单站点云坐标转换参数的线性拟合解算方法  

Linear fitting method for calculateing parameters of coordinate-converting of single-station TLS point clouds

作　　者：刘科利[1] 姚吉利[1] 马宁[1] 贾象阳 曹俊茹[1] 

机构地区：[1]山东理工大学建筑工程学院,山东淄博255049

出　　处：《测绘科学》2015年第8期13-17,17,共5页Science of Surveying and Mapping

摘　　要：针对三维坐标转换公式是非线性的,会造成在旋转角度任意大小时影响平差计算转换参数的问题,该文根据罗德里格矩阵的性质,建立坐标转换的线性拟合模型,计算6个拟合参数,并计算坐标转换参数。利用三维激光扫描仪获得物体形状信息后,需要将点云坐标系统转换成工程测量中使用的国家或地方坐标(指定坐标)系。多站实验后的数据分析表明,平面点位误差为21.27mm,高程中误差为18.96mm,证明此方法可靠有效。An automatic building detection method based on fusing LiDAR and multispectral imagery was pro- posed in this paper. The proposed method first separated the building and tree points from original LiDAR point cloud～ then different features were extracted from LiDAR and multispectral imagery, such as surface curvature, intensity and NDVI value. These features were combined together to form the feature vector; finally, using fea- ture vector derived from LiDAR and multispectral imagery as the input for support vector machine, an automatic building detection result could be acquired. Experiment results showed that the proposed method could efficiently fuse LiDAR and multispectral imagery for automatic building detection based on support vector machine.

关 键 词：扫描坐标系 指定坐标 线性拟合模型 平面点位中误差 高程中误差 

分 类 号：P232[天文地球—摄影测量与遥感]