探地雷达回波信号处理中小波变换域滤波方法的研究

Study of wavelet domain filtering techniques in the processing of the received signals of a ground penetrating radar

探地雷达信号是非平稳信号．在传统的傅氏变换域滤波不能有效地解决滤波和信号高频损失的情况下，文中研究了利用小波变换来处理这一矛盾的方法．系统地讨论了小波变换在信号滤波中的原理和各种实现方法，比较了与傅里叶变换域滤波的差别．最后对实际的探地雷达信号进行了处理．结果表明，基于小波变换域的滤波能滤除８０％～９０％的噪声，同时可基本保持原信号的边缘不变．此外小波变换域滤波还具有直观、简便、快速操作、稳健的优点，完全可用于实时或事后的数据处理．

This paper tries to exploit wavelet domain filtering techniques for handling the problem of the filtering of the received signals from a ground penetrating radar (GPR), which are non stationary, while traditional Fourier transform based filtering can not make an acceptable compromise in getting rid of noises and, at the same time, keeping the sharp edges of the signals. The principles and the methods of the wavelet domain filtering are systematically discussed. A comparison between the wavelet domain filtering and the Fourier transform based filtering is made. Some signals from field measurements are filtered by using the new methods. Results show that wavelet domain based filtering can remove noise up to the extent of 80%～90% and can keep the sharp edges of the signals unchanged. In addition to this, the new scheme has the advantages of intuition, easy operation, fast implementation, robustness and easy utilization in real or post GPR data processing.