Disease Detection in Paddy Crop using CNN Algorithm

Abstract

In this paper, we propose a technique for early detection of the diseases like blast and blight in paddy crop using one of the deep learning algorithms called as CNN algorithm [8]. The entire system works on the Raspberry-pi which acts as processing unit for the proposed system. The well trained keras model to predict the disease in paddy leaf is developed using CNN algorithm [5]. As the symptoms of the blast and blight disease are detected using feature of the paddy leaf, so the model is trained using the diseased image data set. The pi-camera will capture the image of the leaf and the captured image will be processed by the python script running on raspberry pi. The images will be captured at different angle with respect to raspberry-pi. The trained model will predict whether the processed image of the leaf contains the symptoms of the disease. The proposed system also includes the feature which will alert the farmer about spreading of the disease by using the GSM module to send the message to the farmer of the field. Also the pesticides are suggested to farmers to control the further growth of the germs causing the disease. In India paddy is grown as the staple crop. The paddy crop is mostly damaged due to the leaf diseases called as blast and blight disease[4]. About 20%-30% of the yield will be destroyed by the blast and blight diseases respectively [14]. Hence, by implementing the above proposed system the disease can be detected at the early stage and loss can decreased.