An Identifying Plant Leaf Diseases Based on Color, Shape and Texture Features

Abstract

An identification of disease on the plant is a very important key to prevent a heavy loss of yield and the quantity of agricultural product. The symptoms can be observed on the parts of the plants such as leaf, stems, lesions and fruits. The leaf shows the symptoms by changing color, changing shape and texture, by showing the spots, leaf rolling and holes on it. The aim of the project is to identify and classify the disease accurately from the leaf images. The analysis conducted by extracting color, shape and texture features from healthy and unhealthy tomato leaf and black gram leaf images. We propose a disease recognition and classification approach which consists of framework to pre-processing, segmenting, feature extraction, and diseases classification and recognition. In pre-processing diseased leaf image should be enhanced and then segmenting diseased leaf images by using K-means clustering, extracting shape and color features from lesion information, texture feature extracted by LGGP (combining LBP and Gabor filter) technique. Extracted features from segmented images fed to classification. Diseased leaf images classifying by using sparse representation approach. Combining of these proposed techniques efficiency can successfully detect and more accurately classify the examined diseases.