长短时算法:一种视频多目标关联算法

TP391.41; 视频多目标跟踪面临的主要挑战是严重遮挡带来的身份切换问题.解决身份切换的技术为视频目标关联,即识别不同帧中的同一目标并分配身份编号.针对身份切换提出长短时视频目标关联算法,该算法在短时即相邻帧之间使用运动特征进行目标匹配,在长时即不相邻帧之间增加外观特征进行目标匹配,实现对于遮挡后被检出目标的再匹配.此外对卡尔曼滤波进行改进,增加框宽参数,使得预测框更准确;还使用平均外观特征并增加检测置信度作为更新参数,使得外观特征鲁棒性更强,在复杂环境下仍能表现良好.长短时算法在数据集MOT17 上多目标跟踪精度得分 81.3,身份F1 值得分 81.3,实现了严重遮挡场景下的稳定跟踪....

Full description

Saved in:
Bibliographic Details
Published in指挥控制与仿真 Vol. 46; no. 3; pp. 116 - 122
Main Authors 王锐, 丁春山
Format Journal Article
LanguageChinese
Published 江苏自动化研究所,江苏 连云港 222061 2024
Subjects
multi-object tracking
tracking-by-detection
视频多目标跟踪
average appearance feature
先检测后跟踪
平均外观特征
视频目标关联
video object association
Online AccessGet full text
ISSN1673-3819
DOI10.3969/j.issn.1673-3819.2024.03.017

