A hierarchical stratagem for rice leaf disease distinction

India is one of the largest producers in agriculture. Agriculture always bears the prime role in Indian economy. Rice, scientifically known as Oryza Sativa, holds one of the largest producing grains in India. It accounts marketable production. Unfortunately rice cultivation is facing many challenges nowadays due to the attacks of insects and other factors on rice leave causing rice leaf diseases. Inspiring from the fact that, there are fundamental discriminations between diseased and non-diseased rice leave and minute distinction among the diseased ones, this paper has introduced a hierarchical strategy for the diagnosis of rice leaf diseases consisting of two stages, namely, stage 1 and stage 2. Henceforth for the classification of the two major categories — diseased and non diseased rice leave has been attained by extricating the features of the pre-processed image based on color moment in stage 1. Stage 2 deals with the sub classification of infected leave into three categories such as brown spot, blast, and leaf blight through the extraction of textural features from the gray level co-occurrence matrix of the pre-processed image. Some standard classifiers like Support Vector Machine, k-Nearest Neighbour, Bayes' classifier, Multi Layer Perceptron, and Decision Tree have been instantiated in the proposed work.