The Water-injected Beef Internal Water Property Analysis And Nondestructive Detection Technique Application

The meat of livestock,particularly beef cattle,is sold at a high price.To obtain higher profits,unscrupulous practices regarding this kind of meat are quite common.At present,most of these practices can be detected based on their indications,as non-destructive examination techniques have become increasingly widespread.However,there are still some unprincipled practices unable to be detected,for example,water injection into meat.This detection inability is largely due to a lack of defined key indicators for water-injected meat.Specifically,the study on the biological and physicochemical properties of water-injected meat is insufficient,and the essence of water-injected meat in current research is misunderstood.In view of this situation,this research will analyze the physical properties of water-injected meat through designed experiments and data analysis,to clearly define the properties of water-injected meat.On this basis,the key indicators for such meat are proposed as a reference for quantitative research on the differences between water-injected meat and normal meat,laying a foundation for the application of nondestructive examination techniques to the detection of water-injected meat.Therefore,the research,which focuses on key indicators that help distinguish normal meat from water-injected meat,provides a great theoretical innovation and has remarkable practical significance and application values.To sum up,the research is carried out around normal beef and water-injected beef,and mainly covers qualitative study and quantitative analysis of the moisture changes and their results of water-injected meat.It puts forward key indicators that reflect and distinguish normal meat from water-injected meat.Based on the key indicators,the research discusses the application of nondestructive examination techniques for water-injected meat.The research involves the following aspects.Firstly,specialize and enough data were obtained from some experiment such as dropping water loss experiment,water-holding capacity,drying experiment and so on.The experiment was obeyed by national standards such as national standard of Permitted level of moisture in meat of livestock and poultry,National food safety standard of Determination of moisture in foods.The data of moisture content which obtained from the experiment analysis result was showed the difference saliency between normal beef and water-injected beef.In this way,it proved that the conventional measurement was not suitable for water-injected beef.Furthermore,a novel definition for water-injected meat should be confirmed.It was a better distinction between factually water-injected meat and the manufactured one in laboratory.On the other hand,the parameter of moisture content measured was also an important explaining method for water-injected meat which based on current study and technology level.However,another type of rich describer was also very necessary for moisture content concept.In this study,moisture content was a synthesis concept which including many features of internal construct,moisture physical properties,status variation and the different status change of water's ratio.All of which were obtained by experiment with setting conditions of 28 degree Celsius and vacuum.As the result of data analyzed,the major feature of meat moisture which convenient dividing normal beef and water-injected one were deduced according to a ratio value.The value was moisture content ratio of such two types of meat.Secondly,the moisture content of national standard was measured and named class ?,another moisture content which under certain conditions was also measured and named class ?,the ratio value of? and ? was named relative value of moisture content.It was known that different value has significant difference after the calculation.This conclusion provided a certain research route for separating water-injected meat from normal one.The principle of statistics was applied for establishing foundation of data analysis.160 normal meat samples and 160 water-injected meat samples were used in the experiment for getting rate of moisture content values.In the same way,the confidence interval was set for testing meat sample's feature under the condition of significance level of 0.1.In another word,the confidence interval value was the quantifying basement in water-injected research.Accordingly,confidence intervals could be set up for normal meat and water injected meat.The results indicated that the relative value of moisture content was reliable enough to distinguish two kinds of meat products.This key parameter was of significance and had obvious advantage for the study of the problem of water-injected meat.Taking the relative moisture content as the observation value,the spectral data were collected from the two kinds of meat in the near infrared band of 1025-2180nm as the variable matrix,and the partial least squares method were used to establish the regression analysis model,and the nondestructive testing technology of the water injected meat was discussed.In this process,the focus was on the optimization of the spectrum.Full wave modeling,full band extraction of spectral feature modeling,partition band extraction of spectral feature modeling was tried to be used,which included the continuum removal method,Wavelet Transform method,K clustering based on particle swarm optimization and so on.The results showed that the prediction effect of normal meat in class ? was better,while the prediction effect of water-injected meat was better in the relative value of water content.The best result of regression analysis showed that prediction set's correlation coefficient and standard deviation:for normal meat model they were 0.7933 and 0.0306,and for the water-injected model they were 0.8096 and 0.0299.In the actual test of the model with 26 new samples,the correlation coefficient and standard deviation between the actual value of the sample and the predicted value of the model were showed here:For normal meat model they were 0.8023 and 0.0288,for the water-injected model they were 0.8188 and 0.0295.Based on the actual test results,if the confidence interval was constructed strictly under the condition of alpha =0.1,it was estimated that the recognition rate of the two meat models can reach 76.92%,but when the interval was not strictly limited,recognition value was 88.46%.Furthermore,the method of SVM classifier had a certain advantage to identify two different samples,and it was widely used at present.Then wavelet transform was applied in SVM model for selecting the spectral values of two kinds of meat products.At last,the total recognition rate of the 26 new samples of the two meat products was 76.92%.