Main Article Content
Support vector machine (SVM) has been used to identify many plant diseases, but there is no relevant method and data for rice disease identification. In this paper, the computer image processing technology and support vector machine recognition method are used to analyze the characteristics of rice disease, the fast independent component method is used to process the collected rice disease image, the vector median filter is used to remove the noise, and then the statistical pattern recognition method and mathematical morphology are used to segment the disease. Finally, the texture feature parameters, the shape feature parameters and the color feature parameters of the disease rice image are extracted to identify the disease features using SVM. The research shows that the performance of SVM is better than neural network, and the accuracy is between 95% and 97%. The research shows that the method based on support vector machine can identify disease more quickly and accurately than using only single texture or color feature parameters. The research provides a theoretical basis and experimental method for rapid, efficient and nondestructive detection of rice diseases.