摘要
羊肉中的水分含量直接影响着其加工、贮藏和口感,因此对其水分含量的检测具有十分重要的意义。在900~1700 nm的波长范围内采集南疆羊肉的光谱数据,采用一阶导数(first derivative,1-DER)、标准正态变换(standard normal transformation,SNV)、多元散射校正(multivariate scatter correction,MSC)、小波变换(wave transformation,WT)、SG平滑变换(Savitzky Golag smooth transformation,SG)、傅里叶变换(Fourier transform,FT)对原始光谱数据进行预处理。分别采用连续投影算法(successive projection algorithm,SPA)和竞争自适应重加权算法(competitive adaptive reweighted sampling,CARS)进行光谱特征选取,建立偏最小二乘回归(partial least squares regression,PLS)和支持向量机回归(support vector regression,SVR)模型对羊肉水分含量进行预测。结果显示,采用1-DER-CARS-SVR模型,选取参数c为0.7011,g为0.0884时,预测效果最佳,测试集的均方误差为1.2162,拟合优度为0.7395。研究结果为研发羊肉水分含量的无损检测装置提供理论参考。
Detection of moisture content of mutton is very important because it directly affects the processing,storage,and taste.The near-infrared spectrum data of mutton in south Xinjiang were collected in the wavelength range of 900-1700 nm.Noise among the spectrum data was removed to enhance the accuracy and robustness of the model by six pretreated methods,including first derivative,standard normal variate,multiplicative scatter correction,wavelet transforms,SG smoothing transform,and Fourier transform.Then,successive projections algorithm and competitive adaptive reweighted sampling algorithm were used to select characteristic variables respectively.Finally,partial least squares regression(PLS)and support vector regression(SVR)models were established to predict the moisture content of mutton.Results showed that the 1-DER-CARS-SVR model with parameters c was 0.7011 and g was 0.0884 which had the best performance.The mean square error of the test set was 1.2162,the goodness of fit was 0.7395.This provides theoretical references for the further study on the nondestructive testing detection device of moisture content of mutton in south Xinjiang.
作者
张立欣
杨翠芳
张晓
张楠楠
王亚明
ZHANG Lixin;YANG Cuifang;ZHANG Xiao;ZHANG Nannan;WANG Yaming(College of Information Engineering,Tarim University,Alaer 843300,China;School of science,Nanjing University of Science and Technology,Nanjing 210094,China)
出处
《食品与发酵工业》
CAS
CSCD
北大核心
2022年第12期255-260,共6页
Food and Fermentation Industries
基金
国家自然科学基金(31960503)
国家自然科学基金(61662064)
塔里木大学校长基金(TDZKSS202006)
塔里木大学校长基金(TDZKQN201709)。
关键词
近红外光谱
连续投影算法
竞争自适应重加权算法
偏最小二乘回归
支持向量基回归
near infrared spectrum
successive projection algorithm
competitive adaptive reweighted sampling
partial least squares regression
support vector regression
