Detection of Beef pH Using Visible and Near Infrared Spectroscopy Combined with Genetic Algorithm
Rapid and nondestructive detection of pH value of beef stored at 4℃ using visible and near infrared spectroscopy with genetic algorithm was investigated. A laboratory visible and near infrared spectroscopy system was established to collect 60 strip loin cuts samples reflectance spectra in the range of 4001700nm. The samples were stored at 4℃ for 1-18 days after 2 days postmortem. Different pretreatments such as Multiplicative Scatter Correction (MSC), SavitzkyGolay (SG) smoothing method were performed to the original reflectance spectra. The prediction model of multiple linear regression (MLR) , partial least squares regression (PLSR) and least square-support vector machine (LS-SVM) were constructed for prediction of beefs pH value with full-spectrum and effective wavelengths selected by genetic algorithm (GA) . The results showed that the MSC combined with SG smoothing was the best pretreatment, the performance of models established with effective wavelengths selected by GA were better than the fullspectrum models, and the best performance was achieved by PLSR model. PLSR model was able to predict beefs pH value with Rp=0.935 and SEP=0.101, prediction accuracy to that of full- spectrum PLSR model was improved by 29％ and 31％ respectively. The input variables of model were decreased by 87％. This study demonstrated that application of visible and near infrared spectroscopy combined with GA can rapidly and nondestructively evaluate the beefs pH value, and the prediction ability and prediction accuracy can be improved.
visible and near infrared spectroscopy genetic algorithm beef nondestructive detection pH value
Shibang Ma Yang Xu Xiaoyu Tian Xiuying Tang
College of Engineering China Agricultural University Beijing, China Nanyang Institute of Technology College of Engineering China Agricultural University Beijing, China
国际会议
哈尔滨
英文
700-704
2012-06-16（万方平台首次上网日期，不代表论文的发表时间）