Seismic multi-attribute depth fusion-based wave impedance inversion low-frequency model construction method

• Main Authors: ZHANG GUODONG, MA GUANGKE, LI FANG, LIAO YI, LI YONG, LI LEI
• Format: Patent
• Language: Chinese; English
• Published: 08.08.2023
• Subjects: DETECTING MASSES OR OBJECTS; GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; MEASURING; PHYSICS; TESTING

The invention discloses a wave impedance inversion low-frequency model construction method based on seismic multi-attribute depth fusion, and the method comprises the steps: building a well interpolation low-frequency model, carrying out the post-stack inversion, obtaining a band-pass wave impedance inversion body, and extracting other attribute bodies: a three-transient attribute body, a time window frequency attribute body, and a frequency division seismic body; and then, extracting a pseudo well curve, obtaining a wave impedance body through deep learning, generating a residual body, and finally obtaining a low-frequency model with relatively high precision. According to the method, multiple attribute bodies with geological significance are fused through deep learning, the uncertainty caused by well interpolation can be effectively reduced, the precision of a low-frequency model is improved, and a good basis is provided for high-precision seismic wave impedance inversion. 本发明公开了一种基于地震多属性深度融合的波阻抗反演低频模型构建方法，