摘要
机载激光雷达在测绘、勘探等领域有广泛的应用,其数据处理联合激光雷达测距数据和姿态位置信息,解算获得扫描目标的三维坐标并形成三维点云图。为了满足机载激光雷达点云解算的实时性要求,采用基于软硬件协同的设计方法,设计、实现了激光点云解算的SoC。通过使用基于AXI-4的DMA高速传输方式,运用流水线优化和存储优化方法,实现了高性能的硬件加速器。实验结果表明,提出的激光点云解算的SoC能够满足机载平台的实时性处理要求。
Airborne lidar has a wide range of applications in mapping, exploration and so on. Its data processing uses the ranging data of Li DAR and the position and orientation information to obtain high precision surface models. In order to meet the real-time requirements of airborne LiDAR points position cloud calculation, a software-hardware co-design based method is used to design and implement the SoC of points position cloud calculation. A high-performance hardware accelerator is realized by using the AXI-4-based DMA high-speed transmission and using pipeline optimization and storage optimization methods. The experimental results show that the proposed SoC of points position cloud calculation can meet the real-time processing requirements of the airborne platform.
引文
[1]黎荆梅,周梅,李传荣.阵列推扫式机载激光雷达三维点云解算方法研究[J].遥感技术与应用,2013,28(6):1033-1038.
[2]DABNEY P,HARDING D,ABSHIRE J,et al.The slope imaging multi-polarization photon-counting lidar:development and performance results,2010[C].IEEE International Geoscience and Remote Sensing Symposium,2010:653-656.
[3]YU A,KRAINAK M,HARDING D,et al.Development effort of the airborne lidar simulator for the lidar surface topography(LIST)Mission[C].Proceedings of SPIE.The International Society for Optical Engineering,2011.
[4]郭商勇,胡雄,闫召爱,等.国外星载激光雷达研究进展[J].激光技术,2016(5):772-778.
[5]李然,王成,苏国中,等.星载激光雷达的发展与应用[J].科技导报,2007(14):58-63.
[6]MARTIN G.Overview of the MPSoC design challenge,2006[C].Design Automation Conference.ACM/IEEE,2006:274-279.
[7]陈芳园,张冬松,王志英.异构多核处理器体系结构设计研究[J].计算机工程与科学,2011(12):27-36.
[8]马飞,刘琦,包斌.基于FPGA的AXI4总线时序设计与实现[J].电子技术应用,2015,41(6):13-15,19.
[9]曹彦荣,张锐.DMA传输与Cache一致性分析[J].硅谷,2014,7(8):39-40.