基于稀疏变换的大型固定干扰信噪分离技术研究

Signal-to-noise separation of a large-scale fixed interference source based on sparse transformation

在复杂地表区域的地震资料采集中,大型厂矿的机械干扰、大钻、高速公路等固定的干扰源会产生能量极强的固定点震源干扰,且在干扰源附近产生线性或双曲型的相干噪声,通过常规信噪分离方法难以对其进行有效压制,从而对资料的信噪比及后期的成像产生严重影响。野外施工时往往只能停工协调或变观以尽量减少此类噪声的影响,会造成施工成本增加或影响成像质量。利用基于稀疏变换和同步压缩变换的信噪分离技术,对此类噪声从能量上进行有效的压制。针对残留的线性和双曲型相干噪声,在FX域采用非均匀空间采样相干噪声压制法(NUCNS)进行噪声压制,通过上述技术的综合应用,最终实现了强能量固定干扰源的噪声压制处理。通过实际资料处理试验表明,此方法能有效压制强能量的大型干扰源噪声,并能较好地还原淹没在噪声中的有效反射信息。

In the seismic data acquisition of complex surface areas,fixed interference sources,such as the mechanical interference of large factories,large drills,or highways,produce strong fixed-point source interference.Linear or hyperbolic coherent noise is generated near the interference source.It is difficult to effectively suppress it using the conventional signal-noise separation method,which has a bad effect on the signal-to-noise ratio of data and imaging.Usually,work must stop or geometry must be changed to minimize the impact of such noise,which increases the cost or affects the imaging quality.The sparse transform and synchronous compression transform signal-noise separation technology is used to effectively suppress such noise from energy.For the residual linear and hyperbolic coherent noise,the frequency space domain NUCNS is used to analyze and model the noise,and finally,successfully suppress the noise of strong-energy fixed interference sources.The actual data processing test showed that this method could effectively suppress large interference sources with strong energy and restore the effective reflection information submerged in the noise.