地震岩性模式识别研究

模式识别技术引入到地震岩性预测领域,实现了地质、测井和地震综合资料的自动解释。地震岩性模式识别同时采用了统计模式识别和结构模式识别两种技术,并将两者合理地结合为一体,通过地层岩性-地层速度-地震反射波形的模式匹配、地震岩性特征抽取、以及句法分析和判别运算,完成地层岩性的分类和识别。在辽西凹陷北洼应用地震岩性模式识别进行了模式训练和钻前岩性预测,所获结果均表明该法具有较高使用价值。

RESEARCH ON SEISMIC LITHOLOGIC PATTERN RECOGNITION

This paper introduces the technique of pattern recognition into the seismic litho-logic prediction, and realizes autometic interpretation of integrated data consisting of geology, well logging and seismic profiles. Seismic lithologic pattern recognition (SLPR) adopts statistic pattern recognition and structural pattern recognition simultaneously, combining these two methods reasonably. Through pattern matching of lithology-velocity-seismic waveshape, characteristics extracting of seismogram and lithology, sentence analysis and discriminating operation, SLPR achieves classification and recognition of subsurface lithology. SLPR was applied to conduct pattern training and practical lithology predicting in north sag of Liaoxi depression. The application clearly shows that SLPR is of great value to lithology exploration.