A Review of Object Detection Techniques for Automobile Assisted Driving9
Zhaohong Gao, Haiyang Wang, Yongxin Sun, Feilong Wang · 2023
In this paper, the development of current target detection technology and its advantages and disadvantages are described in the context of foreign literature, mainly focusing on the advantages and disadvantages of traditional visual detection algorithms and deep learning detection methods in image target detection algorithms are analyzed and compared. Radar target detection algorithms: traditional point cloud features and deep learning point cloud features target detection algorithms are analyzed and compared. Finally in the analysis of multi-sensor fusion algorithms. Random model fusion methods and artificial intelligence fusion methods are described, respectively. The characteristics and research status of each detection method and fusion method are analyzed together with the literature, and the future development trend of assisted driving target detection methods is prospected: detection methods are lightweight to meet in-vehicle needs, three or more sensors are fused to achieve the purpose of improving detection accuracy, and multi-sensor fusion is the future development trend for target detection of harsh road conditions and small and distant targets.