표적이 일시적으로 가려지는 환경에서 ITS 기법을 이용한 영상 표적 추적 알고리듬 연구

다국어 초록 (Multilingual Abstract)

Automatic tracking in cluttered environment requires the initiation and maintenance of tracks, and track existence probability of true track is kept by Markov Chain Two model of target existence propagation. Unlike Markov Chain One model for target existence propagation, Markov Chain Two model is made up three hypotheses about target existence event which are that the target exist and is detectable, the target exists and is non-detectable through occlusion, and the target does not exist and is non-detectable according to non-existing target. In this paper we present multi-scan single target tracking algorithm based on the target existence, which call the Integrated Track Splitting algorithm with Markov Chain Two model in imaging sensor.

목차 (Table of Contents)

• Abstract
• Ⅰ. 서론
• Ⅱ. 다중스캔 방식 표적 추적, 탐지 기법
• Ⅲ. 시뮬레이션
• Ⅳ. 결론
• 참고문헌

참고문헌 (Reference)

1 T. L. Song, "Tracking through occlusions and track segmentation reduction" 49 (49): 623-631, 2013

2 Yaakov Bar-Shalom, "Tracking and data association" Academic Press Professional, Inc. 1987

3 X. R. Li, "Theoritical analysis and performance prediction of tracking in clutter with strongest neighbor filters" 2758-2763, 1995

4 D. Musicki, "Stankovic, "Integrated probabilistic data association(IPDA), "Prec" 1992

5 T. Kirubarajan, "Probabilistic data association techniques for target tracking in clutter" 32 (32): 536-557, 2004

6 D. Musicki, "Integrated track splitting filter-efficient multi-scan single target tracking in clutter" 43 (43): 2007

7 B. M. Kim, "High speed technique for ensemble classifier using Fern" 1225-1228, 2012

8 T. L. Song, "A probabilistic strongest neighbor filter algorithm for m validated measurements"

9 T. L. Song, "A probabilistic nearest neighbor filter algorithm for m validated measurements" 54 (54): 2006