| China is one of the biggest meat production countries in the world.The meat represented by pork and beef is rich in fat,protein,minerals and vitamins,so it is a kind of important food which is indispensable to people's daily life.But meat food safety in our country is a little serious for that meat poisoning events occur frequently.Meat quality and security problems have increasingly become an extremely important security problem in agricultural products market in China,and they directly affect people's life,health and basic consumption safety.Freshness is the direct reflection of meat rottenness metamorphic grade.It could comprehensively reflect the nutritional,safety and reliability of the meat,thus it is one of the important indicators of the meat quality evaluation.Traditional detection methods of meat freshness including sensory evaluation,physical and chemical experimental analysis methods with the complicated process,low speed,low efficiency,and samples destructive,could not meet the requirement of the rapid developing production and economy.Therefore,developing a rapid,nondestructive,real-time and accurate pork freshness detect method is of great significance.Spectral imaging technology is a new platform technology that it is the integration of the image analysis technology and the spectrum analysis technology,widely applied in agricultural products detection field.Since multispectral imaging technology involves a few wavelengthes,and the selected wavelengthes are characteristic wavelengthes of detection indicators,multispectral imaging data processing speed is faster,and the technology is more targeted,more practical,lower cost than visible/near infrared spectroscopy technology.This research used multispectral imaging techniques to detect pork freshness.In order to study the relationship between the light scattering characteristics and the freshness of pork,an experimental multispectral imaging system was developed to carry out.The system consisted of a halogen tungsten lamp light source unit,image acquisition unit,a data processing unit,afilters,a black box and a loading station.In this research,pork color(L*,a*,b*),pH value and total volatile basic nitrogen(TVB-N)were selected as the evaluation indexes of pork freshness.Prediction models were established beween pork color parameters(L*,a*,b*),pH value,TVB-N and multispectral image scattering information respectively by multiple linear regression(MLR)method combined with stepwise regression method,and discriminant model of pork freshness grade was developed by linear discriminant analysis method.Prediction correlation coefficient of validation set(Rv)for pork TVB-N content was 0.93 with prediction standard error(SEP)of 1.68 mg/100g,and the ratio of standard deviation(SD)to root mean square error of prediction(RPD)for TVB-N content was 2.74;Rv of pH value was 0.90,SEP was 0.03,and RPD was 2.61;for b*,the Rv was 0.94 with SEp of 0.49 and RPD of 2.76;the Rv of L*was 0.91 with SEP of 0.73 and RPD of 2.37;the Rv of a*was 0.93,SEp was 0.72,and RPD was 2.65.The accuracy of pork freshness level discriminant model was more than 85%.Research result showed that using multispectral scattering imaging technology is feasible for detection of raw pork freshness.According to the experimental research and actual production needs,a portable pork freshness fast nondestructive detction device was developed,and a correspondin g software system was also devoped.After the hardware and software system debugging of the device,the detection speed of the device is 4s pur sample,and the accuracy is more than 85%.Using multispectral scattering imaging technology is feasible for detection of raw pork freshness.Development of the corresponding portable detection device helps the technology popularization and application in the field of food and agricultural products detection,and provides a theoretical basis and technical support for more meat indicators testing at the same time using the technology in future. |
PDF Full Text Request