Adaptive DAS coupling noise suppression based on local MCA

The distributed acoustic sensing (DAS) system deploys standard optical fiber cable along the well, which can produce high-precision VSP images, and can implement functions in typical VSP applications, such as checkshots, imaging and time-lapse reservoir monitoring. However, the acquired VSP dataset may suffer from the strong coherent noise, which is mainly caused by the physical placement and the swing of the wireline in the well. This noise may reverberate in the uncoupled area of the cable and the well. Since morphological component analysis (MCA) can separate different components of seismic data input that show distinguishable morphologies with over-complete sparse representation dictionaries, great improvements were shown in suppressing DAS coupling noise. However, the MCA method may ignore the time-varying characteristics of the effective signal and the DAS coupling noise when suppressing the DAS coupling noise, so the MCA algorithm can't completely suppress DAS coupling noise in some areas of the seismic data. In the paper, we propose the local MCA algorithm that using MCA algorithm working on segments of one trace to separate DAS coupling noise instead of working on the whole trace. we automatically select the optimal sampling point length according to the sparsity of the required signal and DAS coupling noise, and then implement an adaptive DAS coupling noise removal method based on local MCA.