Detection and Classification of Diseases of Banana Plant Using Local Binary Pattern and Support Vector Machine

Banana plantation is a commercial agricultural practice of huge significance especially in Asian and African countries. Banana production is affected by natural calamities and plant diseases. But plant diseases present a constant threat to the farmers affecting the quantity and quality of the banana cultivation. From the last decade, the image processing techniques and machine learning algorithms have been broadly used for identification and classification of infections in plants. In this work, texture pattern techniques for identification and classification of diseases in banana plants is introduced. The proposed methodology consists of two primary phases; (a) extraction of texture features from using local binary pattern (LBP); (b) classification of banana plant diseases and healthy banana plant. The texture features using LBP are extracted from an enhanced input image. The extracted features are fed to Support Vector Machine (SVM) and K-nearest neighbor (KNN) for final banana plant disease classification. The proposed technique is tested on the Plant Village dataset for the classification of two different experimental cases (i) Healthy-Black Sigatoka and (ii) Healthy-Cordana leaf spot. The proposed methodology attained an accuracy of 89.1 % and 90.9% for two experimental cases using SVM classifier.