Laser & Optoelectronics Progress, Volume. 60, Issue 4, 0410018(2023)

Adaptive Scale Target Tracking Algorithm Based on Dynamic Template Matching

Fangfang Chen and Daiping Song*
Author Affiliations
  • College of Mechanical and Vehicle Engineering, Chongqing University, Chongqing 400030
  • show less
    Figures & Tables(11)
    Flowchart of the proposed algorithm
    Comparison of template update algorithms. (a) Initial frame; (b) ground truth; (c) traditional algorithm template; (d) proposed algorithm template
    Effect diagram of Kalman filtering algorithm
    Diagram of template matching
    Coincidence rate and accuracy of each tracking algorithm. (a) Coincidence rate; (b) accuracy
    Tracking results of target tracking algorithms. (a) dog 1; (b) freemen 1; (c) david 3; (d) freeman 4; (e) fish; (f) trellis
    Schematic of experimental environment
    Tracking results of target tracking algorithms at real scene. (a) Traditional TM algorithm; (b) proposed algorithm
    • Table 1. Precision of different algorithms under various influencing factors

      View table

      Table 1. Precision of different algorithms under various influencing factors

      CircumstanceTMSSTMSCMCSKTLDStruck
      Scale variations0.5960.8880.9000.6890.6800.832
      Background clutter0.5860.7650.9190.8720.5810.925

      In/Out-plane rotation

      Illumination variations

      Occlusions

      0.539

      0.415

      0.616

      0.821

      0.742

      0.783

      0.890

      0.904

      0.832

      0.713

      0.668

      0.704

      0.737

      0.863

      0.610

      0.850

      0.958

      0.753

      Deformation0.8300.9190.9410.9030.6070.948
    • Table 2. Coincidence rate of different algorithms under various influencing factors

      View table

      Table 2. Coincidence rate of different algorithms under various influencing factors

      CircumstanceTMSSTMSCMCSKTLDStruck
      Scale variations0.4160.7460.8310.4210.5240.577
      Background clutter0.5730.6220.9430.6180.6100.902

      In/Out-plane rotation

      Illumination variations

      Occlusions

      0.402

      0.405

      0.638

      0.707

      0.570

      0.791

      0.823

      0.893

      0.816

      0.530

      0.514

      0.667

      0.603

      0.764

      0.679

      0.623

      0.880

      0.765

      Deformation0.8160.7720.9880.6320.6790.961
    • Table 3. Time consuming of different algorithms on dog 1 test set

      View table

      Table 3. Time consuming of different algorithms on dog 1 test set

      AlgorithmAverage time /msSpeed /(frame·s-1
      SSTM without Kalman67.4815
      SSTM with Kalman19.9150
      TM20.5048

      CSK

      Struck

      SCM

      22.19

      45.28

      104.60

      45

      22

      10

      TLD31.0232
    Tools

    Get Citation

    Copy Citation Text

    Fangfang Chen, Daiping Song. Adaptive Scale Target Tracking Algorithm Based on Dynamic Template Matching[J]. Laser & Optoelectronics Progress, 2023, 60(4): 0410018

    Download Citation

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

    Category: Image Processing

    Received: Dec. 7, 2021

    Accepted: Jan. 5, 2022

    Published Online: Feb. 14, 2023

    The Author Email: Song Daiping (songdp@cqu.edu.cn)

    DOI:10.3788/LOP213163

    Topics