Laser & Optoelectronics Progress, Volume. 59, Issue 22, 2215006(2022)

An Overall Matching Algorithm for Image Feature Points in Visual Navigation

Gaojie Wang1、*, Xiangyang Hao1, and Shufeng Miao2
Author Affiliations
  • 1Institute of Geospatial Information, Information Engineering University, Zhengzhou 450001, Henan, China
  • 2Wuhan Kedao Geographical Information Engineering Co., Ltd., Wuhan 430081, Hubei, China
  • show less
    Figures & Tables(11)
    Image sampling diagram
    Fitting results of detection threshold and feature coefficient
    Method of image area segmentation. (a) Target image;(b) image to be matched
    Diagram of point selection. (a) Target image; (b) image to be matched
    Epipolar constraint and position constraint
    Effect of parameter values on matching results
    Results of feature point extraction by proposed method
    Results of feature point matching by proposed method
    Comparison of average running time for image feature point matching in different data sets. (a) TUM data set; (b) UAV data set;(c) images captured by mobile robots
    • Table 1. Results of different methods for feature point extraction

      View table

      Table 1. Results of different methods for feature point extraction

      Image numberProposed methodFASTORB
      Number

      Repeat

      points

      Repetition

      rate /%

      Number

      Repeat

      points

      Repetition

      rate /%

      Number

      Repeat

      points

      Repetition

      rate /%

      136/342471160/158895636/342265
      235/402571195/2071196135/402366
      338/442566156/165925938/442668
      438/472668192/1811196638/472874
      531/331961156/174865531/332168
      628/311864159/157935928/311864
    • Table 2. Results of different methods for image feature point matching

      View table

      Table 2. Results of different methods for image feature point matching

      Data setImage numberProposed methodFASTORB

      Matching

      number

      Accuracy

      rate /%

      Matching

      number

      Accuracy

      rate /%

      Matching

      number

      Accuracy

      rate /%

      360_kidnap383

      6160/5895

      6882/6764

      6762/6513

      4788/4433

      8687/8635

      9480/9363

      10863/10713

      14340/14196

      5858/5702

      6661/6395

      7383/7116

      95.70

      51418/39895

      66186/49594

      57776/42204

      42681/31089

      40013/32419

      49022/37483

      57008/43434

      90032/73818

      37371/29300

      34124/28268

      37970/31397

      77.59

      38460/35996

      39940/37864

      38720/35125

      29824/26931

      45027/43065

      50182/47815

      47649/46483

      56979/52679

      39265/33434

      38719/36392

      44284/42175

      93.59
      freiburg2_desk40898.2974.9394.80
      large_no_loop39996.3273.0590.72
      pioneer_slam31192.5972.8490.30
      building138099.4081.0295.64
      building245598.7776.4695.28
      building345098.6276.1997.55
      building448799.0081.9992.45
      test134997.3478.4085.15
      test237196.0182.8493.99
      test340296.3882.6995.24
    Tools

    Get Citation

    Copy Citation Text

    Gaojie Wang, Xiangyang Hao, Shufeng Miao. An Overall Matching Algorithm for Image Feature Points in Visual Navigation[J]. Laser & Optoelectronics Progress, 2022, 59(22): 2215006

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Machine Vision

    Received: Jul. 2, 2021

    Accepted: Oct. 19, 2021

    Published Online: Oct. 12, 2022

    The Author Email: Gaojie Wang (gaojiewang2014@163.com)

    DOI:10.3788/LOP202259.2215006

    Topics