Acta Photonica Sinica, Volume. 50, Issue 4, 182(2021)
Research on Visual Odometry Based on Catadioptric Panoramic Camera
Figures & Tables(11)
Fig. 4. Flow chart of monocular initialization for catadioptric panoramic camera
Fig. 9. Comparison of calculated trajectory and real trajectory under each sequence of Multi-FoV data set
Table 1. Pose trajectory error statistics
View tableView in ArticleTable 1. Pose trajectory error statistics
Sequence Length/m Algorithm RMSE/m Std/m Urban canyon 482.59 Ours 0.11 0.04 ORB⁃SLAM2 0.18 0.10 ORB⁃SLAM3 0.14 0.09 DSO 0.12 0.07 Room 9.33 Ours 0.002 0.000 7 ORB⁃SLAM2 0.004 0.002 0 ORB⁃SLAM3 0.004 0.002 0 DSO 0.020 0.009 0
Table 2. Tracking result statistics
View tableView in ArticleTable 2. Tracking result statistics
Algorithm Speed(×2) Speed(×3) Speed(×4) Speed(×5) Ours √ √ √ √ ORB⁃SLAM2 √ √ √ × ORB⁃SLAM3 √ √ √ × DSO √ √ × ×
Tools
Get Citation
Copy Citation Text
Yu ZHANG, Xiping XU, Ning ZHANG, Yaowen LÜ, Yi LU. Research on Visual Odometry Based on Catadioptric Panoramic Camera[J]. Acta Photonica Sinica, 2021, 50(4): 182
Paper Information
Category: Machine Vision
Received: Oct. 27, 2020
Accepted: Dec. 28, 2020
Published Online: May. 11, 2021
The Author Email:
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20