基于小波包分析光声光谱信号的滤波方法研究

Research on the Filtering Method Based on Wavelet Packet Analysis for Signals of Photoacoustic Spectroscopy

针对变压器油中溶解气体在光声光谱检测信号中,存在噪声干扰、随机误差等因素而不能精确测量光声光谱中弱光谱信号的问题,提出了一种基于小波包分析的滤波方法。研究了小波包分析和滤波的原理、不同小波函数和阈值函数的选取,比较了选择不同小波函数、阈值函数对光声光谱信号滤波的效果;选取合适的小波包基、阈值量化,实现了弱光谱信号的小波包滤波处理。研究结果表明:利用小波包滤波能有效减小随机误差,在去除噪声的同时不改变原信号的相位,也不会产生波形的畸变;在提高光声光谱测量精度的同时,有效提高光声光谱在线监测系统的分析精度和模型稳定性。

The filtering method based on wavelet packet analysis is proposed for solving the problems of noise interference, random errors, and inaccurately measure the weak spectral signals of photoacoustic spectroscopy. The principles of wavelet packet analysis and filtering, the selection of wavelet function and threshold function are studied, and the filtering effects are compared among different choices of wavelet function and threshold function. The appropriate wavelet packet and threshold quantization are selected to implement the wavelet packet filtering processing for weak spectral signals. The result of research shows that the wavelet packet filtering effectively decreases the random error,the original phase of the signal is not changed during de - noising, and the waveform is not distorted ; the measurement accuracy of photoacoustic spectroscopy is improved,thus the analysis accuracy and stability model of online monitoring system are enhanced effectively.