Full text loading...
-
Recovery of Blended Data - A Sparse Coding Approach for Seismic Acquisition
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 78th EAGE Conference and Exhibition 2016, May 2016, Volume 2016, p.1 - 5
Abstract
Simultaneous shooting is one of the main levers to increase the cost-effectiveness of seismic data acquisitions - either by decreasing the acquisition time or increasing the shot density. It is then fundamental to design efficient source separation solutions to recover blended data. We address the de-blending problematic in the context of land vibroseismic as its operational models are naturally suited for blended acquisition. In the case of multiple autonomous vibrator trucks, the essential criteria of shooting time randomness is met and at the same time, many constraints encountered with classical acquisition disappear (such as shooting time patterns). Fully unconstrained source acquisitions open the way to unprecedented production rates and shot densities. Simultaneous shooting acquisition can be considered as a case of Compressed Sensing (timely compressed data). Applying concepts and techniques from this field, we design a deblending procedure based on inverse problems in the Curvelet domain. We use a mathematical formulation to address simultaneous source acquisition. The data recovery is based on the search of the sparse code of the "clean" data in the Curvelet domain, through a 11 regularized inverse problem. Our procedure has been successfully used to deblend 3D common receiver gathers from a real blended acquisition.