Plant Disease Detection and Classification Using Deep Learning

Abstract

Tomato leaf disease identification using computer vision is a cutting-edge application of artificial intelligence (AI) and image processing techniques to agriculture. It plays a crucial role in modern farming by enabling early and accurate detection of diseases that affect tomato plants. This technology leverages the power of computer vision to automatically analyze images of tomato leaves and identify any signs of diseases, providing farmers with valuable insights and helping them take timely corrective measures. In this paper, we propose a completely automated end-to-end tomato leaf disease identification frameworks using Convolutional Neural Networks (CNN) and Computer Vision (CV). The proposed framework is trained and evaluated on Plant Village database containing several tomato leaf diseases along with healthy samples. We employed transfer learned CNN as well as developed a custom CNN architecture with extensive hyper-parameter optimization. The proposed framework obtained the highest classification accuracy of 92% on unseen samples. Hence, the system can be efficiently utilized for the task of real-time tomato leaf disease identification and classification.