Wheat Sunn Pest Detection and Species Classification Using Deep Ensemble Learning

The primary objective of this project is to develop an accurate, efficient, and user-friendly intelligent system for early detection of Sunn Pest damage and classification of wheat species. The proposed framework, SunnPest, integrates three high-performance deep learning models — ConvNeXt-Base, EfficientNetV2-M, and their soft-voting Ensemble — to achieve superior classification performance. The system performs binary damage detection (Healthy vs Damaged) and multi-class wheat species classification simultaneously. A secure and intuitive web application built using the Flask framework with MySQL authentication has been developed to enable real-time analysis of wheat leaf images. The project aims to create a reliable, practical, and high-accuracy decision-support tool for precision agriculture and crop protection.