基于地质约束条件的找矿信息加权融合方法研究

Influenced by interaction, superimposition and coupling of multi-geological activities, metallogenic events are often with complexity, non-linearity, self-similarity and scale invariance. Quantitative prediction and assessment for mineral resources are necessary to significantly improve the guidance of metallogenic theories. In order to comprehensively support and increase rationality and reliability of the prediction and assessment, geological theory and geo-information analytical methodology should be organically integrated according to delicate and systematical studies. Focusing on the multi-factor coupling mechanism of complex mineralization processes, the currently proposed study firstly employs non-linear theory to spatially and quantitatively investigate geological constrains on mineralization, and consequently defines weights of controlling factors. Secondly, this study integrates various weighted fusion theory and proposes a general weighted fusion model with enhanced adaptability and effectivity. Finally, integrating the quantitatively weighted geological constrains and the general weighted fusion model, the study intends to propose a geologically weighted predictive information fusion methodology which can rationally analyze, identify and describe predictive information within in complex geological environment.

成矿地质事件受多个地质活动混合、叠加、耦合的作用影响，具有复杂性、非线性、自相似性、尺度不变性。矿产资源定量预测与评价需加强成矿理论指导，通过开展细致研究，将地质概念模型同多元找矿信息分析算法有机融合，以支撑预测与评价的合理性、可靠性。本项目从复杂成矿地质过程多因耦合机制出发，通过非线性理论方法空间定量化研究成矿地质约束条件，定义预测要素权重，研究信息加权融合算法模型，提高加权分析方法的可适应性与有效性，最终建立复杂地质背景条件下分析、识别并合理表述找矿信息的加权融合方法，提出结合地质实际与数据信息结构特征的找矿信息融合广义概念模型。

成矿地质事件受多个地质活动混合、叠加、耦合的作用影响，具有复杂性、非线性、自相似性、尺度不变性。矿产资源定量预测与评价需加强成矿理论指导，本项目通过开展细致研究，将地质概念模型同多元找矿信息分析算法有机融合，从复杂成矿地质过程多因耦合机制出发，研究了控矿地质因素权重定量化、找矿信息加权分析以及地质约束条件下找矿信息加权融合。提出了新的地质约束条件的表达方式：（1）建立了基于构造测量数据的地质条件约束下缓冲区分析方法技术；（2）研发了基于专家知识的地质条件约束下的机器学习方法，为岩性单元识别和矿物识别提供了现实可行的技术流程。进而，将空间U统计方法与奇异性指数算法进行融合，构建了地质约束下的信息挖掘分析窗口，提出了新的地质约束条件下的找矿信息融合方法。