Development of advising system for exploration of geology and mineral resources using artificial intelligence technique.

这项研究的目的是设计一个小型的计算机辅助系统，通过实地地质和实验数据来帮助估计地质地体及其所含矿物资源的形成过程，并评价更普遍和实际使用的可能性。为此，我们采取了三种方法：(1)制定和评估基于规则的系统，该系统可针对现场地质中的小型实际问题进行咨询（自下而上方法）；(2)设计一个处理地质地体及其含矿资源形成过程估计中的一般问题的系统（自顶向下方法）；(3)制作一个用于输入和分析观测数据的底部工具。对于方法(1)，我们建立了一个基于规则的专家系统，帮助用户估计火山碎屑岩的侵位模式。在开发过程中，我们使用了Hyperpress Publishing Corp.的专家shell Intelligent developer，完成了一个系统Volcaniclast，至少可以用于本科生野外火山学教育。方法(2)系统地整理了现有地质地体及其含铀资源形成过程知识，设计了框架结构的知识体系。我们还将一个真实的铀矿的知识融入到系统中，明确了系统的有效性和存在的问题。在方法(3)中，我们构建了一个系统，该系统可以使用数字化仪读取地貌数字数据并分析地表面积、体积和坡度。通过所有这些方法，我们积累了重要的技术诀窍，可以开发一个大型系统，用于勘探地质和矿产资源的一般用途。

This study aims to design a small, computer-aided system which helps to estimate formative processes of a geological terrane and its bearing mineral resources through field geological and experimental data, and to evaluate the possibility for more general and practical use. For this purpose, we made three approaches : (1) making and evaluating a rule-based system which consults on a small, practical problem if field geology (bottom-up approach) ; (2) designing a system which treats a general problem in estimating formative processes of geological terrane and its bearing mineral resources(top-down approach) : (3) making a bottom-ended tool for inputting and analyzing observational data. As for approach (1), we made a rule-based expert system, which helps a user to estimate a mode of emplacement of volcaniclastic rocks. For developing this, we used an expert shell Intelligent developer (Hyperpress Publishing Corp.) and completed a system Volcaniclast, which can at least be used in education of field volcanology for undergraduate students. Concerning approach (2), we systematically reorganized available knowledge of formative processes of a geological terrane and its bearing uranium resources and designed a frame-structured knowledge-based system. We also incorporated knowledge of a real uranium mine into the system and clarified the validity and problems of the system. In approach (3), we constructed a system which can read digital data of geomorphology using a digitizer and analyze surface area, volume, and slope. Through all these approaches, we accumulated significant know-how for developing a large system which reast general purposes in exploring geology and mineral resources.