TinyML for Plant Disease Detection: Efficient Edge AI Solutions for Apple and Mango Leaves

Detecting plant diseases and providing timely treatment are crucial for ensuring sustainable agricultural productivity. With advances in deep learning, computer vision techniques for crop science have seen rapid growth. This paper introduces a TinyML-based model designed for real-time detection of diseases in apple and mango leaves, utilizing the MobileNetV2 architecture, which is optimized for deployment on edge devices. The model’s performance was assessed using key metrics, achieving an accuracy of 93.1%, along with high precision, recall, specificity, and an F1 score of 93.1%. It outperforms other models like ResNet-50, VGGNet-16, and AlexNet in identifying and classifying both healthy and diseased leaves, demonstrating its potential for real-time agricultural use. By optimizing the model for low-power devices, it becomes suitable for use in remote, resource-constrained areas, offering a valuable solution for improving crop management and fostering sustainable farming practices. Future developments will focus on extending the model to detect a wider variety of crops and diseases, integrating sensor data, and conducting field trials to validate its real-world effectiveness.