3D seismic waveform classification study based on high-level semantic feature

Abstract

With the improvement of Natural energy exploration technologies, the Seismic interpretation member need to deal with more and more information and parameters. How to better use seismic characteristic parameter to detect hydrocarbon becomes increasingly complex. In this article, we deeply studied the seismic waveform classification, and propose a seismic waveform classification method based combine various characters. After reducing the dimensions of seismic wave, we classify it using the high-level semantic feature extraction technique in pattern recognition. Experiments proved that, the classification result improved in continuity and details, and reduced the redundancy of seismic signal, increased performance of classification.