Tomato Plant disease detection and classification using leaf images based deep learning

Abstract

Tomato plant diseases detection and classification using leaf images based deep learning revolves around the field of plant pathology and agriculture. The focus is on developing effective methods that leverage deep learning algorithms to enable early and accurate identification of plant diseases. Deep learning has exhibited great potential in precisely recognizing and classifying diverse types of plant diseases using leaf images. This has the capacity to significantly impact agriculture and enhance food security. By devising efficient deep learning-based approaches, this study aims to contribute substantially to the advancement of plant disease detection and classification, promoting sustainable practices in agriculture. In this work, we investigate a hybrid system that use a powerful machine learning algorithm named exponential discriminant analyses (EDA) with transfer learning process based on recent and the powerful deep nets including, Resnet50, Darknet53, Densenet201and Efficientnetb0. Our system is evaluated on two challenging datasets, Plantvillage, and Taiwan tomato leaves with 18,340 and 14,531 tomato images respectively. The experimental results show high competitiveness of the proposed method.