Computer vision based approach to detect rice leaf diseases using texture and color descriptors

One of the major reason behind degradation of quality and quantity of rice crop is pest. The lack of technical and scientific knowledge to prevent pest diseases is the main reason for low production of these commodities. This article aims to develop a computer vision based automatic system for the diagnosis of diseases caused by pests in the rice plants. Automatic disease detection using computer vision approach involves three types of feature extraction in this experiment. Diseased area of the leaf, textural descriptors using gray level co-occurrence matrix (GLCM) and color moments are extracted from diseased and non-diseased leaf images resulting in 21-D feature vector. Genetic algorithm based feature selection approach is employed to select relevant features and to discard redundant features, generating a 14-D feature vector that reduces the complexity. Artificial neural network (ANN) and support vector machine (SVM) is used for classification. The proposed algorithm results in classification accuracy of 92.5% using SVM and 87.5% using ANN.