Space target detection algorithm for star images based on multi-feature Hungarian matching

NIU Haipeng; HUANG Yawei

doi:10.37188/CJLCD.2026-0099

您当前的位置：

首页 >

文章列表页 >

Space target detection algorithm for star images based on multi-feature Hungarian matching

Image Processing | 更新时间：2026-09-12

- Space target detection algorithm for star images based on multi-feature Hungarian matching
- Chinese Journal of Liquid Crystals and Displays Vol. 41, Issue 7, Pages: 1053-1071(2026)
- 作者机构：
  
  中国科学院长春光学精密机械与物理研究所，吉林长春 130033
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(62505316)
- DOI：10.37188/CJLCD.2026-0099
  CLC： TP751.1;TP391.4
- CSTR：32172.14.CJLCD.2026-0099
- Received：02 June 2026，
  
  Revised：2026-06-13，
  
  Published：06 July 2026
- 稿件说明：
移动端阅览
NIU Haipeng, HUANG Yawei. Space target detection algorithm for star images based on multi-feature Hungarian matching[J]. Chinese Journal of Liquid Crystals and Displays, 2026, 41(7): 1053-1071.
DOI：

NIU Haipeng, HUANG Yawei. Space target detection algorithm for star images based on multi-feature Hungarian matching[J]. Chinese Journal of Liquid Crystals and Displays, 2026, 41(7): 1053-1071. DOI： 10.37188/CJLCD.2026-0099. CSTR： 32172.14.CJLCD.2026-0099.

摘要

传统匈牙利算法在星图空间目标检测中存在4方面局限性：缺乏运动预测、全局搜索冗余、匹配特征单一和固定权重缺乏适应性。在高低速目标共存、目标密集逼近、尺寸动态变化及低信噪比等复杂场景下，上述不足导致检测精度和身份一致性显著下降。本文提出一种基于多特征递进改进的匈牙利匹配算法，首先引入卡尔曼滤波器对目标位置和速度进行预测，以预测位置替代静态观测位置构造匹配代价；其次利用卡尔曼预测协方差矩阵构建马氏距离门控，将匈牙利匹配的搜索空间从全局约束至预测椭球内的局部区域；然后设计多特征动态权重机制，根据位置、面积和长宽比的实时可靠性自适应调整各特征的代价贡献；最后引入异常惩罚机制，对面积和长宽比异常突变的候选匹配施加额外代价。在5种代表性仿真星图场景（多速度目标、目标密集逼近、多机制混合、高密度目标和低信噪比）下的实验结果表明，本文算法平均MOTA为88.3%，超出传统匈牙利算法（82.4%）5.9个百分点、PPMHT算法（64.6%）23.7个百分点和最近邻算法（82.3%）6.0个百分点；平均IDF1为65.8%，超出传统匈牙利算法（56.3%）9.5个百分点、PPMHT算法（50.0%）15.8个百分点；平均IDS为1.6次，低于传统匈牙利算法（3.9次）2.3次、低于最近邻算法（5.2次）3.6次；算法平均耗时为41.2 ms/帧，满足实时性要求。消融实验表明，卡尔曼预测、马氏门控、动态权重和异常惩罚4项改进分别贡献MOTA增益8.8、6.5、3.2和1.2个百分点。本文算法通过4个维度递进改进框架，在检测精度、身份一致性和计算效率之间取得了有效平衡，为星图空间目标的可靠帧间关联提供了一种有效方案。

