Apple color automatic grading method based on machine vision

Apples' color classification prior to the market can help determine the price range of the apples. In order to extract the color features of the apples, the acquired apples' image under the RGB model is transformed into the HIS model, and the threshold of the segmented image is determined by the histogram of the hue value of the statistical image. The hue values of apple fruit are divided into six ranges, and the frequency of different ranges is used as the input vector of neural network. Using LVQ neural network to divide apples into three levels. The experimental results show that the apple color grading accuracy based on LVQ neural network can reach 93%, which is slightly higher than SVM.