MobileNet-based disease detection with AWS S3 integration
AgriCure is an end-to-end plant disease detection system designed to assist farmers, agricultural officers, and crop-monitoring services by enabling real-time diagnosis of tomato leaf diseases using a lightweight deep learning model deployable on low-resource devices. The project leverages MobileNet, a highly efficient CNN architecture optimized for mobile deployment, enabling on-field disease identification even in areas with limited computational infrastructure. Farmers can simply upload an image of a tomato leaf, and the system automatically classifies it into one of the target disease categories while simultaneously storing images in a structured cloud repository for downstream analysis. The solution integrates lightweight deep learning (MobileNet) for disease classification, AWS S3 for scalable image storage and dataset management, data preprocessing pipeline for image cleaning, augmentation, and normalization, and a production-ready workflow suitable for smartphone applications and automated monitoring systems.
Python, TensorFlow / Keras, PyTorch, MobileNet, Transfer Learning, ONNX, TensorFlow Lite, AWS S3, AWS IAM, boto3 (AWS SDK), Flask / FastAPI, OpenCV, NumPy, Pandas, Matplotlib, Docker, CI/CD