自动驾驶汽车目标检测方法综述  

Overview of target detection methods for autonomous vehicles

作　　者：李爱娟[1] 巩春鹏 黄欣 曹家平 刘刚 LI Aijuan;GONG Chunpeng;HUANG Xin;CAO Jiaping;LIU Gang(School of Automotive Engineering, Shandong Jiaotong University, Jinan 250357, China;School of Information Science and Electrical Engineering, Shandong Jiaotong University, Jinan 250357, China;Academic Affairs Office, Shandong Jiaotong University, Jinan 250357, China)

机构地区：[1]山东交通学院汽车工程学院,山东济南250357 [2]山东交通学院信息科学与电气工程学院,山东济南250357 [3]山东交通学院教务处,山东济南250357

出　　处：《山东交通学院学报》2022年第3期20-29,共10页Journal of Shandong Jiaotong University

基　　金：国家自然科学基金项目(51505258,61601265,51405272);山东省自然科学基金项目(ZR2015EL019,ZR2020ME126,ZR2021MF131);山东省高等学校“青创科技计划”项目(2019KJB019);国家重点实验室开放课题(1903);河北省交通安全与控制重点实验室开放项目(JTKY2019002)。

摘　　要：目标检测是自动驾驶汽车环境感知的重要内容。综合国内外部分研究文献,基于目标检测用传感器的数量及种类对自动驾驶汽车目标检测方法进行分类,总结分析每种目标检测方法的特点及研究现状,展望自动驾驶汽车目标检测方法的未来发展趋势为:基于视觉检测方法的轻量级车载应用,3种及以上传感器数据融合技术,多传感器信息融合在小目标和集群目标检测及复杂恶劣环境下的目标检测应用。Target detection is an important research content for environment perception of autonomous vehicles.Some literature from domestic and foreign researchers is synthesized to classify target detection methods based on the number and types of sensors used for target detection.The characteristics and research status of each target detection method are summarized and analyzed,and the future development directions of target detection methods for autonomous vehicles are looked forward to,including lightweight on-board applications based on vision detection methods,data fusion technology with three and more sensors,and the applications of multi-sensor information fusion in small target,cluster target detection and target detection under complex and harsh environments.

关 键 词：自动驾驶汽车 目标检测 传感器 多传感器信息融合 

分 类 号：U471.1[机械工程—车辆工程] TP273[交通运输工程—载运工具运用工程]