In contrast to conventional methods for the determination of meat chemical composition ,near infrared reflectance spec-troscopy enables rapid ,simple ,secure and simultaneous assessment of numerous meat properties .The present review focuses on the use of near infrared reflectance spectroscopy to predict meat chemical compositions .The potential of near infrared reflectance spectroscopy to predict crude protein ,intramuscular fat ,fatty acid ,moisture ,ash ,myoglobin and collagen of beef ,pork ,chick-en and lamb is reviewed .This paper discusses existing questions and reasons in the current research .According to the published results ,although published results vary considerably ,they suggest that near-infrared reflectance spectroscopy shows a great po-tential to replace the expensive and time-consuming chemical analysis of meat composition .In particular ,under commercial con-ditions where simultaneous measurements of different chemical components are required ,near infrared reflectance spectroscopy is expected to be the method of choice .The majority of studies selected feature-related wavelengths using principal components re-gression ,developed the calibration model using partial least squares and modified partial least squares ,and estimated the predic-tion accuracy by means of cross-validation using the same sample set previously used for the calibration .Meat fatty acid composi-tion predicted by near-infrared spectroscopy and non-destructive prediction and visualization of chemical composition in meat u-sing near-infrared hyperspectral imaging and multivariate regression are the hot studying field now .On the other hand ,near in-frared reflectance spectroscopy shows great difference for predicting different attributes of meat quality which are closely related to the selection of calibration sample set ,preprocessing of near-infrared spectroscopy and modeling approach .Sample preparation also has an important effect on the reliability of NIR prediction ;in particular ,lack of homogeneity of the meat samples influenced the accuracy of estimation of chemical components .In general the predicting results of intramuscular fat ,fatty acid and moisture are best ,the predicting results of crude protein and myoglobin are better ,while the predicting results of ash and collagen are less accurate .%和传统的肉品化学成分检测方法相比，近红外光谱分析是一种能够快速、简单、安全和可同时测定多种化学成分的肉品检测方法。论述了近年来猪、鸡、牛、羊四种动物肉中粗蛋白质、肌间脂肪、脂肪酸、水分、灰分、肌红蛋白和胶原蛋白七种化学成分的近红外光谱分析检测研究进展，并探讨了目前研究中存在的问题及其原因。从已发表的研究结果可以看出，近红外光谱分析具有替代耗时成本高的现有肉品化学检测方法的巨大潜力，特别是用于商业上同步检测多种化学成分。研究大多采用主成分分析法进行特征光谱筛选，偏最小二乘和修正偏最小二乘法建立校正模型，交互验证进行预测结果评价。目前研究热点集中在肉中脂肪酸含量预测模型和应用近红外高光谱图像和多元回归预测肉中化学成分。近红外光谱预测结果差异性较大，这与样品集选择、光谱预处理和建模方法都有很密切的关系。样品选择对近红外预测精度具有重要的意义，非均质肉样影响了近红外化学成分预测精度。总的来说，肌间脂肪、脂肪酸与水分指标预测效果较好，粗蛋白质和肌红蛋白指标预测效果次之，灰分和胶原蛋白指标预测结果较差。
展开▼
