Three-dimensional Earthquake Fault Identification Method Based On Ant Colony Algorithm

Fault is the key factor of controlling the formation and distribution of oil-gas field. Detection and recognition of fault is one of the most important foundation works in oil and gas exploration. Traditional methods of fault interpretation usually work at vertical seismic section which is assembled on horizontal plane later. Affected by subjective factors, the reliability and accuracy are low as well as working efficiency. Using edge detection techniques, 3D coherent method can effectively recognize fault in 3D seismic data. However, using 3D coherent method simply can only display the position of faults in 3D coherent slices, the 3D configuration can not be obtained, which is not contribute to 3D stratigraphic analysis.Our paper probed into the approach of fault recognition of 3D seismic coherent data. In this approach, fault lines are extracted from a series of coherent slices automatically at first, spliced into 3D fault plane after, finally interactive displayed through vision techniques. Our main works focus on:1. Use ant colony algorithm to obtain thresholding fault image from classifying coherent slice image.2. Use opening and closing operation in Mathematical Morphology to remove the noises in thresholding fault slice data. Then use Zhang thinning algorithm to thinning the slice image. At last, we wipe off furcations in refine images and connect segregate lines.3. Apply Gana's heuristic restructuring algorithm to reconstruct 3D fault plane, and then display it interactively by OpenGL.Our experiments of an actual three dimensional seismic coherent data show that this method can recognize and display the shape of 3D fault plane, raise the recognition efficiency effectively.