基于迭代阈值收缩的高分辨率Radon变换方法效果对比

提高Radon变换分辨率的研究一直是地震数据处理方法研究的热点之一。常用的提高分辨率的算法是在频率域进行的,然而此方法在频域中计算的模型权重是偶然耦合的,会对所有同相轴施加相同的权重,导致高能量同相轴地震道生成假像。此文给出了3种时间域提高Radon变换分辨率的方法,IST(iterative shrinkage thresholding)、FIST(fast iterative shrinkage thresholding)和SRTIS(sparse radon transform iterative shrinkage),并对3种算法的收敛效率和计算效果进行了对比分析。理论和实际数据的测试表明,SRTIS方法在效率和效果的对比上均优于其他两种,并且具有较好的去除多次波能力。

The comparison of effects of high-resolution Radon transform based on iterative shrinkage thresholding

The study of the resolution improvement of Radon transform is one of the research hotspots in seismic data processing area.The commonly-used resolution improvement methods are carried out in the frequency domain.However,the weighting of the Radon model in the frequency domain is coupled,and it will impose the same weight on all the events,leading to the artifacts generated by high energy seismic events.This paper presents three resolution improvement methods of Radon transform in the time domain:Iterative Shrinkage Thresholding(IST),Fast Iterative Shrinkage Thresholding(FIST)and Sparse Radon Transform Iterative Shrinkage(SRTIS),with a comparison of their compute efficiencies and results.Synthetic data and real data test results show that SRTIS is superior to the other two methods in computation effect and efficiency,and it has a better multiple attenuation capability.