沉积与特提斯地质2025，Vol.45Issue(4)：737-750,14.DOI:10.19826/j.cnki.1009-3850.2025.11004
基于残差卷积−异常检测混合架构的智能找矿预测模型构建及其在腾冲−小龙河锡矿集区的应用
Development of a residual convolution-anomaly detection hybrid model for intelligent mineral prospectivity mapping:A case study from the Tengchong-Xiaolonghe tin district
摘要
Abstract
In current large-scale(1∶50 000 or finer)mineral prospectivity modeling scenarios,several critical challenges persist,including disconnection from metallogenic theories,non-uniform sample distribution,and low model training efficiency.These issues often result in reduced anomaly detection accuracy,an increased risk of missing concealed ore bodies,and diminished interpretability of prediction results in practical applications,thereby hindering rapid mineral discovery and exploration breakthroughs.This research focused on the Tengchong-Xiaolonghe tin polymetallic ore district,and a metallogenic system theory-guided sample construction mechanism to harmonize domain knowledge with data-driven approaches was established.On this basis,a novel deep–shallow hybrid architecture was developed.The model leverages deep residual convolutional networks to capture complex nonlinear features,and integrates shallow one-class support vector machines to enable efficient fusion of multi-source geoscientific data and rapid identification of mineralization anomalies.This approach effectively alleviates challenges related to imbalanced sample distribution and low training efficiency under large-scale scenarios.The prediction results indicate that the high-probability prospecting targets exhibit a strong spatial correspondence with 86.7%of the known deposits(occurrences)and show a pronounced coupling relationship with the contact zones between Yanshanian granitoids and wall rocks,as well as with the distribution of greisen veins.Accordingly,11 prospective mineralization zones were delineated,and inspection holes conducted within these zones successfully intercepted Sn-W mineralization.These results,corroborated by both geological evidence and engineering verification,confirm that the residual convolution-anomaly detection hybrid architecture provides reliable and practically meaningful predictions.Moreover,it achieves higher training efficiency while maintaining nonlinear representation capability comparable to that of deep autoencoder networks,demonstrating strong potential for practical applications.关键词
智能找矿预测/腾冲–小龙河锡矿/多源地学信息融合/异常检测/深–浅层混合神经网络Key words
intelligent mineralization prediction/Tengchong-Xiaolonghe tin deposit/multi-source geoscientific data fusion/anomaly detection/deep-shallow hybrid neural network分类
天文与地球科学引用本文复制引用
周放,张玙,周清,马龙,梁虹,韩志婷..基于残差卷积−异常检测混合架构的智能找矿预测模型构建及其在腾冲−小龙河锡矿集区的应用[J].沉积与特提斯地质,2025,45(4):737-750,14.基金项目
中国地质调查局地质调查项目(DD20240070,DD20242602,DD20240069,DD20242494) （DD20240070,DD20242602,DD20240069,DD20242494）
深地国家科技重大专项(2024ZD100320703) （2024ZD100320703）
