Multi-Crop Plant Leaf Disease Detection Using Lite Models

India, being an agricultural nation, significantly relies on agriculture for its economic growth. Farmers in India are facing several problems with plant diseases that directly impact crop yield. Plant diseases reduce both crop yield and quality. Detecting and treating plant diseases is crucial for achieving a high-quality crop yield. Identifying plant leaf disorders with the human eye is challenging and time-consuming. Our research will help farmers identify diseases even in their early stages with high accuracy and low time. We utilized a Convolutional Neural Network (CNN) trained on a large dataset for plant disease detection. Raw photos of normal leaves as well as sick leaves of various species abound in the dataset. The framework extracts features from the leaf image and classifies them based on the color, size, symptoms of the disease, and several factors and identifies the disease in real time. Adding more images and using EfficientNet-Lite has increased accuracy and efficiency with lower computational power. This model has higher potential than conventional transfer learning models. The suggested study outperforms the several machine learning tools currently in use in terms of accuracy. The results of this study show that the proposed model achieved 93% accuracy, indicating superior performance compared to traditional methods. The system's real-time recognition function allows early intervention and minimizes harvest loss. Future research should focus on expanding data records to improve inference period optimization, generalization, and further integration of environmental factors.