[1]王奕辉,余捷,于廷海,等.基于Scan Context回环检测与GICP精匹配改进的SLAM算法[J].福建理工大学学报,2025,23(06):541-549.[doi:10.3969/j.issn.2097-3853.2025.06.005]
　WANG Yihui,YU Jie,YU Tinghai,et al.Improved SLAM algorithm based on Scan Context loop detection and GICP fine matching[J].Journal of Fujian University of Technology;,2025,23(06):541-549.[doi:10.3969/j.issn.2097-3853.2025.06.005]
点击复制

基于Scan Context回环检测与GICP精匹配改进的SLAM算法()
分享到：

《福建理工大学学报》[ISSN:2097-3853/CN:35-1351/Z]

卷:
第23卷
期数:
2025年06期
页码:
541-549
栏目:
出版日期:
2025-12-25

文章信息/Info

Title:
Improved SLAM algorithm based on Scan Context loop detection and GICP fine matching
作者:
王奕辉; 余捷; 于廷海; 戴村供
福建理工大学机械与汽车工程学院
Author(s):
WANG Yihui;  YU Jie;  YU Tinghai;  DAI Cungong
School of Mechanical & Automotive Engineering, Fujian University of Technology
关键词:
激光SLAM; 回环检测; 点云配准; GICP
Keywords:
laser SLAM; loop detection; point cloud registration; GICP
分类号:
TP242.6
DOI:
10.3969/j.issn.2097-3853.2025.06.005
文献标志码:
A
摘要:
为解决同步定位与建图(SLAM)技术在构建点云地图的漂移问题并减少轨迹误差，提出GS-LeGO算法。该算法基于LeGO-LOAM算法的框架，融合扫描上下文优化(Scan Context)与广义迭代最近点(GICP)精匹配。使用Scan Context提取的全局描述符进行循环检测，通过GICP点云配准进行配准，对地图修正得到最终的位姿估计，实现精准回环检测和提高定位精度。使用KITTI数据集的00序列和07序列数据集进行测试。实验结果表明，同LeGO-LOAM算法相比，GS-LeGO算法解决了在急剧转弯时点云地图漂移的问题，回环效果较好，构建的点云地图更为精准，运动轨迹与真实轨迹重合度更高，估计轨迹长度更接近真实轨迹长度。与LeGO-LOAM算法的绝对位姿误差相比，GS-LeGO算法在数据集00中误差平均值减少了24.22%、均方根误差减少了约31.28%；在数据集07中误差平均值减少了43.30%、均方根误差减少了约45.28%。相比于实时激光雷达测程和测绘算法，GS-LeGO算法在数据集00中误差平均值减少了65.10%、均方根误差减少了约68.10%，在数据集07中误差平均值减少了15.38%、均方根误差减少了约13.43%。
Abstract:
In order to solve the drift problem of simultaneous localization and mapping (SLAM) technology in constructing point cloud maps and reduce trajectory errors, a GS-LeGO algorithm, which integrates Scan Context optimization and generalized iterative nearest point (GICP) fine matching with lightweight ground optimized radar odometer and mapping (LeGO-LOAM) algorithm, is proposed. The algorithm uses the global descriptor extracted by Scan Context for cyclic detection, registers the map through GICP point cloud registration, and obtains the final pose estimation to achieve accurate loop detection and improve positioning accuracy. The algorithm also uses KITTI data set 00 sequence data set and 07 sequence data set for testing. Experimental results show that, compared with LeGO-LOAM algorithm, GS-LeGO algorithm solves the problem of point cloud map drift during sharp turns, and the loop effect is better, the constructed point cloud map is more accurate, the motion trajectory and the real trajectory have higher overlap, and the estimated trajectory length is closer to the real trajectory length. Compared with the absolute pose error of LeGO-LOAM algorithm, the mean error is reduced by 24.22% and the root mean error is reduced by about 31.28%. In dataset 07, the mean error is reduced by 43.30% and the root mean error is reduced by about 45.28%. Compared with the real-time Lidar range measurement and mapping algorithm, the mean error of data set 00 is reduced by 65.10% and the root mean square error is reduced by about 68.10%. In dataset 07, the mean error is reduced by 15.38% and the root mean square error is reduced by about 13.43%.