Abstract

Four major limitations exist when the conventional Hungarian algorithm is applied to space target detection in star images， including the lack of motion prediction， redundant global searching， single matching feature， and poor adaptability of fixed weights. In complex scenarios with coexisting high-speed and low-speed targets， densely distributed and closely approaching targets， dynamically varying target sizes， and low signal-to-noise ratios， the above drawbacks significantly degrade tracking accuracy and identity consistency. An improved Hungarian matching algorithm based on progressive multi-feature optimization is proposed. Firstly， a Kalman filter is introduced to predict the position and velocity of space targets， and predicted coordinates are adopted to construct matching costs instead of static observations. Secondly， a Mahalanobis distance gating mechanism is established using the covariance matrix derived from Kalman prediction， which constrains the matching search range from the global image to the local region within the predicted ellipsoid and reduces computational redundancy. Thirdly， a dynamic multi-feature weighting strategy is designed to adaptively adjust the cost contribution of position， area and aspect ratio according to their real-time reliability. Furthermore， an anomaly penalty term is added to impose extra costs on matching pairs with abrupt changes in area or aspect ratio to suppress false associations. Comparative experiments are carried out on five types of simulated star image scenes. The experimental results show that the average MOTA of the proposed algorithm reaches 88.3%， which is 5.9 percentage points higher than the conventional Hungarian algorithm， and 23.7 and 6.0 percentage points higher than PPMHT and the nearest neighbor method， respectively. Its average IDF1 is 65.8%， 9.5 percentage points higher than that of the conventional Hungarian algorithm. The average IDS is only 1.6， which are 2.3 and 3.6 times lower than the conventional Hungarian algorithm and the nearest neighbor method， respectively. The average processing time per frame is 41.2 ms， which meets the real-time operation requirements. Ablation experiments verify that the Kalman prediction module， Mahalanobis distance gating， dynamic weighting and anomaly penalty bring MOTA gains of 8.8， 6.5， 3.2 and 1.2 percentage points， respectively. The proposed four-module progressive optimization framework achieves an effective balance among tracking accuracy， identity consistency and computational efficiency， and provides a reliable solution for inter-frame association of space targets in star images.

关键词

Keywords

references

TURYSHEV S G . Orbital debris in Earth orbit： operations， stability， control， and market formation ［J/OL］. arXiv ， 2026 ： 2603 . 23552 . doi: 10.1016/j.actaastro.2026.05.026 http://dx.doi.org/10.1016/j.actaastro.2026.05.026

BAKER-MCEVILLY B ， BHADAURIA S ， CANALES D ， et al . A comprehensive review on Cislunar expansion and space domain awareness ［J］. Progress in Aerospace Sciences ， 2024 ， 147 ： 101019 . doi: 10.1016/j.paerosci.2024.101019 http://dx.doi.org/10.1016/j.paerosci.2024.101019

牛海鹏，颜昌翔，王一霖，等 . 基于快速局部对比度和目标特征的星图弱小目标检测算法［J］. 液晶与显示， 2024 ， 39 （ 1 ）： 69 - 78 ． doi: 10.37188/cjlcd.2023-0060 http://dx.doi.org/10.37188/cjlcd.2023-0060

NIU H P ， YAN C X ， WANG Y L ， et al . Star map spatial target detection method based on fast local contrast and target features ［J］. Chinese Journal of Liquid Crystals and Displays ， 2024 ， 39 （ 1 ）： 69 - 78 . （in Chinese） . doi: 10.37188/cjlcd.2023-0060 http://dx.doi.org/10.37188/cjlcd.2023-0060

LAN X Q ， XIN B Q ， WANG B S ， et al . Space object detection using multi-frame temporal trajectory completion method ［C］// Proceedings of the 5th International Symposium on Artificial Intelligence and Big Data . Guiyang ： IEEE ， 2025 . doi: 10.1109/aibdf67964.2025.11440810 http://dx.doi.org/10.1109/aibdf67964.2025.11440810

EMAMI P ， PARDALOS P M ， ELEFTERIADOU L ， et al . Machine learning methods for data association in multi-object tracking ［J］. ACM Computing Surveys （CSUR）， 2021 ， 53 （ 4 ）： 69 . doi: 10.1145/3394659 http://dx.doi.org/10.1145/3394659

SU Y Q ， CHEN X ， LIU G R ， et al . Implementation of real-time space target detection and tracking algorithm for space-based surveillance ［J］. Remote Sensing ， 2023 ， 15 （ 12 ）： 3156 . doi: 10.3390/rs15123156 http://dx.doi.org/10.3390/rs15123156

ZHU R ， FU Q ， WEN G Y ， et al . An improved multi-target tracking method for space-based optoelectronic systems ［J］. Remote Sensing ， 2024 ， 16 （ 15 ）： 2847 . doi: 10.3390/rs16152847 http://dx.doi.org/10.3390/rs16152847

ZHU R ， LENG J S ， FU Q ， et al . Transformer-based target tracking algorithm for space-based optoelectronic detection ［J］. Frontiers in Physics ， 2023 ， 11 ： 1266927 . doi: 10.3389/fphy.2023.1266927 http://dx.doi.org/10.3389/fphy.2023.1266927

ZHANG Y F ， SUN P Z ， JIANG Y ， et al . ByteTrack： multi-object tracking by associating every detection box ［C］// Proceedings of the 17th European Conference on Computer Vision . Tel Aviv ： Springer ， 2022 ： 1 - 21 . doi: 10.1007/978-3-031-20047-2_1 http://dx.doi.org/10.1007/978-3-031-20047-2_1

