基于多传感器融合的汽车环境感知与目标识别技术

打开文本图片集

中图分类号：U463.6 文献标识码：A 文章编号：1003-8639（2026）03-0061-03

Automotive Environment Perception and Target Recognition Technology Based on Multi-sensor Fusion

Chen Ruojing （Nanjing Tech University，Nanjing 211816，China）

【Abstract】 This paper proposes a 3D object detection algorithm integrating LiDAR and monocular camera data. First，acoordinateunification modelbasedonjointcalibrationand hard-triggeredtimesynchronizationaddressesspatial and temporaldiscrepancies inheterogeneous data.Second，afeature-level fusion network withatentionmechanisms is designed.Dual-stream feature extraction branches separately obtain image semantic features and point cloud spatial geometric features.Across-modalatention module is incorporated to adaptivelyweightand fuse features，adressing he challenge of aligning sparse point clouds and dense images in feature space.

【Key words】multi-sensor fusion;environmental perception; object recognition

0 引言

伴随着人工智能、传感器技术的迅速发展，自动驾驶汽车正从实验室走向开放道路的应用阶段。（剩余4264字）