Plant Disease Detection using Convolution Neural Network Approach

Plant disease detection is a very crucial research area to identify the important factors that create an impact on plant growth and production or crop yield. As per recent studies, 14.1% of plants die due to plant disease. It is a major issue that can also lead to health issues for humans. As many plants can be used for medicine preparation. Based on this, it is necessary to detect plant disease at an early stage. The present study is based on a Convolutional Neural Network (CNN) to detect plant disease using image data of contaminated and healthy leaves. The database used to train the CNN model in this study contains 87848 images of 25 plant species with 58 pairs of classes [plant, illness]. The best model architecture out of the ones that were trained identified the corresponding [plant, disease] pair with a success percentage of 99.68%. (or healthy plant). The presented model results in a high success rate and proves as an early warning tool for the identification of plant disease.