猪肉中挥发性盐基氮含量光谱检测模型的修正方法.pdf

第4卷第3期
食品安全质量检测学报
VOL 4 No. 3
2013年6月
Journal of Food Salcly and Quality
um..2013
猪肉中挥发性盐基氮含量光谱检測模型的
修正方法
赵政,李小显”,刘洁,文东东,刘娇
(华中农业大学工学院,武汉430070)
摘要:目的研究猪肉新鲜度指标挥发性盐基氮(TVBN)含量检测模型修正方法,以提高光谱校正模型对不
同品种猪肉样品的适用性。方法建立基于偏最小二乘回归(PLSR)的杜长大猪肉TVB-N模型,采用光谱信号
补正与模型更新两种方法对该模型进行修订,比较修正后杜长大模型对恩施山猪样本的预測效果。结果建立
的杜长大猪肉样本模型预测決定系数R2,为088.预測标准差RMSP为1.792,将此模型用于预测恩施山猪
TVB-N值、R2,为0552, RMSEP为4.733修正后的杜长大模型预测恩施山猪TVB-N值时,R,分别提高到0.964
和0.943, RMSEP分别降低为1.329和1.885。结论光谱信号补正和模型更新方法均能有效改善模型预测性能
提高模型适应性
关键词:模型修正;猪肉;挥发性盐基氮;高光谱图像技术;偏最小二乘回归
Correction methods of pork total volatille basic nitrogen content detection
model based on hyperspectral imaging technology
ZHAO Zheng, LI Xiao-yu, LIU Jie, WEN Dong-dong, LIU Jiao
(College of Engineering, l/azhong Agricultural Universit, Wuhan 430070, China)
ABSTRACT: Objective lo sludy correction methods for pork freshness(tvb-n)detection model of dil-
ferent species based on hyperspectral imaging technology and improve the generality of the calibration model
Methods Du changda model was cstablished bascd on partial least squares regrcssion using Du changda
mountain boars as samples. Model updating by adding new typical samples and spectral correction based on
model regression coefficient were adopted to improve the model applicability of the calibration model for I nshi
mountain boars. Results The TVB-N content model, with 0.884 as the coefticient of detcrmination in prediction
set (R, )and 1. 792 as the root mean squared error of prediction (RMSEP), was used to predict the Enshi moun
tain boars, and R, and RMSHP were 0.552 and 4. 733, respectively. While the R,increased to 0.964 and 0.943
and the RMST P decreascd to 1.329 and 1.885 using calibration modcl. Conclusion B3oth methods can improve
the predict performance of model effectively, and enhance the model adaptation
KEY WORDS: model correction methods; pork; total volatile basic nitrogen, hyperspectral imaging technology
partial least squares regression
基金项目:公益性行业(农业)科研专项(201003008)、国家自然科学基金青年基金项目(61205153)
Fund: Supported by the Special Fund for Agro-scicntific Rcscarch in the Public Intcrest(201003008)and the National Natural Scicncc Founda-
tion of China (61205153
通讯作者:李小昱,教授.博士生导师,主要研究方向为农产品无损检测。F-mail: lixiaoyu(@ mail hau.edu. en
Corresponding author II Xiao-yu, Professor Ih. D. Supervisor, Collcgc of Enginccrin, Huazhong Agricultural Univcrsity, No. 1, Shizishan
Strect, Wuhan 430070, China. E-mail: lixiaoyu( mail hzaucdu. cn