Research On The Detection Method Of Appearance Quality Of Germ Rice Based On Embedded Technology

Rice is one of the three major grains in the world and the staple food of most Chinese people.For a long time,the rice people eat is basically pure white rice with low nutritional value-a kind of rice that basically only keeps endosperm.germ rice is a kind of rice that retains more than 80% of the plumule in the rice production process.Compared with the ordinary white rice,germ rice has higher nutritional value.The detection of processing quality of germ rice is to divide the germ rice into different quality categories according to the relevant detection standards,and to detect some specific quality indicators such as chalkiness degree,rice plumule ratio,etc.The traditional manual detection method has the disadvantages of strong subjectivity,easy to be affected by the environment and can not work continuously,while the detection method based on machine vision technology is more objective and consistent and can work continuously.In this article,according to the characteristics of germ rice which is different from refined white rice,I design the quality detection system of germ rice processing based on embedded machine vision technology.First of all,this article studies the image acquisition and preprocessing methods of rice embryo.According to the requirement of image quality,the image acquisition device of rice germ was designed.According to the requirements of the follow-up image processing,the preprocessing methods of rice germ image are studied,such as Gaussian filter,median filter,morphological filter and other image denoising methods;gray level piecewise linear transformation,power-law transformation and other image enhancement methods;image rotation and other image geometric transformation methods.On the basis of image preprocessing,this article studies the segmentation algorithm of germ rice background and adhesion rice.A morphological segmentation algorithm based on region growth is proposed for sticky rice segmentation.Then,according to the characteristics of germ rice,the detection algorithm of chalkiness,chalkiness rate and rice plumule ratio is designed and implemented.The difference between germ rice and ordinary white rice is that most rice grains carry germ.In the process of chalky region segmentation,the germ region is often divided together.For the problem of chalkiness segmentation,this paper proposed a new chalkiness extraction algorithm;for the problem of embryo rate,this article designs and uses the embryo segmentation algorithm based on color feature,the embryo position recognition algorithm and the germ rice recognition algorithm based on area feature.Finally,the software and hardware design of the quality inspection system of rice germ processing based on embedded technology is completed.The hardware structure design realizes the whole structure and the mutual connection between the modules.The software structure design configures the embedded Linux operating system environment and the transplantation of Open CV library,and completes the architecture design and implementation of the quality inspection program for germ rice processing.