时频域自动识别与压制高频噪声

Automatic recognition and suppression of high-frequency noises in time-frequency domain

本文在分析张宇等及钱忠平等提出的高频噪声判别准则基础上,利用夏洪瑞等根据在窄档等频距小波分频之后识别、压制高频噪声的思路,提出在时频域将代表地震道主要能量的几个频带的平均振幅绝对值作为正常子波的平均振幅绝对值,以此来识别和压制高频噪声的异常振幅。实际处理结果表明,此法不仅可以压制高频谐振及高频突发噪声,而且可以保护弱地震反射信号不受损害。

On the basis of analyzing the norm of recognizing high-frequency noises presented by Mr. Zhag Yu and Mr. Qian Zhong-ping and using the idea that recognition and suppression of high-frequency noises after frequency division of wavelet in equal frequency components within a narrow band presented by Mr. Xia Hong-rui et al. ,the paper presented that in time-frequency domain taking absolute values mean amplitude in several bands representing main energy in seismic traces as absolute values of mean amplitude of normal wavelet to recognize and suppress the abnormal amplitudes of high-frequency noises. The results of practical processing showed that the method not only suppresses high-frequency harmonics and high-frequency pulse noise,but also can protect the weak seismic reflection signals from damage.