Abstract:

Aiming at the issues that the conventional TFPF(Time-Frequency Peak Filtering) doesn't consider the correlation between seismic channels and the radial-trace TFPF doesn't consider the real direction of the reflected event and the reflected event is locally linear, we propose self-adaptive radial-trace Time-frequency peak filtering. The new method finds the direction of the reflected event in the local area by utilizing the self-adaptive function and maps the event to radial-trace area along the direction by RTT(Radial-Trace Transform), and can improve the linearity of the signal. Consequently, it is more close to the unbiased TFPF and can reduce the error caused by the low linearity. Experiment results indicate that this method can provide better performance in both
random-noise attenuation and reflected signal preservation. When selecting same window length, the method can increase the SNR(Signal-to Noise Ratio) by 5 dB more than traditional TFPF and 3 dB more than radial-trace TFPF.

Key words: time-frequency peak filtering (TFPF), radial-trace transform (RTT), self-adaptive function;window length (WL), random noise