FORTMANN T E ， BAR-SHALOM Y ， SCHEFFE M . Sonar tracking of multiple targets using joint probabilistic data association ［J］. IEEE Journal of Oceanic Engineering ， 1983 ， 8 （ 3 ）： 173 - 184 . doi: 10.1109/joe.1983.1145560 http://dx.doi.org/10.1109/joe.1983.1145560

张薇，席红霞，宋崇金 . 基于位置预测的多级假设检验空间目标检测方法［J］. 制造业自动化， 2023 ， 45 （ 8 ）： 199 - 203 ．

ZHANG W ， XI H X ， SONG C J . Multistage hypothesis testing method based on position prediction for spatial target detection ［J］. Manufacturing Automation ， 2023 ， 45 （ 8 ）： 199 - 203 . （in Chinese）

GAETJENS H X ， DAVEY S J ， ARULAMPALAM S ， et al . Histogram-PMHT for fluctuating target models ［J］. IET Radar ， Sonar & Navigation， 2017 ， 11 （ 8 ）： 1292 - 1301 . doi: 10.1049/iet-rsn.2016.0618 http://dx.doi.org/10.1049/iet-rsn.2016.0618

张奕群，尹立凡，王硕，等 . 一种改进的直方图概率多假设多目标跟踪方法［J］. 航空学报， 2021 ， 42 （ 11 ）： 524851 ．

ZHANG Y Q ， YIN L F ， WANG S ， et al . An improved histogram PMHT multi-target tracking method ［J］. Acta Aeronautica et Astronautica Sinica ， 2021 ， 42 （ 11 ）： 524851 . （in Chinese）

KUHN H W . The Hungarian method for the assignment problem ［J］. Naval Research Logistics Quarterly ， 1955 ， 2 （ 1/2 ）： 83 - 97 . doi: 10.1002/nav.3800020109 http://dx.doi.org/10.1002/nav.3800020109

BEWLEY A ， GE Z Y ， OTT L ， et al . Simple online and realtime tracking ［C］// Proceedings of 2016 IEEE International Conference on Image Processing . Phoenix ： IEEE ， 2016 ： 3464 - 3468 . doi: 10.1109/icip.2016.7533003 http://dx.doi.org/10.1109/icip.2016.7533003

WOJKE N ， BEWLEY A ， PAULUS D . Simple online and realtime tracking with a deep association metric ［C］// Proceedings of 2017 IEEE International Conference on Image Processing . Beijing ： IEEE ， 2017 ： 3645 - 3649 . doi: 10.1109/icip.2017.8296962 http://dx.doi.org/10.1109/icip.2017.8296962

KALMAN R E . A new approach to linear filtering and prediction problems ［J］. Journal of Basic Engineering ， 1960 ， 82 （ 1 ）： 35 - 45 . doi: 10.1115/1.3662552 http://dx.doi.org/10.1115/1.3662552

CHEN L ， AI H Z ， ZHUANG Z J ， et al . Real-time multiple people tracking with deeply learned candidate selection and person re-identification ［C］// Proceedings of 2018 IEEE International Conference on Multimedia and Expo . San Diego ： IEEE ， 2018 ： 1 - 6 . doi: 10.1109/icme.2018.8486597 http://dx.doi.org/10.1109/icme.2018.8486597

MUNKRES J . Algorithms for the assignment and transportation problems ［J］. Journal of the Society for Industrial and Applied Mathematics ， 1957 ， 5 （ 1 ）： 32 - 38 . doi: 10.1137/0105003 http://dx.doi.org/10.1137/0105003

LIU M ， NIU J W ， LIU Y . UKF-MOT： an unscented Kalman filter-based 3D multi-object tracker ［J］. CAAI Transactions on Intelligence Technology ， 2024 ， 9 （ 4 ）： 1031 - 1041 . doi: 10.1049/cit2.12315 http://dx.doi.org/10.1049/cit2.12315

YI K F ， LUO K ， LUO X L ， et al . UCMCTrack： multi-object tracking with uniform camera motion compensation ［C］// Proceedings of the 38th AAAI Conference on Artificial Intelligence . Vancouver ： AAAI ， 2024 ： 6702 - 6710 . doi: 10.1609/aaai.v38i7.28493 http://dx.doi.org/10.1609/aaai.v38i7.28493

