Identification of Leaf Blast and Brown Spot Diseases on Rice Leaf with YOLO Algorithm
Ma. Kristin Agbulos, Yovito Sarmiento, Jocelyn F. Villaverde
Abstract
Rice Leaf Blast Disease and Brown Spot Disease are among the most significant diseases affecting rice cultivation. Monitoring and maintaining rice plants from many diseases are very important to regulate their production. In this study, the key objective was to develop an application that identifies whether a rice leaf has leaf blast or brown spot disease. The You Only Look Once (YOLO) Algorithm was implemented for the development of the system. The YOLO algorithm was trained with a custom dataset of 200 rice leaf images. It was found out that the device’s accuracy for leaf blast disease was 90.00% while the class 2 brown spot disease was at 70.00%. For the third class, which was the unknown disease, the device’s performance resulted in a 100.00% of accuracy. Therefore, it was concluded that the overall accuracy of the device was at 73.33% and the error of commission was only 26.67%.