计算机与现代化 ›› 2023, Vol. 0 ›› Issue (02): 17-23.

• 算法设计与分析 • 上一篇    下一篇

融合视觉信息的激光SLAM

  

  1. (1.上海电力大学自动化工程学院,上海 200090; 2.上海合时智能科技有限公司,上海 200040)
  • 出版日期:2023-04-10 发布日期:2023-04-10
  • 作者简介:吴松林(1996—),男,河南商丘人,硕士研究生,研究方向:计算机视觉与激光,SLAM,E-mail: wsl0816@qq.com; 张国伟(1970—),男,上海人,副教授,博士,研究方向:信息检测,移动机器人,E-mail: 13917408956@163.com; 卢秋红(1973—),女,上海人,高级工程师,博士,研究方向:计算机视觉,移动机器人,E-mail: 15921492692@163.com; 施建壮(1995—),男,江苏连云港人,硕士研究生,研究方向:激光SLAM,路径规划,E-mail: 1260556598@qq.com; 黄威(1998—),湖北宜昌人,硕士研究生,研究方向:计算机视觉,E-mail: 837178550@qq.com。
  • 基金资助:
    上海市金山区信息化发展专项资金资助项目(2021-XXH-11);上海市闵行区产学研项目(2019MHC083)

Laser SLAM Mapping Method Based on Visual Information

  1. (1.School of Automation Engineering, Shanghai University of Electric Power, Shanghai 200090, China;
    2.Shanghai HRSTEK Co., Ltd,, Shanghai 200040, China)
  • Online:2023-04-10 Published:2023-04-10

摘要: 针对Gmapping SLAM(simultaneous location and mapping)算法在地图构建过程中对里程计定位精度要求较高且存在粒子耗散、退化等问题,本文首先设计出并行视觉识别与定位网络,用视觉特征与定位信息弥补粒子退化与激光点的漂移,强化定位能力,提高语义信息与构图精度;其次优化提议分布,将观测模型从里程计观测模型变换为激光观测模型并进行高斯采样,用更少的粒子覆盖机器人的概率分布;最后通过贝叶斯规则将视觉信息与激光信息融合,利用仿真工具、机器人平台与原算法进行对比,实验结果表明该算法不仅有效地提高地图构建的精确度与鲁棒性而且丰富了地图的语义信息。

关键词: SLAM, 计算机视觉, 激光观测模型, 神经网络, 贝叶斯融合

Abstract: In view of the fact that the Gmapping SLAM algorithm has high requirements on the accuracy of odometry positioning information in the process of map construction, and there are problems such as particle dissipation and degradation, Firstly, a parallel visual recognition and localization network is designed to strengthen the localization ability and improve the semantic information and composition accuracy; Secondly, the optimization proposal distribution is improved, we use the laser observation model to replace the odometrg motion model and perform Gaussian sampling to cover the probability distribution of the robot with fewer particles; Finally, the visual information and laser information are fused by Bayesian rule, and the original algorithm is compared. The results show that the algorithm improves the accuracy and robustness of map construction and enriches the semantic information.

Key words: SLAM, computer vision, laser observation model, neural networks, bayesian fusion