Artificial neural network technique is applied in coal seismic exploration, which has been used to detect faults and to interpret both macrostructures and thickness of coal beds, and has improved the level and precision of seismic data interpretation in coal field.

The production ability and efficiency for mechanized coal faces of lager-scale mines depend on the detecting degree of mining structures. It is a major task for geological exploration in coalfields to detect minor structures in mining blocks. 3D high resolution seismic prospecting is an effective measure for solving this problem.

煤矿瓦斯灾害是一种动态地质现象，受多种因素的影响。传统的研究方法仅考虑个别影响因素，预测结果难以满足实际需要。本文提出一种全新方法，利用地震P波对裂缝性地层所表现出的方位各向异性特征，提取煤层和围岩的裂隙属性。利用GIS作为平台，把煤层和围岩的裂隙属性、煤层厚度、断层 及其它构造分布、煤层埋藏深度、煤层的倾角与露头位置、煤化程度等多源信息进行融合和综合分析后，建立服务于煤矿生产的瓦斯富集带预测模型。

Main functions of the identification system of fault with seismic data are introduced. In this ystem, the computer technique, fuzzy mathematics' theory, pattern recognition technique, fractal technique and artificial neural network technique are combined and used in seismic data structure terpretation. Utilizing this system, the distribution of minor faults with about 5m throw (other minor structure) in mining areas can be detected in the area of good seismic conditions and accurate geological information can be provided for mining excavation.

本文讨论了三种多项式拟合方法在煤田地震勘探中的应用情况。它们可以消除动校拉伸对波形的影响以及普通叠加的低通滤波效应，保持了振幅的真实形态和较高的频率信息，能得到较高分辨率的零炮检距剖面，有利于地震资料的构造解释和岩性研究。