摘要
在自动驾驶场景中,需要感知周围环境中的目标类型、大小、距离及方向等信息,因此3D目标检测技术对于自动驾驶汽车而言尤为重要。为获取周围环境信息,如激光雷达等传感器被应用在自动驾驶汽车上。为此,介绍自动驾驶领域基于激光雷达的3D目标检测现有方法,涵盖最新的模型,并对现有模型进行系统的分类,详细介绍具有代表性和开创性的方法,比较不同方法之间的优劣,并对基于点云的3D目标检测算法的发展进行展望。
In an self-driving scene,it is necessary to perceive information such as the class,size,distance,and orientation of the object in the surrounding environment.Therefore,3 D object detection technology is particularly important for autonomous vehicles.In order to obtain information about the surrounding environment,sensors such as lidar are used in autonomous vehicles.This article aims to introduce the existing methods of lidar-based 3 D object detection in the field of autonomous driving,covering the newest methods,systematically classifying the existing models,introducing in detail representative and pioneering methods,and comparing the pros and cons of the methods.Furthermore,we prospect the development of 3 D object detection algorithms based on point clouds.
作者
刘明宇
严俊
葛万成
LIU Mingyu;YAN Jun;GE Wancheng(Tongji University,Shanghai 200092,China)
出处
《电视技术》
2022年第1期150-155,173,共7页
Video Engineering
