The Seismic Data Interpretation Technology And Application Based On Pattern Recognition

With the increase of mining depth and the development of reconstruction technology, coal mining enterprises need higher standards of 3D seismic exploration technology. In recent years, digital high density 3D seismic exploration technology develop rapidly and can collect more singles of reflected waves from underground impedance interfaces, which makes it possible to find tiny geological structures. Then how to find tiny geological structures comes to be a problem.3D seismic exploration technology can collect singles of reflected waves from underground impedance interface. Analyzing the characteristics of reflected waves contributes to seismic interpretation about geological structure and lithology, which can provide technical support for mining coal.According to the characteristics of pattern recognition algorithms and the geological tasks, this paper establishes the pattern recognition process for identifying subtle geological structure, based on the application of research of pattern recognition and summarizing the advantages and disadvantages of various pattern recognition algorithms. We turn the real seismic data singles to frequency domain based on CWT and STFT, study the response of tiny geological structure.We attach watershed algorithm to the edge detection of the abnormal seismic attribution. We designed the typical geological model of small faults, collapse columns and magmatic intrusive mass to simulate the seismic response to identify the feasibility and necessity of pattern recognition based seismic attributions analytic technique. Analyzed the simulated results and draw the conclusions as follows:higher dominant frequency of seismic data made the response of tiny geological structure clearer; crushed zone blur the response of tiny geological structure; collapse columns and coking coal seam weaken the energy of reflected waves; magmatic intrusive mass strengthen the energy of reflected wave.Applied pattern recognition process to interpret the real seismic data from NS and JCZ coal mining enterprises.Used seismic attributions extracted in different ways to find tiny geological structure based on pattern recognition algorithms and CWT. Spectral decomposition analysis result draw conclusion as follows:the component near the dominant frequency made the response of tiny geological structure clearer than others. Pattern recognition algorithms process can enhance the resolution of seismic interpretation, combined with CWT.