基于DBD-Net的InSAR矿区开采沉陷盆地检测方法  

InSAR mining subsidence basin detection method based on DBD-Net

作　　者：李涛 邹英杰 范洪冬[2] 吝涛 LI Tao;ZOU Yingjie;FAN Hongdong;LIN Tao(Geophysical Prospecting and Surveying Team of Shandong Bureau of Coal Geology,Jinan 250104,China;Key Laboratory of Land Environment and Disaster Monitoring,Ministry of Natural Resources,China University of Mining and Technology,Xuzhou 221116,China)

机构地区：[1]山东省煤田地质局物探测量队,山东济南250104 [2]中国矿业大学自然资源部国土环境与灾害监测重点实验室,江苏徐州221116

出　　处：《煤矿安全》2024年第4期177-186,共10页Safety in Coal Mines

基　　金：山东省煤田地质局科研专项资助项目(鲁煤地科字(2022)46号);国家重点研发计划资助项目(2022YFE0102600)。

摘　　要：目前通过合成孔径雷达干涉测量(Interferometric Synthetic Aperture Radar,InSAR)检测开采沉陷盆地主要依靠地下开采资料或人为目视解译,针对这一问题,提出一种针对大范围InSAR干涉图的开采沉陷盆地检测网络(Deformation Basin Detection Network,DBD-Net);同时,为了训练该网络,利用矿区的真实差分干涉图数据和模拟干涉数据建立了开采沉陷盆地样本库,在神东矿区和兖州矿区各选取3幅不同时间基线的差分干涉影像对网络性能进行验证。结果表明:DBD-Net在大范围InSAR干涉图中对开采沉陷盆地的平均检测准确度为81.87%,漏检和误检区域大多是噪声严重污染和特征不明显的区域。At present,the detection of mining subsidence basins by interferometric synthetic aperture radar(InSAR)mainly relies on underground mining data or human visual interpretation.To solve this problem,this paper proposes a deformation basin detection network(DBD-Net)for large-scale InSAR interferograms.At the same time,in order to train the network,a sample database of mining subsidence basins is established by using real differential interferogram data and simulated interferogram data.In Shendong Mining Area and Yanzhou Mining Area,three differential interference images with different time baselines were selected to verify the network performance.The results show that the detection accuracy of deformation basin detection network(DBD-Net)in large-scale InSAR interferograms for mining subsidence basins is 81.87%.Most of the missed and false detection areas are areas with serious noise pollution and unclear characteristics.

关 键 词：INSAR 卷积神经网络 开采沉陷 变形检测 DBD-Net 

分 类 号：TD325.4[矿业工程—矿井建设]