Sensors-based prediction of plant diseases using neural network

Abstract

Plant diseases could cause a loss to agricultural production and the economy; hence there is a need to develop prediction models for fast plant disease detection and assessment methods. These methods must not affect plant growth when deployed. Not only plant diseases can be detected, but also the production can be improved drastically. Fungal infection is the most dominant but can be controlled by appropriate measures if detected at an early stage. The paper aims to develop an expert system for the prediction of various fungal diseases like powdery mildew, anthracnose, rust, and root rot/leaf blight. Hence, we propose an artificial neural network for the classification of the diseases. This paper validates a real-time dataset, captured at DA-IICT, Gandhinagar, India. The results give high classification accuracy for the proposed model. The implementation of this work proves the feasibility of using this technique for faster plant disease detection at an affordable cost. As soon as early detection is achievable in the field, the damage to the crop due to these diseases could be reduced.