Nondestructive Detection Of Protein And Fat Content Of Tan Sheep And Mutton

Yanchi.Tan sheep is a dominant and characteristic animal breed in Ningxia.Since 2003,the breeding method of Tan sheep has been changed from grazing to drylot feeding,which leads to quality decline of Tan mutton and the loss of the original quality of meat tendeness,gamy light and unique flavor.The reason is that the protein content of Tan mutton is reduced and the fat content is increased.In order to solve this problem,the images of back fat and muscle depth of living Tan sheep were collected by B-ultrasonic technique,and the protein and fat content of Tan mutton were measured.The linear equation between back fat and fat content,muscle depth and protein content was constructed.Using visible/near infrared and near infrared hyperspectral collected spectral image of Tan mutton,combining with the measured values of protein and fat content,the prediction models of protein and fat content were built by pretreatment methods and characteristic wavelength extraction methods.which aims to make raiser adjust diet formula timely and make Tan sheep consume fat by sports in order to improve meat quality of Tan sheep and provide the high quality raw materials of mutton sheep or grade for raw material of meat,and to develope quality control technology of Tan mutton and rapid nondestructive detection system of cold fresh Tan mutton in order to provide theoretical guidance.The main research results are as follows:(1)A linear prediction models of protein and fat content in living Tan sheep were established based on B-ultrasonic technique.The protein content of linear equation was y=0.90x+12.61(y was the protein content of sample,x was the muscle depth)and Pearson correlation was 0.71.The fat content of linear equation was y=0.54x+0.23(y was the fat content of sample,x was back fat)and Pearson correlation was 0.98.(2)The prediction model of protein and fat content of Tan mutton based on visible/near infrared hyperspectral(400-1000 nm)was constructed.A 1-SG-IVISSA-PLSR prediction model for protein content of Tan mutton was established,with Rc and Rp values of 0.91 and 0.82 respectively;RMSEC,RMSEP and RMSECV values of 0.0064,0.0080 and 0.0094 respectively;Normlize-IVISSA-PLSR prediction model for fat content was established with Rc and Rp values of 0.93 and 0.75 respectively;RMSEC,RMSEP and RMSECV values of 0.0046,0.0096 and 0.0088,respectively.(3)The prediction model for protein and fat content of Tan mutton was constructed based on near-infrared hyperspectral(900-1700nm).A De-trending-VCPA-PLSR prediction model for protein content of Tan mutton was established,the Rc value and Rp value were 0.88 and 0.70,RMSEC value,RMSEP value and RMSECV value were 0.0851,0.0972 and 0.1260,respectively.The OS-PLSR prediction model of fat content of Tan mutton was established,and the Rc value and Rp value were 0.88 and 0.70 respectively.RMSEC value,RMSEP value and RMSECV value were 0.0060,0.0101 and 0.0126 respectively.By comparing the prediction models of protein and fat content of Tan mutton between visible/near infrared region and near infrared region,it showed that the model in the region of visible/near infrared was better than the model of near infrared region,which indicated that visible/near infrared region was more suitable for predicting protein and fat content of Tan mutton.