New Method Enhances Seismic Data Denoising

Recent advancements in seismic data processing have been made with the introduction of a method termed "Adaptive Graded Denoising of Seismic Data Based on Noise Estimation and Local Similarity" by Xueting Yang, Yong Li, Zhangquan Liao, Yingtian Liu, and Junheng Peng. This method addresses common issues in seismic data denoising, particularly when dealing with noisy, complex, and uneven data. Traditional techniques often result in either over-denoising or under-denoising, which can compromise the integrity of seismic data analysis.

The proposed approach begins with an assessment of the average noise level across the seismic data. It employs block matching and three-dimensional filtering (BM3D) techniques to denoise the data initially. Following this, the denoised data is compared to the residual data using local similarity metrics to identify areas where noise levels significantly deviate from the average. These identified regions are then retained intact, allowing for a more nuanced approach to denoising.

The researchers conducted experiments using both theoretical models and actual seismic data, demonstrating that their method effectively improves the quality of seismic data with uneven noise characteristics. The findings suggest that this adaptive graded denoising technique could enhance the reliability of seismic data processing, which is crucial for various applications, including resource exploration and geological studies.

The full details of this research can be found in the paper available on arXiv: Adaptive Graded Denoising of Seismic Data Based on Noise Estimation and Local Similarity.