Exploring Applications of Convolutional Neural Networks in Analyzing Multispectral Satellite Imagery A Systematic Review

This project focuses on classifying multispectral satellite imagery into 14 distinct classes, such as agriculture, forestry, and harbor, using machine learning models. The models implemented include MobileNet, ResNet, InceptionNet+RF, and DenseNet+SVM, all of which are trained for precise image classification. The system is developed as a web application using Flask for the backend and HTML/CSS for the frontend. Upon registration and login, users can upload satellite images, which are processed locally using pretrained models to provide image predictions. This approach ensures an efficient, scalable solution for analyzing satellite images and can be applied to fields such as environmental monitoring, agriculture, and natural resource management. By combining deep learning models with satellite imagery, the system offers a robust and user-friendly tool for addressing real-world challenges in land use planning and environmental analysis, leveraging machine learning to automate the classification of complex satellite data.

Keywords:

Convolutional Neural Networks (CNNs), Satellite Imagery, Multispectral Classification, MobileNet, ResNet, InceptionNet+RF, DenseNet+SVM, Image Prediction, Flask Web Application, Agriculture, Forestry, Harbor, Machine Learning