• Laser & Optoelectronics Progress
  • Vol. 56, Issue 21, 211507 (2019)
Fuchun Lin, Yuhong Liu, Jinfan Zhou, Zhinan Ma, Qianqian He, Manman Wang, and Rongfen Zhang*
Author Affiliations
  • College of Big Data and Information Engineering, Guizhou University, Guiyang, Guizhou 550025, China
  • show less
    DOI: 10.3788/LOP56.211507 Cite this Article Set citation alerts
    Fuchun Lin, Yuhong Liu, Jinfan Zhou, Zhinan Ma, Qianqian He, Manman Wang, Rongfen Zhang. Optimization of Visual Odometry Algorithm Based on ORB Feature[J]. Laser & Optoelectronics Progress, 2019, 56(21): 211507 Copy Citation Text show less
    Flowchart of traditional VO method
    Fig. 1. Flowchart of traditional VO method
    Improved VO flowchart based on ORB
    Fig. 2. Improved VO flowchart based on ORB
    FAST feature points
    Fig. 3. FAST feature points
    Diagram of variation of adjacent feature point spacing
    Fig. 4. Diagram of variation of adjacent feature point spacing
    Minimum re-projection error of camera
    Fig. 5. Minimum re-projection error of camera
    ORB feature point extraction
    Fig. 6. ORB feature point extraction
    ORB feature point matching
    Fig. 7. ORB feature point matching
    Triangular networks established by Delauna triangulation. (a) Triangle network of reference frame; (b) triangle network of current frame
    Fig. 8. Triangular networks established by Delauna triangulation. (a) Triangle network of reference frame; (b) triangle network of current frame
    ORB front-end camera track
    Fig. 9. ORB front-end camera track
    ImagesequenceKITTI01KITTI05KITTI07KITTI09
    Scenescale /(m×m)1157×1827479×426191×209465×568
    Number of keyframes1101276111011591
    ORB-SLAMerror /mX8.233.367.62
    Error of proposedalgorithm /m15.667.423.146.58
    Table 1. Root mean square error of camera track
    AlgorithmRMSE /mMean /mMedian /mMin /mMax /m
    ORB-SLAM0.0872740.0821280.0756790.0212580.164835
    Proposed algorithm0.0603090.0625650.0677170.0241820.163003
    Table 2. Error table of camera pose estimation
    Algorithm stationAverage error of improved ORBAverage error of standard ORBReduction ratio of average error
    Average distanceerror /mAverageangle error /(°)Average distanceerror /mAverageangle error /(°)Reduction ratio ofdistance error /mReduction ratio ofangle error /(°)
    Scene 10.082.10.7111.888.7382.20
    Scene 20.134.20.9518.386.3277.05
    Scene 30.529.31.6620.668.6754.84
    Table 3. Camera motion errors in different scenes
    Fuchun Lin, Yuhong Liu, Jinfan Zhou, Zhinan Ma, Qianqian He, Manman Wang, Rongfen Zhang. Optimization of Visual Odometry Algorithm Based on ORB Feature[J]. Laser & Optoelectronics Progress, 2019, 56(21): 211507
    Download Citation