基于NURBS曲线的ICP-AES光谱基线校正研究

Research on ICP-AES Spectral Baseline Correction Method Based on RBF Neural Network and NURBS Curve Model

在ICP-AES测量系统中,光谱基线漂移会引起元素浓度定量分析误差,所以在测量过程中需要对光谱进行基线校正。分析了传统光谱基线校正方法的局限性,并设计了基于RBF神经网络和NURBS曲线模型的ICP-AES光谱基线校正方法。首先利用高斯滤波对原始光谱进行去噪预处理,然后基于RBF神经网络筛选光谱基线点序列,并以此构造NURBS曲线内节点序列;利用NURBS曲线逆向计算模型计算NURBS曲线控制点序列;通过控制点序列和内节点序列拟合NURBS曲线作为相应的光谱基线,达到消除光谱基线漂移的目的;最后利用实测数据对提出的ICP-AES光谱基线校正方法进行验证。实验结果表明,提出的ICP-AES光谱基线校正方法能够有效的扣除光谱基线,为后续的元素含量定量分析提供了技术基础。

In the ICP-AES measurement system,the drift of the spectral baseline will cause errors in the quantitative analysis of element concentration.Therefore,the spectral baseline needs to be corrected during the measurement process.In this paper,the limitations of traditional spectral baseline correction methods were analyzed,and on this basis,an ICP-AES spectral baseline correction method based on RBF neural network and NURBS curve model was designed.First,we used Gaussian filtering to denoise the original spectrum,and then filtered the spectral baseline point sequence based on the RBF neural network,and constructed the internal node sequence of the NURBS curve;Then,we used the NURBS curve inverse calculation model to calculate the NURBS curve control point sequence.NURBS curve was fitted by control point sequence and inner node sequence as the corresponding spectral baseline to eliminate the drift of spectral baseline.Finally,the measured data were used to verify the ICP-AES spectral baseline correction method proposed in this paper.The experimental results show that the ICP-AES spectral baseline correction method proposed in this paper can effectively subtract the spectral baseline,which provides a technical basis for the subsequent quantitative analysis of element content.