摘要
针对探地雷达信号处理中直达波衰减问题,提出了一种基于Curvelet变换的直达波衰减方法。该方法首先将原始信号变换到Curvelet域,通过直达波在Cur-velet域的分布特点,得到其Curvelet系数,并对其进行重构,最后在总信号中减去重构的直达波信号得到有效信号。在处理过程中,利用Curvelet域分尺度阈值处理方法同时可以滤除随机噪声。通过对实测数据的处理和分析,验证了该方法的可行性和有效性,并且证实了该方法与2-D连续小波变换方法相比具有占有存储少和计算速度快的特点。
The direct wave reduction is a crucial problem in GPR（Ground Penetrating Radar） signal processing. In this paper, a direct reduction method using the curvelet transform is proposed. First of all, original signal is transformed into curvelet domain, then, extract curvelet coefficient of direct wave by its distribution feature is obtained and reproduced. Finally, the effective signal is gained by subtracting the direct wave from total signal. During the processing, a threshold method depending on scale of curvelet domain is used to reduce random noise. The real measured data show that the method is feasible and effective in reducing direct wave, and is of less storage requirement and is faster than 2-D continuous wavelet transform.
出处
《电波科学学报》
EI
CSCD
北大核心
2008年第3期449-454,共6页
Chinese Journal of Radio Science
基金
自然科学重点基金项目(69831020)
自然科学基金面上项目(40674064)
