论文部分内容阅读
基于分段光谱特征值提取法和小波变换算法等多个数据预处理方法,分别针对分段基线差异及光谱噪声等严重影响激光诱导击穿光谱(LIBS)信号质量的主要影响因素,开展光谱信号预处理研究。基于实验室LIBS实验装置,通过实验验证,基于多通道光谱仪不同波段光谱特征值提取,提出了一种简单易行的多组数据中特征值点连接的方法,有效地提高了LIBS光谱信号的基线平直度,并得出以小波变换算法进行LIBS谱线信号去噪的最佳算法参数。在上述工作的基础上,使用基于误差反向传播的人工神经网络方法,实现了纯铜和不锈钢等物质种类的有效识别,研究结果表明,综合利用多数据处理方法进行LIBS技术中光谱信号处理可以有效提高谱线分析和识别的质量。
Based on multiple data preprocessing methods, such as eigenvalue extraction and wavelet transform, the main influencing factors of laser induced breakdown spectroscopy (LIBS) signal, such as segmentation baseline difference and spectral noise, Pretreatment studies. Based on the laboratory LIBS experimental apparatus, a method of eigenvalue connection in multiple groups of data based on multichannel spectroscopy is proposed based on the experimental verification. Based on multi-channel spectroscopy, the baseline of LIBS spectral signal is effectively improved Flatness, and get the best algorithm parameters of LIBS spectral line signal denoising by wavelet transform algorithm. Based on the above work, the artificial neural network based on error backpropagation is used to realize the effective identification of material types such as pure copper and stainless steel. The results show that the spectral signal processing in LIBS can be integrated by using multiple data processing methods Effectively improve the quality of line analysis and identification.