Rice Disease Detection using Intensity Moments and Random Forest
Sristy Saha, Sk. Md. Masudul Ahsan
Abstract
Improvement of an automated method for recognizing and categorizing various plant diseases is an evolving research area. Usually, it is very time-consuming to recognize plant diseases in remote areas, because of the communication gap between the farmer and the specialist. A programmed layout can help a farmer to discern rice plant diseases. The automatic system that is referred to here can detect the main three types of rice leaf diseases (Bacterial leaf blight, Leaf blast, and Brown spot) by the Random Forest decision tree classifier. I n tensity moments are needed here for extracting features properly. This proposed system obtains 91.47% accuracy and can classify rice diseases nicely in their primary stage. By adding some more collaborative features, the obtained result can assist the developer to rapidly identify plant diseases. This will also help the agriculturalists in active decision-taking for defending the plant professionally from ample harm.