主办：陕西省汽车工程学会
ISSN 1671-7988  CN 61-1394/TH
创刊：1976年
汽车实用技术

汽车实用技术 ›› 2022, Vol. 47 ›› Issue (1): 45-50.DOI: 10.16638/j.cnki.1671-7988.2022.001.011

• 智能网联汽车 • 上一篇    

基于匈牙利匹配和卡尔曼滤波的动态多目标跟踪

李建国，张 睿，王 凯，曾 杰   

  1. 陕西重型汽车有限公司汽车工程研究院
  • 出版日期:2022-01-15 发布日期:2022-01-15
  • 通讯作者: 李建国
  • 作者简介:李建国，就职于陕西重型汽车有限公司汽车工程研究 院。

Multi-Object Tracking Algorithm Based on Hungarian Matching and Kalman Filtering

LI Jianguo, ZHANG Rui, WANG Kai, ZENG Jie   

  1. Shaanxi Heavy Duty Automobile Co., Ltd., Automotive Engineering Research Institute
  • Online:2022-01-15 Published:2022-01-15
  • Contact: LI Jianguo

PDF

摘要： 为了得到无人驾驶汽车目标检测的准确信息，文章通过对标定后的原始激光雷达点云进行 分割和聚类，利用霍夫直线检测提取障碍物的外包框，最后使用匈牙利算法对被跟踪的障碍物和 新检测出的障碍物进行关联匹配，并求出最优解，最终使用卡尔曼滤波算法对其进行状态优化。 通过实验表明，本算法可以在激光雷达一帧 0.1 s 内，快速、准确地跟踪被检测出的障碍物。

关键词: 匈牙利匹配；霍夫直线检测；卡尔曼滤波

Abstract: In order to obtain accurate information about the target detection of unmanned vehicles,by segmenting and clustering the calibrated original lidar point cloud, the Hough line is used to detect the outsourcing frame of the obstacle, and finally the Hungarian algorithm is used to correlate and match the tracked obstacle and the newly detected obstacle, and find the optimal solution, and finally use the Kalman filter algorithm to optimize its state. Experiments show that the algorithm can quickly and accurately track the detected obstacles within 0.1 s of a lidar frame.

Key words: Hungarian matching; Hough line detection; Kalman filtering