A Multiscale Dip Estimation Method Based on Optimized Finite Difference Coefficients

Local dip estimation is essential for various applications in exploration geophysics, directly impacting seismic interpretation and imaging quality. Conventional methods, including the direct method, plane wave destruction filter (PWD) method, structure tensor (ST) method, and one-lag correlation (OLC) method, each have limitations. The direct method is noise-sensitive and inaccurate, while PWD suffers from high computational costs and noise-induced oscillations. ST offers noise suppression but is highly dependent on parameter selection and lacks adaptability, and OLC improves noise resistance at the cost of resolution. To mitigate these issues, multiscale schemes have been applied to dip estimation, achieving a certain balance between noise suppression and resolution. To further enhance performance, we propose a multiscale dip estimation framework based on optimized finite difference (OFD) coefficients. This method applies multiscale filtering to seismic data, followed by high-order finite difference dip estimation at each scale. We introduce three fusion weights—local consistency weights, cross-consistency weights, and local signal-to-noise ratio (SNR) weights—to balance noise suppression and resolution during dip fusion. Tests on synthetic and real datasets demonstrate that the proposed multiscale OFD method achieves superior noise suppression while preserving fine structural details, providing an effective solution for high-quality local dip estimation.