| Since the beginning of the 21 st century,China's economy has developed rapidly,the living standards of Chinese people have been continuously improved,and the demand of meat has also increased.However,in recent years,with the occurrence of food safety accidents such as water-injected meat and beef extract,people are more concerned about food quality and safety issues.Pork is the first staple of domestic meat,so the quality is particularly important at present.In this paper,the machine vision technology and image processing technology are used to detected and identified the quality of pork.The main research contents and results are as follows:(1)A machine vision system for pork quality detection was established.(2)After comparing the simulation results of the three filtering algorithms,we used the quantum-derived median filtering algorithm to remove the noise of RGB image.Then,the method of Maximum Classes Square Error in red component(OTSU)is used to segment the background of RGB image.(3)The fuzzy C-means Clustering(FCM)has strong dependence on the initial value,the effect of segmentation is not real.Therefore,in this paper,particle swarm optimization and kernel fuzzy c-means clusstering were used to segment the muscle and fat.The segmentation result is better than FCM,finally the mathematical morphology method is used to extract the largest connected muscle region.(4)The extracted pork color characteristics include: the mean and standard deviation of RGB;marbling characteristics include total fat area ratio,large fat particle area ratio,and small fat particle area ratio.The multivariate linear stepwise regression analysis method was used to establish the prediction model of color and marble grain level respectively.The test set samples were used for grade recognition.The correct judgment rate of color grade was92.16%,and the correct judgment rate of marble grain grade was 88.24%.(5)Developed a set of pork quality testing software system based on MATLAB. |
PDF Full Text Request