An Improved VSLAM Algorithm Based on Adaptive Feature Map

被引：0
|
作者：
Zhang J.-N. [1 ]
Su Q.-X. [1 ,2 ]
Liu P.-Y. [1 ]
Zhu Q. [3 ]
Zhang K. [4 ]
机构：
[1] Department of Missile Engineering, Army Engineering University, Shijiazhuang
[2] Army Command College, Nanjing
[3] Test Brigade of Special War and Soldier System, China Baicheng Ordnance Test Center, Baicheng
[4] Key Laboratory of Guided Weapons Test and Evaluation Simulation Technology, China Huayin Ordnance Test Center, Huayin
来源：
Zidonghua Xuebao/Acta Automatica Sinica | 2019年 / 45卷 / 03期
基金：
中国国家自然科学基金;
关键词：
Corner response; General graph optimization (g2o); Map extension; Regional feature supplement; Visual odometry; Visual simultaneous localization and mapping (VSLAM);
D O I：
10.16383/j.aas.c170608
中图分类号：
学科分类号：
摘要：
An improved visual simultaneous localization and mapping (VSLAM) algorithm based on the adaptive feature map is proposed in order to enhance real-time performance. The feature map is divided into sub-regions and structural units are employed to reduce computation cost. After that the most effective feature points, sorted by corner response intensity, are extracted and matched with the current frame. In the case that the adaptive map features are not enough for registration, a method of adding more region feature point supplements and extending the feature map is also proposed, which enables re-matching ability for the visual odometry system. A frame-to-frame and frame-to-map graph optimization method is also implemented to effectively update the internal and external points in the feature map. The results in public dataset show that the location accuracy error of the proposed method is centimeter and that the point cloud map is clear and has less drift. Compared with the original one, the proposed method has better real-time performance, positioning accuracy and the ability to build maps. Copyright © 2019 Acta Automatica Sinica. All rights reserved.
引用
收藏
页码：553 / 565
页数：12
相关论文
共 24 条
  • [1] Garcia-Fidalgo E., Ortiz A., Vision-based topological mapping and localization methods: a survey, Robotics and Autonomous Systems, 64, pp. 1-20, (2015)
  • [2] Carlone L., Tron R., Daniilidis K., Dellaert F., Initialization techniques for 3D SLAM: a survey on rotation estimation and its use in pose graph optimization, Proceedings of the 2015 IEEE International Conference on Robotics and Automation, pp. 4597-4604, (2015)
  • [3] Merriaux P., Dupuis Y., Vasseur P., Savatier X., Wheel odometry-based car localization and tracking on vectorial map, Proceedings of the 17th International Conference on Intelligent Transportation Systems, pp. 1890-1891, (2014)
  • [4] Shen J.L., Tick D., Gans N., Localization through fusion of discrete and continuous epipolar geometry with wheel and IMU odometry, Proceedings of the 2011 American Control Conference, pp. 1292-1298, (2011)
  • [5] Ohno K., Tsubouchi T., Shigematsu B., Yuta S., Differential GPS and odometry-based outdoor navigation of a mobile robot, Advanced Robotics, 18, 6, pp. 611-635, (2004)
  • [6] Fuentes-Pacheco J., Ruiz-Ascencio J., Rendon-Mancha J.M., Visual simultaneous localization and mapping: a survey, Artificial Intelligence Review, 43, 1, pp. 55-81, (2015)
  • [7] Endres F., Hess J., Sturm J., Cremers D., Burgard W., 3-D mapping with an RGB-D camera, IEEE Transactions on Robotics, 30, 1, pp. 177-187, (2014)
  • [8] Scaramuzza D., Fraundorfer F., Visual odometry, IEEE Robotics and Automation Magazine, 18, 4, pp. 80-92, (2011)
  • [9] Nister D., Naroditsky O., Bergen J., Visual odometry, Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. I-652-I-659, (2004)
  • [10] Kerl C., Sturm J., Cremers D., Robust odometry estimation for RGB-D cameras, Proceedings of the 2013 IEEE International Conference on Robotics and Automation, pp. 3748-3754, (2013)
123