Plant disease detection and management using deep learning approach

Plant diseases impede the average growth of plants and are one of the main reasons behind the decline in production, leading to economic loss. Early detection of the diseases helps provide a cure that can restrict the spread of the diseases within the plant. This review provides a framework for classifying plants as diseased/healthy based on leaf health using various deep learning models such as ResNet50, InceptionV3, ResNet152V2, and AlexNet and implementing those models on two different datasets, namely cotton and plant village datasets, improved respective performance. The experimental results show that the proposed models for plant disease detection and management using ResNet50, InceptionV3, ResNet152V2, and AlexNet perform well to provide a feasible solution for predicting disease among various plant leaves. But with the use of ResNet152V2 as compared to other classifiers, which gives an accuracy of 100% on the cotton dataset and 94.32% with the plant village dataset in search of the healthy or diseased plant additional to that if detected leaves are diseased, the model provides information on each disease and feasible solution. The results demonstrate that the proposed prediction mechanism represents a new effective way to predict diseases with better accuracy. The models were deployed to a web application using a flask to give users a medium to interact and utilize the mechanism.