Marine Towed Streamer Data-Reconstruction Based on Compressive Sensing

摘要

In 2011, ConocoPhillips conducted a 3D towed streamer survey in Barents Sea, with optimized shot locations and cable configurations (NUOS design). The purpose of this unconventional survey was to test the ability of compressive sensing to increase the effective spatial bandwidth of the seismic data. In this paper, we describe an alternating direction method (AD-M) combined with a nonmonotone line search technique for seismic data reconstruction. It solves a general analysis-based optimization model derived from compressive sensing. This method is highly robust due to the nature of ADM, and able to quickly approach the global minimum for large-scale problems due to the nonmonotone line search. We applied this method to two acquired data sets in the same area—one with a NUOS design and one with a conventionally towed streamer. The, final imaging results show the significant improvement of resolution for both data sets obtained from applying the technology, and inspire future marine survey design and processing.