基于可控气流激光检测技术的鸡肉嫩度评估方法  

Evaluation of Chicken Tenderness Based on Controlled Air-flow Laser Detection Technique

作　　者：徐虎博 赵庆亮[2] 何珂 李永玉[1] 彭彦昆[1] 汤修映[1] XU Hubo;ZHAO Qingliang;HE Ke;LI Yongyu;PENG Yankun;TANG Xiuying(College of Engineering,China Agricultural University,Beijing 100083,China;Chinese Academy of Agricultural Mechanization Sciences,Beijing 100083,China)

机构地区：[1]中国农业大学工学院,北京100083 [2]中国农业机械化科学研究院,北京100083

出　　处：《农业机械学报》2020年第S02期457-465,共9页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金面上项目(31571921);北京市自然科学基金面上项目(6202020)。

摘　　要：以鸡肉嫩度为研究对象,采用可控气流激光检测技术的瞬态、蠕变回复和应力松弛等动静态检测模态,并使用支持向量机分类器和全局变量偏最小二乘算法,结合不同预处理方法,对鸡肉嫩度进行定性判别和定量预测。结果表明3个激励模态结合不同预处理算法均可实现鸡肉嫩度的定性定量评估。在定性方面,瞬态模态对嫩度具有最佳的分类效果;SG卷积平滑算法表现出最佳的预处理性能,校正集嫩/老分类精度分别为1和0.98,马修斯相关系数为0.97;而验证集分类精度也达到了0.95和0.84。在定量预测方面,SG卷积平滑算法在提升原始数据的信噪比上同样具有最佳效果;瞬态模态校正集和验证集模型相关系数分别为0.948和0.913,均方根误差分别为0.736 N和1.013 N。因此,在组织结构引起的品质预测动态模态较静态模态更适用。本研究开展的可控气流激光技术在鸡肉嫩度评估的应用,为肉品检测领域提供了新的解决方案。The air flow laser fusion technique has the characteristics of non-contact and non-destructive.A controlled air flow laser detection(CAFLD)method was proposed,which was based on the high-precision detection of micro deformation by laser,flexible on-off control and non-contact of the air flow.Five components were included in the air flow laser detection platform:laser ranging system,air force generation system,lifting testing bed system,force sensing system,and control and information processing system.The feasibility of chicken breast tenderness detection by using the CAFLD technique was explored.Three modes:transient(dynamic),creep-recovery(static)and stress relaxation(static)were adopted.The support vector machine and global variable partial least square algorithm were used to qualitatively identify and quantitatively predict the tenderness of chicken breast.The results demonstrated that the three modes combined with different preprocessing algorithms could carry out the qualitative discrimination of chicken tenderness,in which the transient mode had the best classification effect compared with the static modes.SG convolution smoothing algorithm showed the best preprocessing performance.The classification accuracy(tender or hard)of the calibration set was 1 and 0.98,respectively,the Matthews correlation coefficient was 0.97;the classification accuracy(tender or hard)of the verification set was up to 0.95 and 0.84.For quantitative prediction of chicken tenderness,the SG convolution smoothing algorithm was the optimum on improvement of the signal-to-noise ratio.The transient had the best prediction effect,the correlation coefficients of calibration set and validation set were 0.948 and 0.913,respectively,the root mean square error was 0.736 N and 1.013 N,respectively.Because tenderness was the quality of meat which was shown by the difference muscle fiber structure.it can be inferred that the dynamic mode was more suitable than the static mode in predicting the quality caused by tissue structure.The application o

关 键 词：鸡肉 嫩度 气流 激光 无损检测 

分 类 号：TS251.5+5[轻工技术与工程—农产品加工及贮藏工程]