基于LiDAR的目标检测算法与应用研究综述OA

[Objective]Environmental perception is a core task in autonomous driving,and high-precision object detection is essential for ensuring the safety and stability of autonomous driving systems.In recent years,LiDAR has been widely adopted in autonomous driving as a key sensor for three-dimensional perception due to its advantages,such as immunity to lighting conditions and high ranging accuracy.[Methods]This study first reviewed traditional camera-based object detection methods and analyzed their limitations in complex environments,then introduced the development history,working principles,types,and key parameters of LiDAR,followed by a systematic review of object detection methods based on point cloud representations,voxel representations,and multi-sensor fusion strategies.The network architectures,advantages,and challenges of different methods were compared and analyzed,and quantitative performance evaluations were conducted based on experimental results from the KITTI detection dataset.In addition,this study introduced a perception framework based on the bird's eye view(BEV)perspective and the trend of multi-sensor fusion,and analyzed the trade-offs between detection accuracy,real-time performance,and environmental adaptability of the current algorithms.[Results]The advantages of LiDAR-based object detection were summarized,and key future research directions were proposed to address challenges such as point cloud sparsity,high computational overhead,and the complexity of multimodal fusion.[Conclusions]Continuous optimization of algorithms and hardware enhances the accuracy,robustness,and practicality of LiDAR object detection in complex scenes.