A time efficient leaf rust disease detection technique of wheat leaf images using pearson correlation coefficient and rough fuzzy c-means

Abstract

In agricultural sector diagnosis of crop disease is an important issue, since it has a marked influence on the production of agriculture of a nation. It is very essential to diagnose disease in an early stage to control them and to reduce crop losses. This paper presents a time efficient proposed technique to detect the presence of leaf rust disease in wheat leaf using image processing, rough set and fuzzy c-means. The proposed technique is experimented on one hundred standard diseased and non-diseased wheat leaf images and achieved 95 and 94% success rate respectively depending on most three dominated features and single most dominated feature, Ratio of Infected Leaf Area (RILA). The three most dominated features and single most dominated feature are selected out of ten features by the Pearson correlation coefficient. A significant point of the proposed method is that all the features are converted into size invariant features.