Early Detection of Diseases in Coconut Tree Leaves

摘要

Coconut tree has got its name as “functional food” since it is rich in nutrition content. Coconuts are nutritious in fiber, vitamins and minerals. The coconut is most important source of food as it provides oil, water, milk and also medicine. For more than hundreds of years, the coconut fruit of the coconut palm has been a great source of versatility. It has been used by human beings for their day to day requirements. Today, coconut is a valuable source of both food and medicine for people from much diverse culture and religion. Coconut leaves are affected by a number of diseases, some of which are dangerous and the disease gradually reduces the strength of the leaves causing severe losses in the yield. This paper deals with the identification of diseases of Coconut tree leaves. These leaves are affected by the diseases namely leaf rot, leaf blight and leaf yellowing. The database of such leaf images is carried out and preprocessing of them is done. Segmentation and k-means clustering is done to differentiate the fault regions that remain abnormal. From this, the statistical and GLCM features are extracted and Cubic SVM classification is carried out. The accuracy of 97.3% is achieved which is better than the existing methodologies. This would greatly help in increasing the yield to greater extent and reducing the economic losses.