Maturity Analysis of Citrus Based on Machine Vision and Android Mobile Platform Technology

Harvest maturity is an important factor affecting fruit storage quality. In order to realize rapidly and nondestructively discriminate to citrus maturity, the rapid and non-destructive testing of citrus maturity by using machine vision and Android mobile platform technology was studied in this paper. Firstly, the citrus images in the field were captured by the camera provided on the mobile phone. Secondly, the RGB values of the citrus color images were gray-scaled according to the ratio of 0.30: 0.59: 0.11. Thirdly, the images were segmented by the adaptive Ostu algorithm to get the binaryzation image. Fourthly, the binaryzation image was executed AND operation with the color source image to obtain a final image. Finally, the maturity formulas was used to separately calculate the percentage of pixels of the H and S components in the mature image of the citrus region to the whole citrus region in the final image. This process was implemented on the Android phone, through the 150 citrus image maturity analysis, the accuracy rate was attained to 90.7%. The test results showed that the method has good feasibility and can provide technical support for rapid and quantitative detection of citrus maturity.