Rice Disease Classification for Eastern Shan State Using Deep Learning

Rice is the main crop in Myanmar as well as the main regional product of Shan State. Rice plants often suffer from diseases that affect their growth and quality. The production and quality of crops are declined due to the rice diseases. Timely identification of these diseases can significantly benefit the agricultural production and help farmers make decisions about their crops. Accoding to the recent survey, there are many rice diseases in argiculture sector. Among them, this research focus on bacterial leaf blight, brown spot, and leaf smut rice diseases because of they are mostly found in Eastern Shan Sate. This research intends to implement rice leaf diseases classification model using VGG 19 and RELIEF. In this system, The relief algorithm (RELIEF) is combined to VGG 19 based transfer learning technique. To achieve image accession, we utilized a dataset containing 316 leaf images depicting various diseases. These images were categorized into four classes: healthy, Bacterial Leaf Blight, Brown Spot, and Leaf Smut. As the result of experiment, the accuracy in classifying things was 98.9%. The proposed classification model will be used the data from Kaggle and Eastern Shan State to improve the classification result. In this paper, the accuracy of proposed classification model is compared to other methods that do similar work in order to show how much better it is.