应用低场核磁共振技术检测注水肉的探讨

Discussion on Detection of Water-injected Meat Using Low Field Nuclear Magnetic Resonance Technique

采用CPMG序列检测不同注水比例肉样品产生的核磁共振信号,运用反演软件得到每个肉样品的横向弛豫谱及相关参数。选择总峰面积S,自由水比例P_(23)和自由水峰终点对应的横向弛豫时间T_(23e)作为预测变量,对试验数据进行判别分析,建立检测注水肉的Bayes模型。采用回代验证法验证的正确率为94.3%,采用交叉验证法验证的正确率91.4%。该模型可以为肉品监管部门建立快速、准确区分和检测不同注水比例注水肉的方法提供依据。

The CPMG sequence was used to detect the nuclear magnetic resonance（NMR）signals from meat samples with different water-injection ratios. The transverse relaxation spectra and the relative parameters of each meat sample were obtained by inverse software. The total peak area（S）,free water ratio（P23）and transverse relaxation time（T23e）corresponding to the end point of free water peak were selected as predictive variables. The experimental data were analyzed by discriminant analysis,and the Bayes model was established for the detection of water-injected meat. The correct rate of verification by resubstitution verification was 94.3% and that by cross-verification was 91.4%. The model could provide experimental data support for the establishment of a fast and accurate method to detect the water-injected meat with different ratios.