Plant Disease Detection System

Plant diseases are an ongoing challenge for emerging farmers, which threatens money and food security. Recent changes in the penetration of smartphones and computer viewing models have created an opportunity for the separation of images in agriculture. Convolutional Neural Networks (CNN) are regarded as technological standards in image recognition and provide the ability to provide a quick and clear diagnosis. In this paper, the effectiveness of the ResNet34 model previously trained in diagnosing plant diseases is being investigated. The advanced model is still distributed as a web application and can detect 7 plant diseases with healthy leaf tissue. Database containing 8,685 leaf photographs; installed in a controlled environment, it is established to train and validate the model. Verification results show that the proposed method can reach 97.2% accuracy and an F1 rating above 96.5%. This shows the technological potential of CNNs in the identification of plant diseases and paves the way for AI solutions for emerging farmers.