Application of Image Processing in Diagnosing Guava Leaf Diseases

Plant and leaf diseases in Guava result in poor plant growth and reduced fruit yields. Leaves are indicators of the health and growth of the Guava shrub/short tree that has its origin in tropical and subtropical regions. Precise diagnosing of diseases is vital, as remedies rely on it. Image processing in the place of manual/visual detection of Guava leaf diseases relieves from difficulties experienced, time consumed and inaccuracy resulted. In the present work resized leaf images with improved contrasts are subject to region growing segmentation, colour transformation (YCbCr,CIELAB), and Scale Invariant Feature Transform(SIFT). Support Vector Machine (SVM) and kNearest Neighbor (k-NN) classifiers have been evaluated for their disease-wise classifying accuracies. 125 leaf samples at 25 per disease class and 128 texture features per sample were used in the study. Though both SVM and k-NN perform reasonably well, the former is slightly superior in terms of accuracy