A judging method of rice milling degree based on the color characteristic and BP neural network

The paper has proposed a method to discriminate the rice milling degree based on color characteristic and BP neural network. A device for rice milling degree detection based on machine vision has been designed to collect images of rice; the rice images were treated by image processing techniques into acquisition as the target image; a circle of the radius R in the abdomen of the rice was determined to be a color feature extraction area and was divided into five concentric sub-domains by the average area; extracted the R, G, B color value of each sub-region and transformed them to H value as color feature values to describe the surface milling degree of rice; the 5 color feature values as input values were detected by BP neural network to judge the surface milling degree of rice. The experiment results showed that the average accuracy of the method could be 92.17% when beings used to discriminate the 4 types of rice of different milling degrees. ©, 2015, Editorial Department, Chinese Cereals and Oils Association. All right reserved.