Rice Diseases Multi-Classification: An Image Resizing Deep Learning Approach

Rice disease detection (RDD) has become an active area in the research community. In recent years, development and using deep learning (DL) models for RDD has proved to be an efficient area. Keeping this in mind, DL model based convolutional neural networks (CNN) has been developed to categories four types of rice plants with a dataset having images of rice blast, hispa, brown spot, and healthy rice plant. A total dataset of 1600 images has been collected from Punjab, India. An image rescaling-based technique has been used to analyze the performance of the model based on three different dimensions of the input image which results detection accuracy (AC) of 92.83% in detecting hispa disease with image input dimension as 512*512 pixels. This work will be primarily helpful for the farmer community of India at large to detect the disease in rice fields. It will also be helpful for the research community working on RDD using the DL approach.