DAS VSP Flow Noise Attenuation Based on Matching Pursuit

Production-related noise on DAS VSP data generally not only obscures the underneath seismic signal but also results in a distorted VSP image. This paper presents an effective flow noise attenuation workflow for DAS VSP data based on Matching Pursuit. It decomposes DAS VSP traces into Gabor atoms followed by a reconstruction of the flow noise model using a linear combination of selected components. Flow noise attenuation is achieved by subtracting the reconstructed flow noise from the input data. The authors demonstrate performance on field DAS VSP data recorded in a producing well. Examples on field data with strong flow noise illustrate the effectiveness of the method.