【作者】 谭征兵；

【导师】 田明中； 胡凤翔；

【作者基本信息】 中国地质大学（北京） ， 第四纪地质学， 2003， 博士

【摘要】 矿产资源评价与成矿预测这一地学中的重要研究对象随着空间信息技术的迅猛发展和广泛应用，正朝着信息化方向迈进。以GIS为主要支撑技术，分析矿产资源的空间分布，研究多元成矿信息定量模型，发展矿床模型专家系统成为矿产资源评价及预测的基本趋势。本文紧扣矿产资源GIS评价工作中地学数据模型创建、成矿信息GIS提取与集成技术、矿产资源综合评价模型的构造这三个关键问题，以面向对象的程序设计思想，创造性的利用ActiveX技术和GIS功能组件MapObjects，在Visual C++开发环境下，重点研制了矿产资源GIS评价研究中网格单元划分、地质变量智能化提取、空间数据库管理和成矿预测BP模型等核心模块，实现了矿产资源评价、GIS技术和人工神经网络技术的紧密结合，开发出一套界面友好、功能完善、平台独立的矿产资源GIS评价系统。 文中系统分析与总结了矿产资源综合评价及成矿预测的通用工作流程，采用面向对象的软件工程技术和方法，将矿产资源综合评价工作表达为计算机软件工程，并在完成矿产资源GIS评价需求分析和总体结构设计的基础上，详细划分出系统的六大功能模块：文件输入模块、评价单元划分模块、空间分析模块、地质变量智能提取模块、定位预测模块和属性管理模块。 首次将美国ESRI公司的MapObjects应用于矿产资源评价系统的研制开发中，使评价系统摆脱了对基础GIS平台的依赖，具有较好的独立性和可移植性。以矢量数据结构为主，实现了对点、线、面地质体的GIS空间分析功能，以便对多元矢量数据信息进行叠加分析和相关分析，发现和归纳研究区域的成矿规律，建立成矿地质模型。并提供人机交互式定制网格单元的功能，以网格单元作为基本评价单元，开发出多属性、多变量的成矿信息智能化提取模块，完成对矢量图形数据、定量专题数据的自动数值转换。 为有效的组织和管理属性数据，系统采用了DAO数据库访问系统来访问和管理属性数据，并以列表视图控件为基础建立了属性数据浏览视图，允许用户在系统运行的过程中动态创建或选择DAO数据源为数据库应用程序提供要管理的属性数据对象。实现了空间实体数据和属性数据的关联，以及对属性数据的浏览、编辑、查询和管理功能。 论述了人工神经网络BP模型在矿产资源综合评价及成矿预测中应用的可行性和优越性，实现了BP网络的通用算法，建立了矿产资源综合评价及成矿预测BP模型，从而可以对成矿信息进一步进行智能化知识发现和信息挖掘，自动评估各地质变量对成矿的贡献，得到区域性的成矿规律和成矿模式，并圈定出成矿靶区。并提供友好的人机交互式界面，使用户可以自己创建、设计和管理成矿预测BP模型。 在矿产资源GIS评价系统中，集成矿产资源评价方法、GIS技术和人工神经网络技术，实现了矿产资源综合评价及成矿预测的工作流程。以计算机可视化技术便捷、直观的展现出矿产资源综合评价的全过程；利用GIS强大的数据管理和空间分析功能，有效的解决了多源、多类型、多属性地质数据的处理和分析问题；利用人工神经网络技术突破了统计数学模型对矿产资源综合评价及成矿预测的约束和限制，为矿产资源综合评价提供了一种新思路和有效的方法。 本文着眼于矿产资源GIS评价工作中的难点和关键问题，强调矿产资源评价方法、GIS技术和人工神经网络技术的紧密结合，实现了对多层次、多类型、多属性信息的提取和集成，为矿产资源综合评价提供了GIS软件系统技术支撑，对推动矿产资源评价工作的信息化发展具有重要意义。更多还原

【Abstract】 With the great development and wide application of spatial informational technologies, mineral resources assessment and metallogenic prognosis, an important branch of geosciences, is striding forward in the informational direction. Supported by the GIS technology, the research field of the mineral resources assessment and metallogenic prognosis is tending to develop Expert Systems of mineral deposit model by dealing with the temporal and spatial distribution of mineral resources and building quantitative models of multiple metallogenic information. Aiming at the three pivotal problems in research on mineral resources assessment with GIS about designing data models of geology, GIS technologies for extracting and integrating mineralizing information, synthetic assessment models of mineral resources, and being guided by the object-oriented programming theory, the GIS-based Mineral Resources Assessment system was developed with a user-friendly interface, perfect functions and an independent platform. The ActiveX technology, the functional component of GIS named MapObjects and the artificial neural network technology were adopted in the developing environment of Visual C++ to fulfill the system whose core modules of gridding partition, intelligentized extraction of geologic variables, spatial database management and BP model for metallogenic prognosis were developed emphatically.The universal process which was expressed in a computer software project with the object-oriented programming technology and methods was analyzed and concluded systematically for mineral resources synthetic assessment and metallogenic prognosis. Furthermore, the GIS-based Mineral Resources Assessment System was divided in detail after demand analysis and structural design into six functional modules: the files input module, gridding partition module, spatial analysis module, intelligentized extracion of geologic variables module, orientation prediction module and attribution management module.The functional component of GIS named MapObjects published by America ESRI Corporation was applied unprecedentedly in the development of mineral resources assessment systems, which make the system independent of basic GIS platform and transplantable. The GIS-based Mineral Resources Assessment System deals with vector data structure mainly. And its spatial analytic functions such as overlay analysis and buffer analysis that can be used to extract multiple variables and information from geologic bodies organized in map layers in point, line or polygon shape are in favor of discovering ore-forming factors, summarizing regional metallogenic regularities and constructing geologic metallogenic models. Users can use the system to customize sizes of grid units and partition digital maps in grid. And users can also use the module of intelligentized extraction of geologic variables to extract and integrate metallogenic information included in various geologic attributions of variables in order to quantify vector graphic data in numerical value.The DAO database access system was adopted to access and manage attribution data effectively in the GIS-based Mineral Resources Assessment System Based on a list control, a view was created to show users attribution data. Users can create or select dynamically a DAO data source when the system is in a running process to define an attribution data object for the application program Moreover, the system provides users some functions to browse, edit, query and manage attribute data.It is feasible and advantaged to apply a BP model, one of artificial neural networks, to evaluate synthetically mineral resources and do metallogenic prognoses. And a BP model formetallogenic prognosis can be constructed as the general algorithm of BP network was programmed. Users can create, design and manage BP models for metallogenic prognoses by interacting with a computer through a user-friendly interface. When BP models are taken advantage of to do mineral resources assessment and metallogenic prognoses, metallogenic informati更多还原