Analysing the Classification of Rice Leaf Diseases on the Basis of Deep Convolutional Neural Network Architectures

Abstract

In the field of agriculture, timely investigation and recognition of plant leaf diseases assures high crop quality and yield. Due to a lack of knowledge about the most cutting-edge sophisticated approaches in the field of leaf disease detection, one of the largest obstacles for rice farmers is the identification of leaf diseases. Due to the frequency of rice leaf diseases, a large portion of rice growth is disrupted. Early detection of rice leaf diseases is now done manually by farmers, which is extremely time-consuming and labor-intensive. However, the requirement of automatic disease detection in rice leaves aids farmers in more effectively preserving their agricultural harvests. In this review, the major focus is on performance analysis of detection of rice leaf diseases based on the architectures employed. Convolutional neural networks are the best method for classifying rice leaf diseases, and advances in computer vision and deep learning satisfy predictions and turn out to be the greatest method for doing so. Numerous CNN architectures have been analyzed for finding best classification performance based on training from scratch, fine tuning or through transfer learning. Here, right selection of Deep CNN architectures for classification purposes provides high performance rates based on the type of learning employed.