高光谱成像技术定量可视化检测熟牛肉中挥发性盐基氮的含量

发布时间：2018-02-28 06:36

本文关键词： 高光谱成像熟牛肉挥发性盐基氮特征提取可视化　出处：《现代食品科技》2017年09期 　论文类型：期刊论文

【摘要】：为了能够快速、准确的检测出熟牛肉在冷藏过程中的新鲜状况,尝试利用高光谱成像技术对熟牛肉中的挥发性盐基氮(TVB-N)含量进行定量可视化分析。采集400~1000 nm范围内样品高光谱图像,采用变量组合集群分析法(VCPA)提取出6个光谱特征波段变量,针对特征波段图像,利用Tamura算法共提取出18个纹理特征变量,基于RGB颜色模型,分别计算出R、G和B分量图中共9个颜色特征变量。利用粒子群优化最小二乘支持向量机(PSO-LS-SVM)算法分别建立了不同变量组合的TVB-N含量预测模型。经分析比较,基于光谱与颜色特征融合的PSO-LS-SVM模型展现出最优的预测能力,预测集决定系数(R2p)和均方根误差(RMSEP)分别为0.955和1.093。利用最优模型将TVB-N含量进行可视化表达。结果表明,融合高光谱图像中光谱与颜色特征并结合PSO-LS-SVM算法对熟牛肉中TVB-N含量进行准确的预测与可视化表达是可行的,该研究可为其它肉及肉制品新鲜度检测提供理论参考。
[Abstract]:In order to quickly and accurately detect the fresh condition of cooked beef during cold storage, This paper attempts to use hyperspectral imaging technology to quantitatively visualize the volatile base nitrogen (VB-N) content in cooked beef. The hyperspectral images of samples in the range of 400 nm are collected and six spectral characteristic band variables are extracted by variable combination cluster analysis (VCPA). For the feature band image, 18 texture feature variables are extracted by using Tamura algorithm, which is based on RGB color model. Nine color characteristic variables of RGG and B component diagrams were calculated respectively. The TVB-N content prediction models with different combinations of variables were established by using PSO LS-SVM algorithm. The PSO-LS-SVM model based on the fusion of spectral and color features shows the optimal prediction ability. The prediction set determination coefficient (R2p) and root mean square error (RMS EPV) are 0.955 and 1.093, respectively. The TVB-N content is visualized by the optimal model. The results show that the prediction set determination coefficient (R2p) and the root-mean-square error (RMS) are 0.955 and 1.093, respectively. It is feasible to accurately predict and visualize the content of TVB-N in cooked beef by combining the spectral and color features of hyperspectral images with PSO-LS-SVM algorithm. This study can provide a theoretical reference for the detection of freshness of other meat and meat products.
【作者单位】：北京农业质量标准与检测技术研究中心;沈阳农业大学信息与电气工程学院;
【基金】：国家科技支撑计划项目(2014BAD04B05-2)
【分类号】：TP391.41;TS251.52

【相似文献】