摘要
建立土壤重金属含量的X射线荧光(X-ray Fluorescence, XRF)定量模型,为土壤重金属快速测量提供方法。采集87个土壤样品的光谱信号,通过BOSS算法建立土壤重金属含量的BOSS-PLS预测模型,在此基础上为进一步优化模型,使用ICO算法对光谱进行变量筛选,保留强信息变量与弱信息变量,建立ICO-BOSS-PLS模型,并与常规波长选择算法建立的预测模型进行对比分析。结果表明,ICO-BOSS-PLS模型的预测结果要优于常规波长选择算法的建模,其中,ICO-BOSS-PLS模型预测的结果,其重金属的相关系数R2达到0.97以上,均方根误差达到25。综上,使用ICO-BOSS-PLS建模是一种有效的XRF光谱定量建模方法,为土壤重金属快速检测模型的建立提供了一定的技术支持。
An X-ray Fluorescence(XRF) quantitative model of soil heavy metal content was established to provide a method for rapid measurement of soil heavy metals. The spectral signals of 87 soil samples were collected, and the BOSS-PLS prediction model of soil heavy metal content was established by the BOSS algorithm. On this basis, in order to further optimize the model,the ICO algorithm was used to screen the spectrum variables, and the strong and weak information variables were retained. The ICO-BOSS-PLS model was established and compared with the prediction model established by the conventional wavelength selection algorithm. The results show that the prediction results of the ICO-BOSS-PLS model are better than the modeling results of the conventional wavelength selection algorithm. 25. In conclusion, the use of ICO-BOSS-PLS modeling is an effective XRF spectral quantitative modeling method, which provides certain technical support for the establishment of a rapid detection model for heavy metals in soil.
作者
陆旻波
LU Minbo(China Three Gorges University,Hubei Yichang 443000)
出处
《长江信息通信》
2022年第5期5-7,共3页
Changjiang Information & Communications
