Computer Vision Technique for Detection of Grape Esca (Black Measles) Disease from Grape Leaf Samples

Grape Cultivation requires a periodic and preventive monitoring and adoption of diagnosis mechanisms for the management of diseases and for reducing the aesthetic and economic damages which are induced by the plant diseases. Grapevine Measles, also called Esca, Black Measles or Spanish Measles, is one of cataclysmic disease found in grape plant. This paper proposes an automatic computer vision method to discover the Black Measles disease from the image samples of grape leaves. Multi-channel analysis is performed to metamorphose the leaves from backdrop, followed by the identification and subjugation of the disease affected portion in the leaf image for diagnosis. Global Thresholding is followed by some mathematical morphological operations in order to precisely decline the noisy pixels. Statistical features extraction is performed and fed to a SVM classifier. The propound algorithm successfully acknowledged the affected parts present in the grape leaf images and achieved an accuracy of 97. 2% which makes the proposed algorithm robust and systematized for identification of grape leaves diseases using image processing methods.