Multi-scale Hessian Matrix-Based Seismic Attribute Fracture Detection Technology

To achieve high-precision detection of small-scale fractures and quantitative characterization of fracture parameters, this study proposes a seismic attribute-based fracture detection technology utilizing multi-scale Hessian matrices. Building upon post-stack seismic fracture detection attributes, the algorithm implementation comprises three key steps: (1) Linear enhancement of fracture detection attributes through Hessian matrices and multi-scale Jerman filtering; (2) Fracture centerline detection using the Steger algorithm to refine enhanced fracture attributes, followed by Hough transform-based fracture line vectorization; (3) Discrete fracture statistical analysis via bounding box algorithms, enabling quantitative computation of fracture attributes including length, density, and orientation. Application to field data demonstrates that predicted fracture density and orientation align consistently with fracture density interpretations from imaging logs, validating the method’s accuracy. When integrated with conventional seismic multi-scale fracture prediction techniques, this technology effectively enhances fracture characterization precision and enables quantitative characterization of fracture parameters.