K-Means cluster based leaf disease identification in cotton plants Vani K.1, Poongodi S.2, Harikrishna B.2 1Assistant Professor, Department of ECE, CMR Engineering College, Hyderabad 2Professor, Department of ECE, CMR Engineering College, Hyderabad Online published on 1 November, 2018. Abstract Pests are found to be most common and continuous threat for the agricultural crops in India which would affect the entire plant including leaves and roots. The quality of plant products can be improvised by introducing the early diagnosis techniques for plant diseases. Last year analysis of cotton crops found that the bulk of crops are lost due to increased infestations created by pests and insects. In this proposed method, detection of pests and types of disease occurred in cotton plants by introducing the SVM classifier. Initially image capturing devices are integrated in the crop yields to gather the plant images periodically. These time serious images are preprocessed first using median filter which is then segmented to attain the required image part using k means clustering method. After segmentation feature extraction is done where the following features are extracted: Color features (mean, skewness), texture features (energy, entropy, correlation, contrast, and edges). This texture feature extraction is done by using Gray Scale Co-occurrence Matrix (GSCM) which is then matched with the cotton leaf image without disease. The experimental evaluations are carried out using MATLAB software. Top Keywords K-means clustering method, Gray scale co-occurrence matrix, SVM classifier. Top |