Cover

Abstract TP391.41; 视频多目标跟踪面临的主要挑战是严重遮挡带来的身份切换问题.解决身份切换的技术为视频目标关联,即识别不同帧中的同一目标并分配身份编号.针对身份切换提出长短时视频目标关联算法,该算法在短时即相邻帧之间使用运动特征进行目标匹配,在长时即不相邻帧之间增加外观特征进行目标匹配,实现对于遮挡后被检出目标的再匹配.此外对卡尔曼滤波进行改进,增加框宽参数,使得预测框更准确;还使用平均外观特征并增加检测置信度作为更新参数,使得外观特征鲁棒性更强,在复杂环境下仍能表现良好.长短时算法在数据集MOT17 上多目标跟踪精度得分 81.3,身份F1 值得分 81.3,实现了严重遮挡场景下的稳定跟踪.
AbstractList TP391.41; 视频多目标跟踪面临的主要挑战是严重遮挡带来的身份切换问题.解决身份切换的技术为视频目标关联,即识别不同帧中的同一目标并分配身份编号.针对身份切换提出长短时视频目标关联算法,该算法在短时即相邻帧之间使用运动特征进行目标匹配,在长时即不相邻帧之间增加外观特征进行目标匹配,实现对于遮挡后被检出目标的再匹配.此外对卡尔曼滤波进行改进,增加框宽参数,使得预测框更准确;还使用平均外观特征并增加检测置信度作为更新参数,使得外观特征鲁棒性更强,在复杂环境下仍能表现良好.长短时算法在数据集MOT17 上多目标跟踪精度得分 81.3,身份F1 值得分 81.3,实现了严重遮挡场景下的稳定跟踪.
Abstract_FL The main challenge of multi-object tracking(MOT)is identity switch caused by severe occlusion.The solution to identity switching is video object association,which assigns an identity number to the same target in different frames.In this paper,a long-short time association algorithm is proposed for identity switching.In the short-time,that is,the motion fea-tures between adjacent frames are used to match,and in the long-time,that is,the non-adjacent frames are directly added to the appearance features for association to rematch the object detected after occlusion.Besides,the Kalman filter is improved and the frame width parameter is added to make the predicted frame more accurate;appearance features use average appear-ance features and increase detection confidence as update parameters to make appearance more robust and can still work in complex scenes.The new tracker,LSATrack,achieves 81.3MOTA and 81.3IDF1 in the MOT17 and achieves stable tracking in severe occlusion scenarios.
Author 王锐
丁春山
AuthorAffiliation 江苏自动化研究所,江苏 连云港 222061
AuthorAffiliation_xml – name: 江苏自动化研究所,江苏 连云港 222061
Author_FL DING Chunshan
WANG Rui
Author_FL_xml – sequence: 1
  fullname: WANG Rui
– sequence: 2
  fullname: DING Chunshan
Author_xml – sequence: 1
  fullname: 王锐
– sequence: 2
  fullname: 丁春山
BookMark eNrjYmDJy89LZWBQMTTQM7Y0s9TP0sssLs7TMzQzN9Y1tjC01DMyMDLRMzDWMzA0Z2HghItzMPAWF2cmGRibmRibGRuYcDIYvpy6__n8tc-mb3u-bvqzzVOtnuxoeL6898XytpeLJj5dMuv57HXPFrQ_bd38onEKRAUPA2taYk5xKi-U5mbQc3MNcfbQLU_MS0vMS4_Pyi8tygPKxBcmVWVkV1WUVKZVZRWDXGRgDHSPMckaAG3UV5g
ClassificationCodes TP391.41
ContentType Journal Article
Copyright Copyright © Wanfang Data Co. Ltd. All Rights Reserved.
Copyright_xml – notice: Copyright © Wanfang Data Co. Ltd. All Rights Reserved.
DBID 2B.
4A8
92I
93N
PSX
TCJ
DOI 10.3969/j.issn.1673-3819.2024.03.017
DatabaseName Wanfang Data Journals - Hong Kong
WANFANG Data Centre
Wanfang Data Journals
万方数据期刊 - 香港版
China Online Journals (COJ)
China Online Journals (COJ)
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
DocumentTitle_FL Long-short time association algorithm:a robust data association algorithm
EndPage 122
ExternalDocumentID qbzhkzxtyfzjs202403017
GroupedDBID -0Y
2B.
4A8
92I
93N
ALMA_UNASSIGNED_HOLDINGS
CCEZO
CUBFJ
GROUPED_DOAJ
PSX
TCJ
ID FETCH-wanfang_journals_qbzhkzxtyfzjs2024030173
ISSN 1673-3819
IngestDate Thu May 29 04:10:19 EDT 2025
IsPeerReviewed false
IsScholarly true
Issue 3
Keywords multi-object tracking
tracking-by-detection
视频多目标跟踪
average appearance feature
先检测后跟踪
平均外观特征
视频目标关联
video object association
Language Chinese
LinkModel OpenURL
MergedId FETCHMERGED-wanfang_journals_qbzhkzxtyfzjs2024030173
ParticipantIDs wanfang_journals_qbzhkzxtyfzjs202403017
PublicationCentury 2000
PublicationDate 2024
PublicationDateYYYYMMDD 2024-01-01
PublicationDate_xml – year: 2024
  text: 2024
PublicationDecade 2020
PublicationTitle 指挥控制与仿真
PublicationTitle_FL Command Control & Simulation
PublicationYear 2024
Publisher 江苏自动化研究所,江苏 连云港 222061
Publisher_xml – name: 江苏自动化研究所,江苏 连云港 222061
SSID ssib036436304
ssib051372449
ssib051327542
ssib023646202
ssib001102748
ssj0002912699
ssib000968374
Score 4.679006
Snippet TP391.41;...
SourceID wanfang
SourceType Aggregation Database
StartPage 116
Title 长短时算法:一种视频多目标关联算法
URI https://d.wanfangdata.com.cn/periodical/qbzhkzxtyfzjs202403017
Volume 46
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnR1NaxNBdGhTEC9FUVGrpYcOgiVxZ2Z3PrzNphuKoKcKvZVskrUqRLQpaE56UE8eBEtBD4oIPdlLQbz5Z2zS_gvfm9lslrZIFMLy9s3M-8zOvPfYmSVkURouRVM1q6mB3CTUvF1tCpFVRTsI27yZqSzD3ch378mV--GdtWhtavpm6a2lrV5aa_VP3VfyP14FHPgVd8n-g2cLooAAGPwLV_AwXCfyMU0MNRGNGzRR1DSoXaaJpEbRWCLGJggDJhbYTViahDTW-HIDtiqqob92gERSllPDaBJRG1JjHc0YiQAFG1CtsElHSA1GaUZNeJJLOdhFpK67gQ6wkQMS5IiktJPTi5Q4IB7pUgd1Rv8FREC7jp26IFowbvGDmVNbOwYgH4NfuZrBx3VMJydDW6EKMdUeAC2sE8lSqx0AAngbBiM1jZMW5DdoQF4_SWoJIVDAeGWss6ZECW1jCSKkwB8In68BUokqJrLlRSKvkz4s1xDcjM-YLAUPzG-yPr4uCSONW5eQQa1gUEMLuDN2_f7VYyd_P037G4_7z3svsv6jTeyK6auaJjNcKRZVyIyNl-NGORnVQpUP98OSQzG74pcCJB8f9ga3QopxMh4xwYvPIft7BbGfKSqW3DAu3QdZCwXOkMVcu1t_081tiOtmze6DUuy2eo7M5knXgvVP0Hky1d-4QNjR9q_h5--DnR_DvZ3B_vbt3z9fDnffHe6-Ofr6_uDbx-GnvcGXtwev9w9fffA9LpJaI1mtr1RzNuv5Q7q5froNxSVS6T7pdi6ThUjBnNQO2mnYMSFYUBvd4llLi6zDNYvSK-TGhESvTtxzjpxF2FfyrpFK79lW5zrEtr10PnfqvKsN_QE0GX0r
linkProvider EBSCOhost
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=%E9%95%BF%E7%9F%AD%E6%97%B6%E7%AE%97%E6%B3%95%3A%E4%B8%80%E7%A7%8D%E8%A7%86%E9%A2%91%E5%A4%9A%E7%9B%AE%E6%A0%87%E5%85%B3%E8%81%94%E7%AE%97%E6%B3%95&rft.jtitle=%E6%8C%87%E6%8C%A5%E6%8E%A7%E5%88%B6%E4%B8%8E%E4%BB%BF%E7%9C%9F&rft.au=%E7%8E%8B%E9%94%90&rft.au=%E4%B8%81%E6%98%A5%E5%B1%B1&rft.date=2024&rft.pub=%E6%B1%9F%E8%8B%8F%E8%87%AA%E5%8A%A8%E5%8C%96%E7%A0%94%E7%A9%B6%E6%89%80%2C%E6%B1%9F%E8%8B%8F+%E8%BF%9E%E4%BA%91%E6%B8%AF+222061&rft.issn=1673-3819&rft.volume=46&rft.issue=3&rft.spage=116&rft.epage=122&rft_id=info:doi/10.3969%2Fj.issn.1673-3819.2024.03.017&rft.externalDocID=qbzhkzxtyfzjs202403017
thumbnail_s http://utb.summon.serialssolutions.com/2.0.0/image/custom?url=http%3A%2F%2Fwww.wanfangdata.com.cn%2Fimages%2FPeriodicalImages%2Fqbzhkzxtyfzjs%2Fqbzhkzxtyfzjs.jpg