论文部分内容阅读
为提高激光诱导击穿光谱技术(Laser-induced breakdown spectroscopy,LIBS)对鲜肉品种的识别率,采用支持向量机结合主成分分析算法辅助LIBS技术对鲜肉品种进行识别.对鲜肉切片用载玻片压平,采用LIBS技术对鲜肉组织(猪肉、牛肉和鸡肉)表面进行光谱数据的采集,每种鲜肉采集150幅光谱并进行随机排列,取前75幅光谱作为训练集建立模型,后75幅作为测试集测试建模结果.研究选取K、Ca、Na、Mg、Al、H、O等元素的49条归一化谱线数据进行主成分分析,并用所得数据建立支持向量机分类模型.结果表明,通过主成分分析降维,输入变量从49个优化减少到18个,模型建模速度从88.91 s降至55.52 s,提高了支持向量机的建模效率;并使预测集的平均识别率提高到89.11%.本研究为激光诱导击穿光谱技术在鲜肉品种快速分类领域提供了方法和数据参考.“,”To improve the classification accuracy of fresh meat species using laser-induced breakdown spectroscopy ( LIBS ) , the support vector machine ( SVM ) and principal component analysis ( PCA ) were combined to classify fresh meat species ( including pork, beef, and chicken) . A simple sample preparation to flatten fresh meat by glass slides was proposed. For each meat sample, 150 spectra were recorded and randomly arranged. The first 75 spectra were used to train a model while the others were used for model validation. By analyzing the 49 normalized spectral lines ( K, Ca, Na, Mg, Al, H, O, etc. ) in the different tissues, the classification model was built. The results showed that the dimensionality of input variables was decreased from 49 to 10 and modeling time was reduced from 89. 11 s to 55. 52 s using PCA, thus improving the modeling efficiency. The mean classification accuracy of 89. 11% was achieved. The method and reference data are provided for further study of fresh meat classification by laser-induced breakdown spectroscopy technique.