Image processing for paddy disease detection using K-Means Clustering and GLCM Algorithm

Abstract

The traditional human-based visual quality inspection approach in agriculture is unreliable and uneven due to various variables, including human errors. In addition to the lengthy processing durations, the traditional method necessitates plant disease diagnostic experts. On the other hand, existing image processing approaches in agriculture produce low-quality output images despite having a faster computation time. As a result, a more comprehensive set of image processing algorithms was used to improve plant disease detection. This research aims to develop an efficient method for detecting leaf diseases using image processing techniques. In this work, identifying paddy diseases based on their leaves involved a number of image-processing stages, including image pre-processing, image segmentation, feature extraction, and eventually paddy leaf disease classification. The proposed work targeted the segmentation step, whereby an input image is segmented using the K-Means clustering with image scaling and colour conversion technique in the pre-processing stage. In addition, the Gray Level Co-occurrence Matrix technique (GLCM) is used to extract the features of the segmented images, which are used to compare the images for classification. The experiment is implemented in MATLAB software and PC hardware to process infected paddy leaf images. Results have shown that K-Means Clustering and GLCM are capable without using the hybrid algorithm on each image processing phase and are suitable for paddy disease detection.