中文摘要：

摘要大数据是海量、高增长率和多样化的信息资产，是未来找矿靶区预测的不可或缺的技术。大数据一三维成矿预测研究面临机遇与挑战，其涉及的地学大数据除了数据来源众多、比例尺不同、数据量大、非结构化管理、时效性强、空间数据与非空间数据协同管理等复杂特点外，还必须具有适应进行三维建模及空间分析的数据结构。本文分析了地学空间大数据的特点，对多源地学综合信息的管理需求进行研究，参考国家及行业标准，建立了可满足三维成矿预测需求的多源地学空间数据库模型，并依据实际划分为勘查控制钻孔地质数据库、空间属性数据库和地球物理数据库，各数据库可在多源地学空间索引库的支持下协同工作。本文以大数据应用的典型实例——钟姑矿田作为研究对象，系统收集了矿田内勘查成果资料，建立了钟姑矿田多源地学空间数据库，并在此基础上进行了控矿要素的有效提取，可进一步支持三维成矿预测。研究结果表明，本文提出的多源地学空间数据库可有效管理地学空间大数据，是大数据一三维成矿预测的重要解决方案，是进行三维成矿预测的重要数据支持。

英文摘要：

Big data is massive, high growth rate and diversified information assets. It is an indispensable technology for the future prediction of ore prospecting targets. The research on the big data-3D metallogenic prediction is facing opportunities and challenges. The geo-spatial big data have some complex characteristics such as multitudinous data sources, different scales, large data volume, unstructured management, timeliness, co-management with spatial data and non-spatial data. Besides, data structure must be adapted to the 3D modeling and spatial analysis. This paper analyzes the characteristics of geo-spatial data and studies the requirements of muhi-source geoscience information management. Based on the national and industry standards, a multi-source geoscience spatial database model is established which can meet the requirements of 3D mineralization prediction. There are several components such as control of drilling geology database, spatial attribute database and geophysical database. The multi-source geo-spatial index library is set up to support those databases work together. In this paper, a typical example of big data application is used as the research object. The data of the exploration results in the ore field are collected and the spatial database of the multi-souree geology of Zhonggu ore field is established. On this basis, the effective extraction of ore- controlling elements, ean further support the 3D mineralization predietion. The results show that the proposed multisource geespatial database can effectively manage big data of geoscience, which is an important solution for big data-3D metallogenic prediction and it is an important support for 3D metallogenic prediction.