Artificial Intelligence Based WhatsApp Bot for Plant Disease Detection

Plant disease lowers the quantity and quality of food, fiber, and biofuel crops, as agriculture tries to feed the world's fast-expanding population. As a result of plant diseases and infections, large-scale farmers and fun-time gardeners face growing concerns about the growth of their plants. Further, many areas do not have access to agricultural testing laboratories or people who can treat the disease. So, we proposed a method wherein any user having a smartphone with WhatsApp installed, can take pictures of the diseased part of the plant and send them to the WhatsApp Bot and the bot sends back the disease-affected areas segmented with the disease type and the probability. The proposed system makes use of HSV Filtering, MobilenetV2, and YoloV8 for disease identification. The algorithms together offer the best and most accurate result. Instance segmentation was used to identify more than one disease appearing on a plant part. As a result, the Chatbot can help the user recognize plant diseases and its cure. The classification model has an accuracy of 95.66% and the segmentation model showed a mAP of 0.756 respectively. Overall, the solution offers portability as farmers need not go to any physical agriculture centers to get plant remedies which can aid the early detection and cure of plant detection with no extra cost.