Plant Disease Detection System for Smart Agriculture

Indian economy relies on agriculture to a greater extent. In traditional agriculture, the farmers identify the crop diseases with the help of an expert either by direct visual inspection or by sending the diseased images to experts through online services. Also, continuous monitoring cannot be done manually. The main objective is to develop an android application which identifies and classifies three major diseases - Black horse riding, Brown spot and Bacterial leaf steak. In the plant disease detection system, image to be tested is acquired, pre-processed, segmented and classified based on the disease type. The classification is performed using probabilistic linear classifier called Naive Bayes. The application is developed using Android Studio and the programming language used for the development is Java. This application identifies the plant disease based on pixel intensities, predicts the plant growth and sunlight condition if it is good or not. It suggests suitable fertilizers and pesticides depending on the disease type. The average accuracy of the developed application is about 80% and the implementation of this system reduces manpower and increases productivity.