小波变换和神经网络的X射线衍射波谱滤噪技术比较

摘要

The spectrum of X-ray polycrystalline diffraction contains background and noise, while noise is the main factors affecting the resolution of diffraction peak. In this paper, the X-ray polycrystalline diffraction spectrum of an iron ore is noise-filtered by means of neural network, simulink and wavelet analysis through introducing it into MATLAB. The effects of the three means were compared, and it was concluded that the effect of wavelet analysis is the best.%多晶X射线衍射扫描所形成的波谱会附带一些背景和噪声,而噪声是影响衍射谱峰分辨率的主要因素.本文通过多晶X射线衍射仪扫描某铁矿石矿物组成所形成的图谱,导入到MATLAB,分别采用神经网络、系统仿真和小波分析三种手段,对该图谱进行滤噪.最终对这三种滤噪方式的滤噪效果进行比较,得出小波分析的方法滤噪效果最佳.