Rice Pest and Disease Detection Using Convolutional Neural Network

Abstract

Detection of rice pest and diseases, and proper management and control of pest infested rice fields may result to a higher rice crop production. According to the International Rice Research Institute, farmers lose an average of 37% of their rice crops due to pest and diseases, yearly. Using modern technologies, like smart phones, farmers can be aided in detecting and identifying the type of pests and diseases found in their rice fields. This study proposed an application that will help farmers in detecting rice insect pests and diseases using Convolutional Neural Network(CNN) and image processing. It looked into the different pests that attack rice fields; information on how they can be controlled and managed was considered; farmers' knowledge in different rice pests and diseases, and how they control these pests was regarded in this study; the study also looked into the reporting mechanism of farmers to government agencies. Using CNN and image processing, the application that detects rice pests and diseases was developed. The searching and comparison of captured images to a stack of rice pest images was implemented using a model based on CNN. Collected images were pre-processed and were used in training the model. The model was able to achieve a final training accuracy of 90.9 percent. Cross-entropy was low, which implies that the trained model can perform prediction or can classify images with low percentage of error. Through the developed application, farmers were provided with information and procedures on how to control and manage rice pest infestation. Future researchers may look into multiple pest comparison to a stack of images for faster retrieval of information.