An Image Processing Framework To Identify Rice Blast

Abstract

An early detection of rice plant disease especially rice plant leaves disease detection can assist farmers to take necessary precaution at the early stage and can achieve better quality of crops. Rice plant can be affected by various types of fungal infectious diseases and among them rice blast is a common one. There are a numerous image processing approaches available today which can analyze rice plant leaves disease. Existing most approaches considered binary threshold based segmentation approach although input images are always RGB color images. To develop an automated system to identify and classify rice blast diseases it is always beneficial to use RGB color images as input and to provide analysis results in RGB color images as well. This study proposed a suitable frame work where enhancement, filter, color segmentation and color feature for classification steps were incorporated for identification. CNN classifier was applied to increase the identified accuracy rate. Compared to all other existing approaches this study proposed framework provides an acceptable accuracy rate of 97.43%.