Noise attenuation technology on wide azimuth seismic data

In wide azimuth seismic data geometry, the ratio of crossline offset to inline offset is relatively large, and the spatial sampling rate of seismic data is irregular. The main features are: (1) The distance between adjacent traces is less than the designed receiver interval. (2) The greater the cross-line offset, the smaller the distance between adjacent traces, the spatial sampling rate varies rapidly with the crossline offsets. The time-distance curve of linear noise with large crossline offset, such as surface waves and refraction waves change into hyperbolic. Conventional linear noise suppression techniques, such as FK transform, are suitable only for a relative regular spatial sampling rate, but cannot completely remove the linear noises on the wide azimuth seismic data. In this article, we discuss two methods: 2-D Fourier transform filter based on linear moveout correction and 3-D conical filter based on orthogonal subset. The two methods make the spatial samples regular by different ways, which meet with the algorithm requirements of different domains. It is proved that the two methods have good results in linear noise attenuation on 3-D seismic data, especially on wide azimuth seismic data.