Grape Maturity’s Nondestructive Detection Research

In recent years, the demand is growing quickly for fresh grapes andits most important byproduct—wine in our country. At the same time, moreand more consumers are concerned about t he quality of fresh grapes and wine.In order to improve the quality of grape and wine, domestic grape and wineindustry has invested large amounts of funds for research. The research hasfound that the grape mature degree is an important factor to determine thequality of grape and wine. So how to determine the mature degree of thegrapes quickly and effectively is what is needed to solve the technical problemnow.This paper attempts to use nondestructive detection technology to detectthe Jinhua Jufeng grape mature degree. Studies have shown that the color ofgrapes’ skin reflects the mature degree of grapes’ inner pulp. So it canconsider the feature value of grape skin color as the basis of judging grapemature degree and classifly grape mature degree.The core content of thispaper is as follows:1) Train on the129string Jufeng Grape samples of different growth stateby using C-means clustering algorithm. Combinating with changes in theirskin color in the growth process of Jufeng grapes, it determines7groupclassification benchmark. Develop a grape mature degree detection system byusing VC++6.0on a computer. It uses an image transducer to acquire detectedgrape color images and extracts color feature values (H、S、I) of detected grapecolor images’ target region (Grape region in the image). Calculate euclideandistance between detected grape color feature and the7group classificationbenchmark.2)5kinds of green,5kinds of yellow and5kinds o f purple artificiallyselected are classified using the support vector machine algorithm to designcolor classifier, each color as a class. It focuses on purple classifiers forclassification of nearly mature grape. Put the color feature values (R、G、B) ofdetected Jufeng grapes into designed purple classifiers and compare the output values of the5designed purple classifiers and the detected Jufeng grapes willbe classified as the class of maximum output value.3) Make a small portable grape mature degree detection device to acquirefreshly picked grapes image and extract color feature value of freshly pickedgrapes and judge their mature degree. This device is developed in ARMembedded development board and it is made of image acquisition equipmentand image processing chip and data display equipment. The display equipmentof the device can display the data of grapes’ color feature value in real timeand the result of judgment for grapes’ maturity.The final results show that, the method of judging the Jinhua Jufeng grapematurity is feasible and effective.