Disease detection of Cercospora Leaf Spot in sugar beet by robust template matching

First study introduces orientation code matching (OCM) for detecting foliar disease in plants.Robust, continuous, and site-specific observations of CLS changes in sugar beet.Pixel-wise CLS classification under varying light conditions.Classification accuracies of mean precision, recall, and F-measure were above 0.87.Promising to be applied in real field for fungicide-spraying management. In a multidisciplinary scheme linking computer science with agricultural engineering, a novel approach based on orientation code matching (OCM) for robust, continuous, and site-specific observations of disease development in sugar beet plants is presented. Differing from conventional plant disease detection approaches, we introduce the robust template matching method of OCM in this paper to not only realize continuous and site-specific observations of disease progress, but also to demonstrate its excellent robustness for non-rigid plant object searching in scene illumination, translation, slight rotation, and occlusion changes. Furthermore, a single-feature two-dimensional xy-color histogram is proposed and input into support vector machine (SVM) classifier for pixel-wise disease classification and quantification. Experimental results with high precision and recall rates demonstrate the feasibility and potential of our proposed algorithm, which could be further implemented in real sugar beet fields with robust detection and precise quantization of foliar disease development, for better analysis of disease mechanism and optimal fungicide-spraying management.