This project involves the development and deployment of a deep learning-based image classification system for solar panel images. It follows a structured machine learning workflow that includes dataset understanding, preprocessing, model architecture development, training, experimentation, performance evaluation, and deployment.
The project evaluates multiple deep learning architectures, including a baseline model, MobileNetV2, and EfficientNet, to explore their effectiveness for the classification task. Different training configurations and experimental trials are conducted to analyze model performance and identify an appropriate model for the application.
The selected model is integrated into a Streamlit-based web interface that enables users to interact with the classification system. The application is subsequently deployed on an AWS EC2 instance, demonstrating how a trained deep learning model can be made accessible through a cloud-hosted application.
The project provides practical exposure to computer vision, deep learning experimentation, model evaluation, interactive application development, and cloud deployment.