitoring of forest regions using satellite imagery is essential for identifying vegetation loss and supporting sustainable forest management. Conventional monitoring methods, including manual surveys and visual interpretation of satellite images, are time-consuming, labor-intensive, and unsuitable for large-scale analysis. This paper presents a deep learning-based deforestation analysis system that uses satellite imagery for automated environmental monitoring. The proposed system is implemented as a Flask-based web application where users upload satellite images for analysis. Initially, a Convolutional Neural Network (CNN) validates whether the uploaded image is a genuine satellite image. The validated image is then classified into DIM, MEDIUM, or BRIGHT categories according to its luminosity characteristics. A luminosity index is calculated to estimate surface brightness, and K-Means with Expectation-Maximization (EM) clustering is applied to segment similar land-cover regions. The clustered output assists in identifying vegetation cover, barren land, exposed soil, and potential deforested regions. The system is developed using Python, TensorFlow/Keras, Flask, OpenCV, NumPy, SciPy, scikit-learn, Pillow, and Matplotlib. Experimental results demonstrate that the proposed system provides efficient satellite image validation, classification, luminosity estimation, and clustered visualization for supporting deforestation effect analysis and environmental monitoring.