A CNN Model For Plant Leaf Disease Detection

The paper proposes a system that utilizes CNNs to accurately identify plant diseases with the help of images of leaves and provide remedies to prevent and mitigate the disease’s impact. The proposed system uses a grid-like topology with nineteen layers using VGG19 architecture. This architecture consists of sixteen convolutional layers and followed by three fully-connected layers. Where the captured image of a plant’s leaf acts as input to predict the plant’s condition, the proposed system can revolutionize how plant diseases are diagnosed and treated, offering a cost-effective solution for farmers to monitor and manage crop health. The system’s ability to accurately identify plant diseases can significantly reduce crop loss, and the remedies provided can be taken as a preventive measure to protect crops from getting affected by the disease. It is a valuable tool for farmers to protect their crops from disease and increase yield and quality. Here, we can get truthfully promising output with an optimum accuracy of 94.7%. Some remedies are provided through this system that can be taken as a cure. In this system, we used a data set collected from Kaggle containing 38 types of diseases with 3GB storage.