Plant diseased leaf segmentation and recognition by fusion of superpixel, K-means and PHOG

Abstract

An Internet of things (IOT) based plant diseased leaf segmentation and recognition method is proposed based on Fusion of Super-pixel clustering, K-mean clustering and pyramid of histograms of orientation gradients (PHOG) algorithms. Firstly, the color diseased leaf image is divided into a few compact super-pixels by super-pixel clustering algorithm. Then K-means clustering algorithm is employed to segment the lesion image from each super-pixel. Finally, the PHOG features are extracted from three color components of each segmented lesion image and its grayscale image, and concatenate four PHOG descriptors as a vector. The experiment results on two plant diseased leaf image databases indicate that the proposed method is effective. This paper provides a feasible solution for plant diseased leaf image segmentation and plant disease recognition.