基于证据信度模型的致矿地质异常信息提取与集成

地质异常信息是找矿靶区圈定的重要依据，适当的异常提取方法和决策模型能从多元地学数据中获得准确可靠的致矿地质异常信息。以祁连山玉石沟-川刺沟铜铅锌矿区为研究区，应用证据信度模型评价岩性、断裂、遥感蚀变、化探元素的空间信息：首先由各致矿因子与矿点分布的关系计算证据层的可信度、怀疑度、不确定性3项指标，并采用分形、主成分分析等非线性提取方法剔除假异常；然后通过Dempster法则集成得到成矿综合信度图。结果显示,96%的矿点位于信度异常值区域（0.17~0.60），且异常分布与实地勘查信息一致。模型增大证据因子输入参数的可靠性，减弱了证据因子间独立性假设的影响，有效地提取了单一因子的致矿地质异常及综合异常。

Extraction and Integration of Geo-Anomalies Associated with Mineralization Based on Evidential Belief Model

Geo-anomalies are significant foundation of mineral prospectivity,appropriate methods of extraction of geo-anomalies and a decisionmodel can be applied to get accurate and reliable geo-anomalies from various data.In the study,evidential belief model is used to evaluate the spatial information of lithological, belts, alteration of remote sensing and geochemical elements in the Cu-Pb-Zn mineral area of Yushigou-Chuancigou,Qilianshan:The indicators(belief,disbelief,uncertainty)of evidences are calculated according to the space relationship between distribution of mineral depositsand evidences,and methods such as fractal and PCA are adopted to eliminate the fake anomalies,then cmineralization beliefs are calculated based on Dempster rules.The result shows that 96% deposits located in the site of the peak of belief, and the distribution of geo-anomaly is consistent to the reality. The model weakens the impact by conditional independence when integrated,and conclusively increases the accuracy and efficiency of extraction and integration of geo-anomalies.