Adaptive sub-images BSR pattern recognition system for seismic reflection images

Abstract Commonly, the Bottom-Simulating Reflector (BSR) location is done by hand-operation by a trained interpreter. Recently, a pattern recognition system named the 2D multiresolution automated system for detecting BSRs on post-stack seismic reflection images has been presented. That system focuses on locating BSRs parallel and non-parallel to the Sea-Bottom Reflector (SBR). However, in that system, the seismic reflection sub-images (containing a BSR) are of the same size. Also, the time contribution of the SBR is eliminated by semi hand-operation. The present work proposes a methodology that uses adaptive sub-images to get a more efficient BSR pattern recognition system. The lateral continuity, intensity, and the inclination angle of the BSR determine the number of traces in each adaptive sub-image. Also, this technique removes the SBR influence automatically. The system was tested using seismic reflection images from offshore the southern Baja California Peninsula. The adaptive sub-image BSR pattern recognition system locates the BSRs accurately in all the cases, even when the BSR cuts lithologic reflectors. Low time was employed to obtain the BSR position, which allows us to implement this methodology in desktop or laptop machines.