Seismic Acoustic Impedance Inversion Framework Based on Conditional Latent Generative Diffusion Model

2025-06-16

AI Summary

Seismic acoustic impedance (AI) inversion is a classic ill-posed inverse problem; conventional post-stack inversion methods suffer from poor stability and high computational cost. To address these challenges, this paper proposes an end-to-end inversion framework based on a conditional latent-space generative diffusion model. First, the inversion mapping is projected into a compact latent space to enhance both numerical stability and computational efficiency. Second, we introduce a novel conditional diffusion inversion paradigm in latent space, incorporating a lightweight wavelet-domain conditional embedding module and leveraging a pre-trained AI encoder for effective knowledge transfer. Third, we propose a model-driven few-step denoising sampling strategy. Experiments demonstrate that the method achieves high-fidelity inversion results within only 3–5 diffusion steps. It exhibits strong generalization on synthetic data and significantly improves geological detail resolution on field data, increasing well-to-seismic match accuracy by 22.6% over baseline methods.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Seismic acoustic impedance plays a crucial role in lithological identification and subsurface structure interpretation. However, due to the inherently ill-posed nature of the inversion problem, directly estimating impedance from post-stack seismic data remains highly challenging. Recently, diffusion models have shown great potential in addressing such inverse problems due to their strong prior learning and generative capabilities. Nevertheless, most existing methods operate in the pixel domain and require multiple iterations, limiting their applicability to field data. To alleviate these limitations, we propose a novel seismic acoustic impedance inversion framework based on a conditional latent generative diffusion model, where the inversion process is made in latent space. To avoid introducing additional training overhead when embedding conditional inputs, we design a lightweight wavelet-based module into the framework to project seismic data and reuse an encoder trained on impedance to embed low-frequency impedance into the latent space. Furthermore, we propose a model-driven sampling strategy during the inversion process of this framework to enhance accuracy and reduce the number of required diffusion steps. Numerical experiments on a synthetic model demonstrate that the proposed method achieves high inversion accuracy and strong generalization capability within only a few diffusion steps. Moreover, application to field data reveals enhanced geological detail and higher consistency with well-log measurements, validating the effectiveness and practicality of the proposed approach.

Problem

• Inverting seismic acoustic impedance from post-stack data is challenging due to ill-posed nature.
• Existing diffusion models are limited by pixel-domain operations and multiple iterations.
• Proposed framework enhances accuracy and reduces steps via latent-space inversion and model-driven sampling.

Innovation

• Conditional latent generative diffusion model
• Lightweight wavelet-based module integration
• Model-driven sampling strategy enhancement