作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2019, Vol. 45 ›› Issue (6): 12-20. doi: 10.19678/j.issn.1000-3428.0052284

所属专题: 智能交通专题

• 智能交通专题 • 上一篇    下一篇

基于部件融合特征的车辆重识别算法

李熙莹1,2,3,4,周智豪1,2,3,4,邱铭凯1,2,3,4   

  1. 1. 中山大学 智能工程学院,广州 510006
    2. 广东省智能交通系统重点实验室,广州 510006
    3. 视频图像智能分析与应用技术公安部重点实验室,广州 510006
    4. 视频图像信息智能分析与共享应用技术国家工程实验室,北京 100048
  • 收稿日期:2018-08-03 修回日期:2018-09-05 出版日期:2019-06-15 发布日期:2019-06-17
  • 作者简介:李熙莹(1972—),女,副教授、博士,主研方向为交通信息技术、视频图像、视频大数据技术|周智豪、邱铭凯,硕士研究生。
  • 基金资助:
    国家自然科学基金“视频大数据高效表达、深度分析与综合利用”(U1611461)

Vehicle Re-identification Algorithm Based on Component Fusion Feature

Xiying LI1,2,3,4,Zhihao ZHOU1,2,3,4,Mingkai QIU1,2,3,4   

  1. 1. School of Intelligent Systems Engineering,Sun Yat-sen University,Guangzhou 510006,China
    2. Guangdong Province Key Laboratory of Intelligent Transportation System,Guangzhou 510006,China
    3. Key Laboratory of Video and Image Intelligent Analysis and Application Technology,Ministry of Public Security of PRC,Guangzhou 510006,China
    4. National Engineering Laboratory of Video and Image Information Intelligent Analysis and Sharing Application Technology,Beijing 100048,China
  • Received:2018-08-03 Revised:2018-09-05 Online:2019-06-15 Published:2019-06-17

摘要:

针对车辆型号相同但车辆个体不同的重识别问题,提出一种新的车辆重识别算法。运用部件检测算法获取不同车辆之间差异较大的车窗和车脸区域,对检测到的车窗和车脸区域进行特征提取并进行融合,生成新的融合特征,计算图像特征之间距离度量进行分类识别。在中山大学公开数据集VRID-1上进行测试,结果表明,该算法的Rank1匹配率达到66.67%,明显优于经典的传统特征表征算法,从而验证该算法是可行且有效的。

关键词: 车辆重识别, 部件检测, 特征提取, 特征融合, 距离度量

Abstract:

To address the re-identification problem of different individual vehicles with identical types,a new vehicle re-identification algorithm is proposed.According to the component detection algorithm,the window and the vehicle face region with large differences between different vehicles are obtained,and the vehicle features of the detected vehicle window and the vehicle face region are extracted and merged to generate new fusion features.The distance measurement between image features is calculated for classification and recognition.The test is carried out on the public dataset VRID-1 of Sun Yat-sen university and results show that the Rank1 matching rate of the algorithm reaches 66.67%,which is obviously better than the classical traditional feature representation algorithm,thus verifies the feasibility and validity of the algorithm.

Key words: vehicle re-identification, component detection, feature extraction, feature fusion, distance measurement