Identification of Tomato Plant Disease Using Faster R-CNN and RetinaNet

Abstract

Plant disease plays a significant role in the low productivity of tomatoes which leads to huge loss to the farmer and the country's economy. Identification of plant disease at an early stage can play a major role in producing good amounts and even good quality tomatoes. Identification of the disease sometimes gets difficult because of lack of knowledge or having multiple diseases or even wrong prediction of disease. The chapter contains implementations of two classification algorithms Faster R-CNN and RetinaNet. Both the algorithms are initiated from transfer learning and tested on different hyperparameters for better results. The work produces impressive results with average precision (AP)-50 as FR 93.11%, RN 95.54%. The preliminary results look promising and can be helpful for harvest quality and precision agriculture.