Decision Tree算法在地质建模岩性识别中的研究

Application of Decision Tree Algorithm in Lithologic Identification of Geological Modeling

针对地质建模时,人工识别山体内部岩石的局限性、低效率且易受主观因素影响等问题,提出了基于地震波反射信号的岩石类型自动识别技术。通过处理地震波反射信号获得岩石力学参数,采用Decision Tree ID3算法,提取岩石密度、波速、弹性模量、剪切模量,构建岩性识别模型分类器。通过该分类器对某山体内部岩石类型进行判断,研究结果证明:研究区内部多为辉长岩,玄武岩最少,通过模型分类结果与研究区真实地质对比分析,玄武岩正判率达到93%,安山岩、闪长岩正判率达到100%,花岗岩正判率达到88%,决策树建立的分类器模型能够基于地震波反射信号高效、准确地识别岩石岩性。

For the geological modeling,the problems of artificially recognizing the limitations of rock inside the mountain,low efficiency,and vulnerability to subjective factors,etc.,propose an automatic rock type recognition technology based on seismic wave reflection signals.The rock mechanics parameters were obtained by processing the seismic wave reflection signals,and the lithology identification model classifier was constructed by using the Decision Tree ID3 algorithm to extract the rock density,wave velocity,elastic modulus,and shear modulus.By using this classifier to determine the rock types within a mountain,the research results show that the interior of the study area is mainly gabbro,and the basalt is the least.Through the model classification results and the real geology of the study area,the positive rate of basalt is 93%.And esite the positive rate of diorite is 100%,and the positive rate of granite is 88%.The classifier model built using decision trees can efficiently and accurately identify rock lithology based on seismic reflection signals.