论文部分内容阅读
目的通过采集鲢鱼的近红外光谱数据和测定鱼肉营养成分含量探索鲢鱼营养成分的快速分析方法。方法采集254个鲢鱼鱼肉样品的近红外光谱数据,经过多元散射校正、正交信号校正、数据标准化等20种方法预处理,在1000~1799 nm光谱范围内,结合化学实测值分别采用偏最小二乘法、主成分分析和BP人工神经网络技术、偏最小二乘法和BP人工神经网络技术建立鲢鱼营养成分近红外定量模型。结果鲢鱼鱼肉粗蛋白含量为12.05%~19.05%,粗脂肪含量为0.24%~5.27%,水分含量为72.62%~80.58%,灰分含量为0.46%~1.50%,数据范围较大,可满足建模要求。在3种建模方法中,近红外光谱数据结合偏最小二乘法建立的鲢鱼营养成分模型最优,所得的粗蛋白、粗脂肪、水分和灰分的近红外定量模型的相关系数分别为0.9969、0.9925、0.9831和0.9976。结论采用近红外光谱数据和偏最小二乘法建立的模型具有较好的预测能力,能较为准确、快速地分析出鲢鱼鱼肉粗蛋白、粗脂肪、水分和灰分的含量。
OBJECTIVE: To analyze the nutrient contents of silver carp by means of near-infrared spectral data of silver carp and the determination of nutritional content of silver carp. Methods Near infrared spectroscopy data of 254 silver carp fish samples were collected and pretreated with 20 methods of multivariate scatter correction, orthogonal signal correction and data standardization. The results were in the range of 1000 ~ 1799 nm with the minimum value Two-component method, principal component analysis and BP artificial neural network technology, partial least-squares method and BP artificial neural network technology to establish a quantitative model of nutrients in silver carp near-infrared. Results The crude protein content of silver carp was 12.05% ~ 19.05%, the crude fat content was 0.24% ~ 5.27%, the moisture content was 72.62% ~ 80.58%, the ash content was 0.46% ~ 1.50% Mold requirements. Among the three modeling methods, the model of nutrient composition of silver carp was established by near-infrared spectroscopy combined with partial least squares. The correlation coefficients of crude protein, crude fat, moisture and ash near-infrared quantitative model were 0.9969, 0.9925, 0.9831 and 0.9976. Conclusion The model established by near-infrared spectroscopy and partial least-squares method has good predictive ability and can accurately and quickly analyze the crude protein, crude fat, moisture and ash content of silver carp.