| 摘要: |
| 针对传统肉类新鲜度检测方法复杂、速度慢等问题，在试验室常温条件下采集猪肉反射率光谱数据，对肉类新鲜度的光谱检测方法进行研究，并分析不同建模方法以及不同样本采样间隔对预测精度的影响。结果表明：1）基于光谱角距离测度准则构建的加权模型，肉类新鲜度检测结果精度最高，决定系数（R2）为0.999 7，均方根误差（RMSE）为3.427 5；2）基于欧氏距离测度准则和三次多项式拟合法所构建的模型，其R2分别为0.998 1和0.998 1，RMSE分别为8.572 5和8.473 5。基于光谱角距离测度准则构建的加权内插模型，可以作为肉类新鲜度光谱快速检测模型使用，其检测精度在3种方法中最高。 |
| 关键词: 肉类新鲜度 光谱角距离 欧氏距离 光谱分析 光谱检测 |
| DOI：10.11841/j.issn.1007-4333.2019.11.15 |
| 投稿时间：2019-03-06 |
| 基金项目:湖南省大学生研究性学习和创新性试验计划项目（201710542032）；中国国家留学基金项目（201806725009） |
|
| Construction of meat freshness detection model based on spectral technology and comparative analysis on the error |
|
HU Shunshi, ZHANG Chenlu, CHEN Zihan, CHEN Yuchi, LI Xinyi
|
| (College of Resources and Environmental Sciences, Hunan Normal University, Changsha 410081, China) |
| Abstract: |
| Traditional meat freshness detection methods need longer time and complex procedure to get the results.In this study,the spectral reflectance data of pork were collected under normal temperature conditions in the laboratory,and the spectral detection methods of meat freshness were investigated.The effects of different modelling methods and sampling intervals on prediction accuracy were also analyzed.The results showed that:1) The weighted model based on spectral angular distance criterion had the highest accuracy,whose coefficient of determination (R2) was 0.999 7 and the root mean square error (RMSE) was 3.427 5;2) The weighted model based on Euclidean distance criterion and cubic polynomial fitting model are next to the above method,whose R2 were 0.998 1 and 0.998 1,and RMSE were 8.572 5 and 8.473 5,respectively.In conclusion,the weighted interpolation model based on spectral angular distance measurement could be used for rapid meat freshness detection with the highest accuracy among the above 3 methods. |
| Key words: meat freshness spectral angular distance Euclidean distance spectral analysis spectral detection |
