【摘要】：結合激光誘導擊穿光譜和線性回歸對豬肉中Cr元素進行定量分析,比較了Cr特征譜線單變量分析與多變量分析對豬肉中Cr元素檢測精度的影響。先采用特征譜線Cr I425.43nm進行單變量分析,得到Cr濃度與LIBS強度的線性相關系數(shù)為0.9434;再利用Cr的三特征譜線Cr I425.43nm,Cr I427.48nm,Cr I428.97nm進行多變量分析,研究了Cr元素預測質量分數(shù)與實際質量分數(shù)之間的線性相關性。結果顯示,單變量分析與多變量分析的線性相關性分別為0.9749和0.9769,定標集平均預測相對誤差分別為4.8%和4.2%,驗證集AREP分別為5.3%和3.3%。結果表明,選取目標元素Cr的單特征譜線與多特征譜線均能對樣品Cr含量進行有效預測,多變量分析能在一定程度上提高Cr元素含量預測的準確性但與單變量比較差異性不大,且采用單一最靈敏線定量分析更簡便可行。
[Abstract]:Combined with laser induced breakdown spectroscopy and linear regression, the quantitative analysis of Cr in pork was carried out, and the effects of single variable analysis and multivariate analysis on the detection accuracy of Cr in pork were compared. The linear correlation coefficient between Cr concentration and LIBS strength is 0.9434 by using characteristic spectral line Cr I425.43nm, and then multivariate analysis is carried out by using Cr I 425.43 nm ~ (-1) Cr I _ (427.48) nm ~ (m) I428.97nm, and Cr I _ (425.43) nm ~ (m) ~ (-1) Cr I428.97nm. The linear correlation between the predicted mass fraction of Cr element and the actual mass fraction was studied. The results show that the linear correlation between univariate analysis and multivariate analysis is 0.9749 and 0.9769, the average relative error of calibration set is 4.8% and 4.2%, and the AREP of verification set is 5.3% and 3.3%, respectively. The results show that both single and multiple characteristic lines of the target element Cr can effectively predict the Cr content of the sample. Multivariate analysis can improve the accuracy of the prediction of Cr content to some extent, but there is little difference compared with the single variable. And the single most sensitive line quantitative analysis is more convenient and feasible.
【作者單位】： 江西農業(yè)大學工學院;江西省高校生物光電及應用重點實驗室;江西農業(yè)大學生物科學與工程學院;
【分類號】：TN249
，

本文編號：2158453