汽车工程 ›› 2024, Vol. 46 ›› Issue (4): 557-563.doi: 10.19562/j.chinasae.qcgc.2024.04.001

• •    

面向车载相机采集图像的智能汽车测试场景关键性量化模型

朱冰,黄殷梓,赵健(),张培兴,薛经纬   

  1. 吉林大学,汽车仿真与控制国家重点实验室,长春 130022
  • 收稿日期:2023-08-09 修回日期:2023-08-24 出版日期:2024-04-25 发布日期:2024-04-24
  • 通讯作者: 赵健 E-mail:zhaojian@jlu.edu.cn
  • 基金资助:
    国家重点研发计划(2022YFB2503402);国家自然科学基金(U22A20247);中国博士后面上项目(2023M741354)

A Criticality Assessment Model for the Intelligent Vehicle Test Scenario Based on the Onboard Camera Images

Bing Zhu,Yinzi Huang,Jian Zhao(),Peixing Zhang,Jingwei Xue   

  1. Jilin University,State Key Laboratory of Automotive Simulation and Control,Changchun 130022
  • Received:2023-08-09 Revised:2023-08-24 Online:2024-04-25 Published:2024-04-24
  • Contact: Jian Zhao E-mail:zhaojian@jlu.edu.cn

摘要:

车载相机图像是构建智能汽车测试场景库的主要数据来源,但其中关键测试场景发生概率低,大部分场景的测试价值小,若将其直接应用于智能汽车测试会浪费大量测试资源。本文提出一种面向车载相机采集图像的智能汽车测试场景关键性量化模型。首先,基于实车相机参数对实车采集的图像进行处理,输出对行车安全具有影响的参数;其次,基于风险场理论将参数整合,输出测试场景关键性量化结果;最后,对实车采集的图像进行测试场景关键性量化验证,结果表明本文模型可以精确输出测试场景关键性的具体数值,进而对比不同场景的测试价值,证明本文提出的模型可以有效筛选智能汽车关键测试场景。

关键词: 智能汽车, 测试场景, 关键性量化模型, 车载相机图像, 风险场理论

Abstract:

Onboard camera images are the main data sources for constructing the intelligent vehicle test scenario library, but the probability of critical test scenarios occurring in the actual collected onboard camera images is very low, and most of the scenarios have little test value. If it is directly applied to the intelligent vehicle test, it will waste a lot of test resources. In this paper, a criticality assessment model for the intelligent vehicle test scenario based on the onboard camera images is proposed. Firstly, the images collected from real vehicles are processed based on the camera parameters to output parameters that have impact on driving safety. Then, the parameters are integrated using the risk field theory to output the criticality assessment results of the intelligent vehicle test scenario. Finally, the criticality assessment validation is conducted on the images collected from the actual vehicle. The results show that the proposed model can accurately output the specific values of the criticality of the test scenarios in order to compare the test values of different scenarios, proving that the model proposed in this paper can effectively screen the intelligent vehicle critical test scenarios.

Key words: intelligent vehicle, test scenario, criticality assessment model, onboard camera images, risk field theory