Development of Classification System of Rice Disease Using Artificial Intelligence

Currently, rice diseases caused by various weather disorders and pests are serious problems. In this research, we will develop a system to classify healthy rice plants and diseased rice plants by processing images from rice planted in paddy field. Symptoms appear clearly in rice diseases, we focused on their color information. The 256 level intensities of Red, Blue, Green of a pixel are employed as a feature amount, and a classifier using SVM was constructed. In addition, we reduced the learning time by using principal component analysis, and obtained accuracy exceeding 90%.