A novel approach to classify and detect bean diseases based on image processing

摘要

The early detection of plant diseases is one of the main reasons that can reduce the world crop losses. It requires a tremendous amount of effort, money, and time. Therefore, the algorithms used in image processing make it possible to detect plant leaf diseases automatically. This paper focuses on detecting two types of bean leaf diseases including bacterial brown spot and powdery mildew. The detecting process involves acquisition, preprocessing, segmentation, feature extraction, and classification. The training and testing images are taken from a public database. The developed methodology can successfully detect the two types of plant leaf diseases with an accuracy of 100%.