基于时间窗边线检测和多级中值滤波的混采数据分离

Separation of blended data based on time window edge line detection and multilevel median filter

针对多源地震采集技术中的混采数据分离,提出了一种时间窗边线检测和多级中值滤波方法.该方法利用了伪分离记录在非共炮点域,混叠噪声呈随机分布的特性.首先设计一个合适的时间窗口并且设定时间窗边线阈值,然后对转换到非共炮点域的伪分离记录进行检测,对不满足阈值条件的窗内记录进行滤除,接着对伪分离记录进行多级中值滤波,进一步压制残余噪声,并且利用时间窗边线检测和多级中值滤波对过滤信号进行有效信号提取,最后把伪分离记录转换回共炮点域从而实现混采数据分离.与常规的混采数据分离技术相比较,采用时间窗边线检测和多级中值滤波方法拥有更高的信噪比.

Multi-source blended seismic acquisition is an efficient seismic acquisition technology, and the deblending is the key to multi-source blended seismic acquisition technology. In order to achieve the deblending in multi-source blended seismic acquisition technology, a new method called time window edge line detection and multilevel median filter is proposed. Generally, it can get the pseudo-deblended record from the blended data through pseudo-deblended. And the blending noise of pseudo-deblended record is continuously distributed in common source gathers when it is randomly distributed in common receiver gathers, common mid-point gathers and common offset gathers. This method uses the distribution feature of the blending noise in common receiver gathers, common mid-point gathers and common offset gathers. Firstly, an appropriate time window is designed and the time window edge line threshold is set. Secondly, it detects the pseudo-deblended record that is switched to common receiver gathers, common mid-point gathers and common offset gathers. When the record in window does not conform to the threshold condition is filtered out. Gradually shrink the time window until the record in window is no longer filtered out effective. Next, the multilevel median filter is applied to the pseudo-deblended record, further suppressing the residual noise. Then it uses the time window edge line detection and multilevel median filter on the residual record for effective signal extraction. Add the extraction of effective signal to the pseudo-deblended record. It can compensate for the missing details of pseudo-deblended record. Finally, the pseudo-deblended record is converted to the common source gathers to realize the deblending. The experiment results of synthetic data and field data show that the method of time window edge line detection and multilevel median filter can complete the deblending well. And compared to the conventional deblending technology, the method of time window edge line detection and multilevel median filter has a higher signal-to-noise ratio.