Rice Disease Detection Based on Image Processing Technique

Abstract

Rice plant disease detection and monitoring is a critical issue. An accurate and timely detection of diseases in rice plants can help farmers in applying timely treatment on the plants and thereby can reduce the economic losses substantially. So, we want to develop a system on python which can detect the disease of rice plant. We studied various techniques and algorithms likes Linear Regression, Logistic Regression, CNN architectures like Resnet, AlexNet, LeeNet, VggNet and KNN regarding this issue. After many discussion and comparison between these machine learning algorithms we chose CNN architecture. We use sequential model based on CNN architecture because this architecture performs best for image classification and detection compared to other architecture or algorithm. We use a large amount of dataset for training our model and for detection. The results show that our proposed method successfully classify and find out the rice leaf diseases based on image processing techniques. Our experimental results show that we achieve 97% test accuracy with our proposed model, while other models have less than 80% accuracy