奚晓梁，周晓东 . 基于小区域滤波的快速星图目标检测方法［J］. 光电工程， 2010 ， 37 （ 3 ）： 35 - 40 ．

XI X L ， ZHOU X D . Fast star map target detection method based on small region filtering ［J］. Opto-Electronic Engineering ， 2010 ， 37 （ 3 ）： 35 - 40 . （in Chinese）

刘颖，孙海江，赵勇先 . 基于注意力机制的复杂背景下红外弱小目标检测方法研究［J］. 液晶与显示， 2023 ， 38 （ 11 ）： 1455 - 1467 ． doi: 10.37188/cjlcd.2023-0030 http://dx.doi.org/10.37188/cjlcd.2023-0030

LIU Y ， SUN H J ， ZHAO Y X . Infrared dim-small target detection under complex background based on attention mechanism ［J］. Chinese Journal of Liquid Crystals and Displays ， 2023 ， 38 （ 11 ）： 1455 - 1467 . （in Chinese） . doi: 10.37188/cjlcd.2023-0030 http://dx.doi.org/10.37188/cjlcd.2023-0030

LI X ， WANG K J ， WANG W ， et al . A multiple object tracking method using Kalman filter ［C］// Proceedings of 2010 IEEE International Conference on Information and Automation . Harbin ： IEEE ， 2010 ： 1862 - 1866 . doi: 10.1109/icinfa.2010.5512258 http://dx.doi.org/10.1109/icinfa.2010.5512258

桑海瑞，郑然，程会艳，等 . 基于多特征融合的空间多目标在轨检测方法［J］. 空间控制技术与应用， 2025 ， 51 （ 1 ）： 105 - 114 ．

SANG H R ， ZHENG R ， CHENG H Y ， et al . A method of space multi-targets on-orbit detection based on multi-feature fusion ［J］. Aerospace Control and Application ， 2025 ， 51 （ 1 ）： 105 - 114 . （in Chinese）

HOSSAIN M A ， HOSSAIN M I ， HOSSAIN M D ， et al . DFC-D： a dynamic weight-based multiple features combination for real-time moving object detection ［J］. Multimedia Tools and Applications ， 2022 ， 81 （ 22 ）： 32549 - 32580 . doi: 10.1007/s11042-022-12446-6 http://dx.doi.org/10.1007/s11042-022-12446-6

BERNARDIN K ， STIEFELHAGEN R . Evaluating multiple object tracking performance： the CLEAR MOT metrics ［J］. EURASIP Journal on Image and Video Processing ， 2008 ， 2008 （ 1 ）： 246309 . doi: 10.1155/2008/246309 http://dx.doi.org/10.1155/2008/246309

刘时，朱明 . 基于Swin-Transformer改进的目标跟踪算法［J］. 液晶与显示， 2024 ， 39 （ 11 ）： 1569 - 1580 ． doi: 10.37188/cjlcd.2024-0118 http://dx.doi.org/10.37188/cjlcd.2024-0118

LIU S ， ZHU M . Improved target tracking algorithm based on Swin-Transformer ［J］. Chinese Journal of Liquid Crystals and Displays ， 2024 ， 39 （ 11 ）： 1569 - 1580 . （in Chinese） . doi: 10.37188/cjlcd.2024-0118 http://dx.doi.org/10.37188/cjlcd.2024-0118

Views

156

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

UAV aerial target tracking with attention mechanism and anti-occlusion strategies

Hand real-time tracking method based on neural network and Kalman filter

Parameter correction of unmanned aerial vehicle based on beidou/inertial navigation and multi-sensor fusion

Software design and implement of flight track radar modeling system

Driving response and two-stage waveform control of quantum-dot electrowetting display devices

Related Author

HUANG Yawei

SUN He

ZHENG Binxi

ZHAO Meng

MA Mingyang

SONG Ce

BAI Jincheng

Gong-ren ZENG

Related Institution

Shenyang Area Office of the Naval Equipment Department， People’s Liberation Army Navy

National & Local United Engineering Laboratory of Flat Panel Display Technology, School of Physics and Information Engineering, Fuzhou University

Jinjiang RichSense Electronic Technology Co., Ltd., Jinjiang

School of Electronic and Information Engineering, Suzhou University of Science and Technology

Changchun Institute of Optics, Fine Mehcanics and Physics, Chinese Academy of Sciences

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China 130033 Postal code：130033
Tel：0431-86176059 Email：yjxs@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