参考文献/References:

［1］ CADENA C，CARLONE L，CARRILLO H，et al. Past，present，and future of simultaneous localization and mapping：toward the robust-perception age［J］. IEEE Transactions on Robotics，2016，32（6）：1309-1332.［2］ 刘诤轩，王亮，李和平，等. 融合双目视觉和2D激光雷达的室外定位［J］. 控制与决策，2023，38（7）：1861-1868. ［3］ LIU W B，SUN W，LIU Y. DLOAM：real-time and robust LiDAR SLAM system based on CNN in dynamic urban environments［J］. IEEE Open Journal of Intelligent Transportation Systems，2021, PP(99):1.［4］ ZHANG X G，ZHANG Z H，WANG Q，et al. Using a two-stage method to reject false loop closures and improve the accuracy of collaborative SLAM systems［J］. Electronics，2021，10（21）：2638. ［5］ CUI Y G，CHEN X，ZHANG Y L，et al. BoW3D：bag of words for real-time loop closing in 3D LiDAR SLAM［J］. IEEE Robotics and Automation Letters，2023，8（5）：2828-2835.［6］ ZHANG J，SINGH S. LOAM：lidar odometry and mapping in real-time［C］∥Robotics：Science and Systems X. Robotics：Science and Systems Foundation，2014：1-9.［7］ SHAN T X，ENGLOT B. LeGOLOAM：lightweight and ground-optimized lidar odometry and mapping on variable terrain［C］∥2018 IEEE/RSJ International Conference on Intelligent Robots and Systems （IROS）. Madrid: IEEE，2018：4758-4765.［8］ PAN H M，LIU D F，REN J Z，et al. LiDARIMU tightly-coupled SLAM method based on IEKF and loop closure detection［J］. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing，2024，17：6986-7001.［9］ XIANG H D，ZHU X S，SHI W Z，et al. DeLightLCD：a deep and lightweight network for loop closure detection in LiDAR SLAM［J］. IEEE Sensors Journal，2022，22（21）：20761-20772. ［10］ SHI Y，LI R，SHI Y J，et al. A robust and lightweight loop closure detection approach for challenging environments［J］. Drones，2024，8（7）：322.［11］ ZHANG X，ZHANG H J，QIAN C，et al. A LiDAR-intensity SLAM and loop closure detection method using an intensity cylindrical-projection shape context descriptor［J］. International Journal of Applied Earth Observation and Geoinformation，2023，122：103419.［12］ CATTANEO D，VAGHI M，VALADA A. LCDNet：deep loop closure detection and point cloud registration for LiDAR SLAM［J］. IEEE Transactions on Robotics，2022，38（4）：2074-2093.［13］ CHEN Z，SUN K，YANG F，et al. SC （2）-PCR++：rethinking the generation and selection for efficient and robust point cloud registration［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence，2023，45（10）：12358-12376［14］ BESL P J，MCKAY N D. Method for registration of 3-D shapes［J］∥IEEE Transactions on Pattern Analysis and Machine Intelligence, 14(2): 239-256.［15］ ZHANG S T，WANG H T，WANG C，et al. An improved RANSAC-ICP method for registration of SLAM and UAV-LiDAR point cloud at plot scale［J］. Forests，2024，15（6）：893 ［16］ SHEN Z W，LU X H，WANG Y N，et al. SOCR：simultaneous overlap prediction and correspondence estimation for point cloud registration in real-time［J］. IEEE Robotics and Automation Letters，2024，9（9）：7955-7962.［17］ YANG L，MA H W，NIE Z，et al. 3D LiDAR point cloud registration based on IMU preintegration in coal mine roadways［J］. Sensors，2023，23（7）：3473.［18］ GEIGER A，LENZ P，STILLER C，et al. Vision meets robotics：the KITTI dataset［J］. The International Journal of Robotics Research，2013，32（11）：1231-1237.

更新日期/Last Update: 2025-12-25