Diagnosis of Plant’s Illness in Agriculture by CNN Technique in Deep Learning: An Analytical Review

Abstract

Multiple diseases have the potential to drastically impair crop productivity, seriously a threat for food security. After recognizing diseases by different symptoms, taking action is made much easier by automatic systems for classifying plant illnesses. Therefore, it is critical and vital to accurately detect plant illnesses. In cultivation, early illness identification becomes crucial ensuring a successful harvest production. Traditional categorization techniques like visual analysis have lot of drawbacks, including the need for a lot of time and accuracy. Currently, classifying numerous plant diseases has proven to be very successful when using Deep learning (DL), machine learning (ML), and convolutional neural networks (CNN). CNN based on DL techniques, in particular, have found considerable use during classifying different plant illness or diseases. We demonstrated that (NN) Neural Networks, when used for diagnosis, can capture the hues, saturation, texture of contusion particular to a given disease, simulating human judgment. They represent cutting-edge technology in this area and have partially overcome the issues with conventional categorization methods. In this study, we analyzed the most recent CNN networks techniques and algorithms relevant to classifying plant leaf diseases and reviewed CNN models for automatic feature extraction and classifying technique. Additionally, we outlined CNN's primary issues and their associated fixes for classifying plant diseases. We also talked about the classification of plant diseases' future development. The findings encourage the end users for betterment in the diagnosis procedure, which will result in an even more effective application of DL for identifying plant illness.