An Efficient Undersampled High-Resolution Radon Transform for Exploration Seismic Data Processing

摘要

Radon transforms have been widely utilized for exploration seismic data processing. They have been part of the seismic data processing workflow for the last few decades. They are robust, easy to compute, and mathematically well established. This paper suggests a new method for obtaining an undersampled high-resolution Radon transform. The proposed method is based on a nonlinear sampling technique known as compressive sensing, which assumes that seismic data is sparse in a certain domain. The Radon transform domain can be sparse for exploration seismic data. The proposed method was applied for different seismic data processing applications including: 1) attenuation of multiple reflections; 2) first-arrival picking; and 3) seismic denoising. The method was tested on synthetic as well as real seismic data. Additionally, it was compared with existing methods for low- and high-resolution Radon transforms. From the simulation results, it is clear that the proposed method not only reduces the number of measurements needed but also produces high-resolution Radon transforms with less computational time. Therefore, it is believed that the proposed method is an appropriate alternative to some of the existing methods for efficient high-resolution sparse Radon transform computation.