Identification of Rice Storage Quality Based on Computer Vision

Abstract

Computer vision technology was introduced in this paper to discriminate the rice storage quality. The rice images of ten categories were obtained by the computer vision hardware device, and the total number of images was 160, of which 100 were used as training samples and 60 as test samples. The 20 feature parameters of color and texture features were extracted from all the samples, which were classified by the BP neural network, and probabilistic neural network (PNN) respectively, and the rate of identification was 85%, 91.67%. Compared to the BP neural network, PNN has better classification results. The experimental results showed that the method based on color and texture feature is effective in the identification of rice storage quality.