• Laser & Optoelectronics Progress
  • Vol. 57, Issue 20, 201501 (2020)
Guangfu Xu1、2, Jichao Zeng1、2, and Xixiang Liu1、2、*
Author Affiliations
  • 1School of Instrument Science and Engineering, Southeast University, Nanjing, Jiangsu 210096, China
  • 2Key Laboratory of Micro-Inertial Instrument and Advanced Navigation Technology, Ministry of Education, Nanjing, Jiangsu 210096, China
  • show less
    DOI: 10.3788/LOP57.201501 Cite this Article Set citation alerts
    Guangfu Xu, Jichao Zeng, Xixiang Liu. Visual Odometer Based on Optical Flow Method and Feature Matching[J]. Laser & Optoelectronics Progress, 2020, 57(20): 201501 Copy Citation Text show less
    Flow chart of algorithm
    Fig. 1. Flow chart of algorithm
    Schematic of referenced-frame/current-frame VO
    Fig. 2. Schematic of referenced-frame/current-frame VO
    Schematic of optical-flow/feature-matching fusion. (a) Optical flow method; (b) feature matching introduced in optical flow method
    Fig. 3. Schematic of optical-flow/feature-matching fusion. (a) Optical flow method; (b) feature matching introduced in optical flow method
    Relationship among different frame rates
    Fig. 4. Relationship among different frame rates
    Schematic of reprojection error
    Fig. 5. Schematic of reprojection error
    Trace map of proposed algorithm
    Fig. 6. Trace map of proposed algorithm
    Estimated traces of different algorithms on EuRoC dataset. (a) V1_01_easy image; (b) V1_02_medium image
    Fig. 7. Estimated traces of different algorithms on EuRoC dataset. (a) V1_01_easy image; (b) V1_02_medium image
    Time consumption test results of three methods. (a) Feature method; (b) optical flow method; (c) proposed algorithm; (d) time average of three methods
    Fig. 8. Time consumption test results of three methods. (a) Feature method; (b) optical flow method; (c) proposed algorithm; (d) time average of three methods
    ImageProposed algorithmORBSLAM2LK
    V1_01_easy0.218419430.184329320.32365693
    V1_02_medium0.401359020.493164230.50254364
    V1_03_difficult0.4663932030.443013040.62435742
    V2_01_easy0.1748694030.170469820.36542354
    V2_02_medium0.5093123520.435916420.71354231
    V2_03_difficult0.3925698430.453306340.48636982
    Table 1. RMSEs of three algorithms unit: m
    Guangfu Xu, Jichao Zeng, Xixiang Liu. Visual Odometer Based on Optical Flow Method and Feature Matching[J]. Laser & Optoelectronics Progress, 2020, 57(20): 201501
    Download Citation