DCT based Plant Leaf Disease Recognition Algorithm using Subtractive Clustering

The main focus of the study is to implement and evaluates the DCT based plant leaf disease recognition using subtractive clustering for automatic recognition and classification. The proposed methodology of the study includes image processing and recognition by classification. The method consists of four phases: First, capturing plant leaf disease images and perform color space transformation, in the second phase, images are segmented using Otsu’s method. Next, in the third phase, we computed DCT based features of leaf infected segmented area and noninfected segmented area. Finally, the extracted features submitted to subtractive clustering based fuzzy classification. For testing and training purpose tomato late blight disease taken. The results of the proposed approach indicate that the leaf disease can recognize and classify by using image processing techniques significantly. The developed fuzzy based classifier can recognize and classify the examined disease with a precision of around 80.3%.