본문 바로가기
  • Home

Design of a Tracking Filter for Preventing Miss-Tracking in Moving Target Tracking

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2025, 30(2), pp.163-170
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : January 6, 2025
  • Accepted : January 27, 2025
  • Published : February 28, 2025

Jeongseok Kim 1 Dae-Yeon Kim 2

1국방과학연구소(대전)
2국방과학연구소

Accredited

ABSTRACT

This paper proposes a tracking filter design method that rapidly removes false targets (miss-detections) in image-based moving object detection and tracking while maintaining stable tracking of the true target. The method continuously monitors the moving distance and direction of the target in consecutive frames to remove tracking gates with irregular trajectories or distances. By dynamically adjusting the gate's speed and expansion range based on slant range information, it achieves robust performance under varying distance conditions. Overlapping filters are resolved by retaining only the highest-reliability filter, and a variable threshold for final track selection, set according to the initial slant range, allows for more thorough validation at greater distances and faster confirmation at shorter distances. Experimental results on real IR and visible-light data demonstrate the effectiveness of the proposed algorithm in swiftly removing false targets while reliably tracking only the true target. Comparisons with existing PDA filters confirm a reduced false alarm rate and earlier target confirmation.

Citation status

* References for papers published after 2023 are currently being built.