Detection of cotton leaf diseases using image processing

Abstract

Agriculture is the main occupation of the Indian population. Plant diseases are the alarming threat to the livelihood of agriculturists. Hence, proper detection and classification of the diseases will help the farmers to take precautionary measures and to increase the crop yield. The paper focusses on applying image processing and neural network techniques in identifying the diseased cotton leaves from the set of images taken from the village plant dataset. The leaves are first pre-processed to filter the noise and then converted to gray scale image. Image enhancement is done to sharpen the image followed by clustering and segmenting the diseased parts. Then, a set of features are extracted followed by the classification of the disease using Radial Basis Function (RBF) neural network classifier. Training and testing are carried out on a large number of samples and the results are compared with that of SVM classifier. The proposed method outweighs the SVM classification technique in providing better accuracy.