基于LS-SVM的无人机匹配点云数据修补  

UAV Matching Point Cloud Data Repair Based on LS-SVM

作　　者：邢鹏威 唐诗华 张曦 张跃 何广焕 XING Pengwei;TANG Shihua;ZHANG Xi;ZHANG Yue;HE Guanghuan(Guangxi Key Laboratory of Spatial Information and Geomatics,Guilin 541006,Guangxi,China;College of Geomatics and Geoinformation,Guilin University of Technology,Guilin 541006,Guangxi,China;College of Mining and Geomatics,Hebei University of Engineering,Handan 056038,Hebei,China)

机构地区：[1]广西空间信息与测绘重点实验室,广西桂林541006 [2]桂林理工大学测绘地理信息学院,广西桂林541006 [3]河北工程大学矿业与测绘工程学院,河北邯郸056038

出　　处：《水力发电》2020年第11期51-55,60,共6页Water Power

基　　金：国家自然科学基金资助项目(41864002);广西空间信息与测绘重点实验室基金项目(16-380-25-25、16-380-25-13、15-140-07-05)。

摘　　要：针对高山区过滤后点云数据缺失问题,提出一种基于LS-SVM的点云漏洞修补方法。以典型高山区地形为试验案例,采用4种常规插值方法与LS-SVM预测方法对数据进行处理分析,将处理数据与CORS动态测量获取的实测数据进行比较研究。研究表明,与4种常规方法相比,采用LS-SVM算法预测出的点云所构建的DEM模型精度有较大提高,模型MAE=-0.148 m、RMSE=0.250 m、R 2=0.9995,能够实现1∶500 A级高山区的高精度DEM生产,同时也增强了DEM在水利、建筑等行业设计初期的应用价值。Aiming at the problem of point cloud data missing after filtering in mountain area,a point cloud vulnerability repair method based on LS-SVM is proposed.Taking a typical mountainous terrain as the experimental case,four conventional interpolation methods and LS-SVM prediction method are used to process and analyze the data respectively,and the processed data is compared with the measured data obtained from CORS dynamic measurement.The research shows that,compared with four conventional methods,the accuracy of DEM model constructed by point cloud predicted by LS-SVM algorithm is greatly improved with MAE=-0.148 m,RMSE=0.250 m and R 2=0.9995.This model can achieve high-precision DEM production of 1∶500 A scale in the mountain area,and enhance the application value of DEM in the early design stage of water resources,construction and other projects.

关 键 词：DEM数据 匹配点云 LS-SVM算法 插值 

分 类 号：P231[天文地球—摄影测量与遥感]