K-means Clustering and SVM for Plant Leaf Disease Detection and Classification

摘要

The role image processing role is inevitable in the computer vision, robotics, agriculture, and medical field. This work proposes K-means clustering algorithm for the detection of leaf disease and classification. Prior to segmentation and classification, preprocessing was performed by the color median filter. The image in the RGB color model was converted into L*a*b model for segmentation. The color texture feature is extracted and fed to multiclass SVM classifier. The algorithms are developed in MATLAB 2015a and tested on real time images. The classification accuracy on an average for SVM was found to be greater than 95%.