煤矿沉陷水域水体深度多源数据融合反演  

Multi-source Data Fusion Inversion of Water Depth in Coal Mine Subsidence Waters

作　　者：王明达 徐良骥[1,2] 叶伟 刘潇鹏 WANG Mingda;XU Liangji;YE Wei;LIU Xiaopeng(School of Spatial Information and Geomatics Engineering,Anhui University of Science and Technology,Huainan,Anhui 232001,China;National Key Experiment of Mining Response and Disaster Prevention and Control in Deep Coal Mine,Anhui University of Science and Technology,Huainan,Anhui 232001,China)

机构地区：[1]安徽理工大学空间信息与测绘工程学院,安徽淮南232001 [2]安徽理工大学深部煤矿采动响应与灾害防控国家重点实验室,安徽淮南232001

出　　处：《遥感信息》2022年第1期88-93,共6页Remote Sensing Information

基　　金：国家自然科学基金项目(41472323);安徽省重点研发计划国际科技合作专项(201904b11020015)。

摘　　要：为准确高效地获取沉陷水域水深数据,提出了一种煤矿沉陷水域水体深度多源数据融合反演模型。基于Sentinel-2遥感数据和无人船水深实测数据,分别构建单波段模型、多波段模型和对数比值模型。利用支持向量机(support vector machine,SVM)分别对三种经典模型中拟合优度最高的模型进行优化。结果表明:蓝色单波段模型、多波段模型、蓝色与绿色比值对数模型三种模型的拟合优度最靠前,分别为0.635、0.675、0.517;均方根误差分别为1.32 m、1.24 m、1.53 m。对经典模型进行SVM优化后,反演精度均有提高,拟合优度分别为0.724、0.759、0.522;均方根误差分别为1.15 m、1.07 m、1.52 m;SVM多波段模型的反演精度最高。研究表明,利用多源数据融合反演沉陷水域水体水深能够达到一定的精度要求,可以为矿区沉陷区水域治理提供依据。In order to obtain the water depth data of subsidence waters accurately and efficiently,this paper presents a multi-source data fusion inversion model of water depth in coal mine subsidence waters. Based on Sentinel-2 remote sensing data and unmanned ship water depth measurement data,a single-band model,a multi-band model and a logarithmic ratio model are constructed,respectively. The model with the highest fit degree of three classical models is optimized by support vector machine(SVM). The results show that:the blue single-band model,the multi-band model,and the blue-to-green ratio logarithmic model have the highest goodness of fit,which are 0.635,0.675,and 0.517,respectively;the root mean square error is 1.32 m,1.24 m,1.53 m,respectively. The classic model is optimized by SVM,the inversion accuracy is improved,and the goodness of fit is 0.724,0.759,0.522;the root mean square error is 1.15 m,1.07 m,and 1.52 m;the inversion accuracy of the SVM multi-band model is the highest. The results show that:using multi-source data fusion inversion of water depth in subsidence area can achieve a certain accuracy requirements,which can provide a basis for water governance in mining subsidence area.

关 键 词：沉陷水域 多光谱遥感 水深反演 Sentinel-2 支持向量机 

分 类 号：X87[环境科学与工程—环境工程] TD327[矿业工程—矿井建设]