Identification of Rice Plant Disease Using Image Processing and Machine Learning Techniques

Abstract

India is known for its rice cultivation. Rice being the major crop has it’s huge impact on people’s life as majority people cultivate it for their livelihood, it generate employment too and also many small scale industries directly or indirectly depend in its cultivation. This cultivation is affected by various disease and pets which may result in huge loss. With the help of modern science and technologies many research work and methods are proposed for better yielding of rice crop. Through this research paper a method is proposed which helps early detection of 8 major rice disease namely; Brown spots, Leaf blast, Leaf smut, Sheath rot, Tungro, Sheath blight and Gudi rotten. Proposed model was designed using Raspberry pi3b+ and various sensor’s like camera, temperature and moisture. CNN classifier was employed to train the proposed model. This model was proficient to identify the disease with an overall accuracy of 99.7%. Image captured were pushed to the cloud along with details like disease name, temperature, moisture, humidity and the image timestamp.