Research On Key Technologies Of Visualization And Interaction Based On Seismic Data

The visualization of seismic data and the seismic event pickup are the two basic means of interpretation of seismic exploration data.Excellent visualization results and accurate seismic event pickup results can effectively improve the reliability of seismic interpretation results and guide specific field exploration operations.In recent years,due to the incomplete details and interactive limitations of traditional seismic data two-dimensional visualization,it has been gradually replaced by three-dimensional visualization of seismic data.To this end,this paper proposes a new seismic data visualization method for rendering 2D seismic data into 3D scenes by means of VTK.The method uses three orthogonal planes and enclosing surfaces to reveal seismic data stored in SEGY format.VTK visualization in 3D virtual stereo scenes and support 360-degree interactive operation.In the specific implementation,the filling hole phenomenon in the variable area graph is combined with the interpolation algorithm and the small trapezoid polygon filling algorithm.For the mosaic phenomenon that may appear in the variable density map,the bilinear interpolation algorithm is used to solve the problem.Comprehensive optimization of specific seismic data visualization methods to enhance its fidelity,comfort and interaction friendliness.At the same time,in the specific user interaction,the picking of the seismic event is gradually becoming more automatic and intelligent,but the application of the existing seismic event pickup method is not satisfactory and has a lot of limitations.Aiming at this problem,this paper proposes a Richer Convolutional Features of edge detection(RCF)algorithm to realize the automatic picking of seismic event based on the principle of deep learning.By combining multi-scale and multi-level feature information acquired from the deep learning network,the algorithm can effectively identify and locate the edge of objects in complex scenes.According to the specific application of the actual seismic event pickup problem,this paper improves the existing convolutional network structure and realizes the automatic seismic event pickup through the combination of edge connection and central axis extraction algorithm.Through specific application examples,the specific application effects of the seismic event pickup based on the traditional method and the deep learning method are compared from qualitative and quantitative perspectives.The comparison results show that the seismic event picked up by the improved RCF has the advantages of more accurate,more continuous,smoother at the same application in complex scenes.It proves the feasibility of using the improved RCF edge detection method for seismic event pickup.