Coconut Leaf Disease Research Paper

Classification of coconut leaf diseases using SVM P. Anu bharathy
Department of ECE
Renganaygi Varatharaj College of Engineering
Sivakasi, Tamilnadu [e-mail] P. Anu bharathy
Department of ECE
Kalasalingam Institute of Technology
Srivilliputhur, Tamilnadu [e-mail] Abstract— In order to prevent the coconut yield loss, the coconut leaf disease identification is very important. Monitoring the diseases in coconut leaf manually is quite difficult and time consuming. The coconut leaf diseases are classified based on the lesions present on the coconut leaf.

By averaging all of the pixels in the cluster, the cluster centers are computed. Until the convergence is attained, repeat the steps 2 and 3. After k-means clustering segmentation is performed, the green color pixels are masked. If the green component [6] of pixel intensities is less than the computed threshold value, then the pixels with zero components and the pixels on the boundaries of the infected clusters got deleted.
D. Feature extraction
In classification of coconut leaf diseases, the feature extraction is very important. The method used for feature extraction is color co-occurrence method which considers both the color and the texture of an image. The matrix which has same number of rows and columns as the number of gray levels in an image, this matrix is known as gray level co-occurrence matrix. The textural features derived from the GLCM are correlation, contrast, energy, homogeneity, entropy, variance, smoothness and IDM. The color features are mean, standard deviation, kurtosis and skewness. The feature descriptions are briefly described below:

CONTRAST: The contrast (1) returns a measure of the intensity contrast between a pixel and its neighbor over the whole image derived from GLCM given