Compressive seismic reconstruction with extended POCS for arbitrary irregular acquisition

Jiang, Tao (In-depth Geophysical, Inc.) | Gong, Bin (In-depth Geophysical, Inc.) | Qiao, Feng (In-depth Geophysical, Inc.) | Jiang, Yifeng (In-depth Geophysical, Inc.) | Chen, Anbo (In-depth Geophysical, Inc.) | Hren, David (In-depth Geophysical, Inc.) | Meng, Zhaobo (In-depth Geophysical, Inc.)


Large-scale seismic acquisition based on compressive sensing (CS) has been made practical recently, thanks to the advance in CS algorithms and increasing computation power (Mosher, 2014; Li, 2013). Compressive seismic acquisition (CSA) can substantially reduce time and costs with proper compressive seismic de-blending (CSD) and especially compressive seismic reconstruction (CSR). Projection onto convex sets (POCS) is a very popular choice among various mathematical approaches for CSR, because of its efficiency, simple parameterization, and flexibility to expand to high dimensions. However, the efficiency of POCS by utilizing FFT leads to its main drawback of relying on regular grids and 4D binning before 5D interpolation. In this paper, we propose an extended POCS (EPOCS) algorithm to incorporate an interpolating operator into the sampling operator and extend its usage for under-sampled arbitrary irregular acquisition. This is important because CSR relies on grids with certain designed randomness. Both synthetic and real data tests demonstrate the effectiveness of the CSR with EPOCS.

Presentation Date: Tuesday, September 26, 2017

Start Time: 4:20 PM

Location: 360A

Presentation Type: ORAL