无人机机载激光雷达系统航带拼接方法研究  

Research of the Unmanned Plane Airborne Lidar System Strip Stitching Method

作　　者：赵大伟[1,2] 裴海龙[1,2] 丁洁[1,2] 张铖[1,2] 

机构地区：[1]华南理工大学自主系统与网络控制教育部重点实验室,广东广州510640 [2]华南理工大学自动化科学与工程学院,广东广州510640

出　　处：《中国激光》2015年第1期315-322,共8页Chinese Journal of Lasers

基　　金：国家973计划(2014CB845301/2/3);国家自然科学基金(61174053);华南理工大学中央高校基本科研业务费自助项目(2014ZP0021)

摘　　要：为了减少机载激光雷达(Li DAR)系统中系统误差和随机误差造成的航带间三维(3D)空间偏移,提高数据精度,选取基于数据驱动的"六参数"航带平差方法,实现无人机机载激光雷达系统的航带拼接。在分析了机载激光扫描系统的数据特征的基础上利用改进的3D正态分布变换(3D-NDT)进行航带配准,得到航带间的变换关系参数。通过具体实验对常见的迭代最近点(ICP)算法与3D正态分布变换算法进行比较,验证了该方法实现航带拼接,具有速度快、精度高、稳健性好等特点,非常适合于工程实际应用。In order to avoid the misregistration in three- dimensional （3D） space between flight strips caused by the system error and random error in airborne light detection and ranging （LiDAR） and improve the data accuracy, a data- driven“six parameters”strip adjustment method is selected for the realization of airborne LiDAR system strips registration. An improved 3D normal distributions transform （3D- NDT） is used for the registration of strips to obtain the transformation parameters between the strips on the basis of the analysis of the data of LiDAR. A detailed comparison between the popular iterative closest point （ICP） algorithm and the 3D normal distributions transform is presented through specific experiment. The result shows that this approach can realize flight strip registration rapidly with high accuracy and stronger robustness, which is very suitable for practical engineering.

关 键 词：激光光学 激光雷达 航带平差 正态分布变换 点云配准 

分 类 号：TN958.98[电子电信—信号与信息处理]