Application of Singular Value Decomposition (SVD) in Seismic Imaging: Theory,Methods, and a Synthetic Case Study

Abstract

Singular Value Decomposition (SVD) is a powerful linear algebra technique widely applied in geophysical imaging, especially seismic data processing. Its utility in solving ill-posed inverse problems, denoising, and data compression makes it indispensable for interpreting large and noisy datasets. This study explores the theoretical foundations of SVD and its application in seismic imaging, with emphasis on inversion and noise attenuation. A synthetic seismic dataset is used to demonstrate how truncated SVD improves the quality of subsurface models. The results highlight the effectiveness of SVD in stabilizing inversion processes and enhancing resolution in seismic sections. We also discuss the implications of singular value truncation on data fidelity and inversion accuracy, setting the stage for future work combining SVD with machine learning and real-time imaging systems.

Dhaka Univ. J. Sci. 74(2): 201–206, 2026 (July)