无人机多源异构感知融合的交通事故现场
三维重构方法

doi:10.16638/j.cnki.1671-7988.2026.020.023

汽车实用技术 ›› 2026, Vol. 51 ›› Issue (20): 141-146.DOI: 10.16638/j.cnki.1671-7988.2026.020.023

• 标准·法规·管理 • 上一篇

无人机多源异构感知融合的交通事故现场三维重构方法

张心怡 1，李凯璇 2，张泽江 2*

1.山西工程科技职业大学汽车工程学院； 2.衢州技师学院智慧交通学院

发布日期:2026-10-22
通讯作者: 张泽江
作者简介:张心怡（1994－），女，博士研究生，讲师，研究方向为智能网联汽车、无人机应用通信作者：张泽江（1999－），男，硕士，讲师，研究方向为智能网联汽车、无人机应用
基金资助:
山西工程科技职业大学校科研基金项目（KJ202509）

Three-dimensional reconstruction method for traffic accident scenes using unmanned aerial vehicle multi-source heterogeneous perception fusion

ZHANG Xinyi1 , LI Kaixuan2 , ZHANG Zejiang2*

1.School of Automotive Engineering, Shanxi Vocational University of Engineering Science and Technology, Jinzhong 030619, China; 2.Smart Transportation College, Quzhou Technical College

Published:2026-10-22
Contact: ZHANG Zejiang

摘要/Abstract

摘要： 针对道路交通事故现场人工勘察耗时长、单一航拍影像缺失空间几何结构信息、单一激光雷达点云缺少纹理信息等问题，文章提出一种基于无人机多源异构感知融合的事故现场三维重构方法。从事故现场取证需求出发，应用无人机负载激光雷达、相机和组合惯导采集事故现场的原始数据。以空间几何位置和色彩纹理为主要信息记录车辆姿态、地面散落物和周围环境等信息。通过点云预处理、空间配准和影像-点云关联实现点云赋色，完成事故现场空间几何结构、颜色纹理的三维模型重构。开展模拟事故现场实地实验，结果表明，融合重构模型均方根误差（RMSE）为 0.040 2 m，较单一航拍建模降低 50.92%，可为事故现场快速取证和二次勘察提供数字化支撑。

关键词: 道路交通事故；无人机；多源数据融合；三维重构

Abstract: To address the problems of time-consuming manual investigation, the lack of spatial geometric structure information in a single aerial image, and the lack of texture information in single light detection and ranging point clouds, this paper proposes a three-dimensional reconstruction method for traffic accident scenes based on unmanned aerial vehicle multi-source heterogeneous perception fusion. Accident evidence objects are organized around vehicle posture, road traces,scattered objects and road markings. Image data, light detection and ranging point clouds and global navigation satellite system/inertial measurement unit pose information are then associated through point cloud preprocessing, spatial registration, image-point cloud correspondence and point cloud coloring. A field experiment is conducted on a simulated accident scene, and the results show that the root mean square error (RMSE) of the fused reconstruction model is 0.040 2 m, which is 50.92% lower than that of single-image photogrammetry, and can provide digital support for rapid evidence collection and secondary investigation of traffic accident scenes.

Key words: traffic accident; unmanned aerial vehicle; multi-source data fusion; three-dimensional reconstruction

张心怡 1，李凯璇 2，张泽江 2*. 无人机多源异构感知融合的交通事故现场三维重构方法[J]. 汽车实用技术, 2026, 51(20): 141-146.

ZHANG Xinyi1 , LI Kaixuan2 , ZHANG Zejiang2*. Three-dimensional reconstruction method for traffic accident scenes using unmanned aerial vehicle multi-source heterogeneous perception fusion[J]. Automobile Applied Technology, 2026, 51(20): 141-146.

参考文献

[ 1 ] 安蕾冰.应用型课程建设研究:以“交通事故现场勘查”课程为例[J].科教导刊,2024(14):109-111. [ 2 ] 杨晓波,王亚州,庞雪纯,等.基于道路剩余通行能力的交通事故等级划分与管控[J].科学技术与工程, 2026,26(15):6600-6608. [ 3 ] 冯浩,李威,周闽琪,等.三维勘查技术在道路交通事故现场的应用[J].交通与运输,2025,41(5):16-20. [ 4 ] 安蕾冰.应用型课程建设研究:以“交通事故现场勘查”课程为例[J].科教导刊,2024(14):109-111. [ 5 ] 关闯,冯浩,衡威威,等.地面三维激光扫描在交通事故勘查及重建中的应用进展[J].刑事技术,2022,47 (1):76-82. [ 6 ] 王正礼,唐子墨,郑振杰.融合检索增强生成与思维链的交通事故责任认定大模型构建[J].交通运输系统工程与信息,2026,26(3):83-92. [ 7 ] 陈兵硕,张高强,文涛,等.基于无人机智慧巡检的高速公路交通事件应急处置技术体系研究[J].交通工程,2026,26(2):19-26,33. [ 8 ] Liu Xiaofeng,Guan Zhiwei,Fan Qiuhan,et al.Remote sensing and scene reconstruction of traffic accident based on unmanned aerial vehicle platform[C]//CICTP 2019:Transportation in China－Connecting the World. Reston:American Society of Civil Engineers,2019: 3331-3342. [ 9 ] 陈强,许洪国,谭立东.基于小型无人机摄影测量的交通事故现场勘察[J].吉林大学学报(工学版),2016,46 (5):1439-1446. [10] 李凯璇,刘晓锋,陈强,等.YOLOv8-GAIS:一种改进的无人机航拍目标检测算法[J].光电工程,2025,52(4): 77-90. [11] 张泽江,刘晓锋,刘军黎.基于机载激光雷达的低照度交通事故现场重建及实证研究[J].国外电子测量技术,2024,43(2):174-182. [12] Zhang Zejiang,Liu Xiaofeng,Chen Qiang,et al.Improving accident scene monitoring with multisource data fusion under low-brightness and occlusion conditions[J].IEEE Access,2024,12:136026-136040. [13] Chen Qiang,Li Da,Huang Bin,et al.SUAV image mosaic based on rectification for use in traffic accident scene diagramming[C]//2020 IEEE 3rd International Conference of Safe Production and Informatization. Piscataway:IEEE,2020:430-433. [14] Zhang Xinyi,Guan Zhiwei,Liu Xiaofeng,et al.3-D modeling of traffic accidents scenes using a UAV-borne SSL-camera system[J].IEEE Access,2025,13:214440- 214451. [15] Zhang Xinyi,Guan Zhiwei,Liu Xiaofeng,et al.Digital reconstruction method for low-illumination road traffic accident scenes using UAV and auxiliary equipment [J].World Electric Vehicle Journal,2025,16(3):171. [16] 徐志博,吕秋娟,甘鑫斌,等.基于最优邻域特征加权的点云引导滤波算法[J].激光与光电子学进展,2024,61 (14):229-236. [17] Lim H,Hwang S,Myung H.ERASOR:Egocentric ratio of pseudo occupancy-based dynamic object removal for static 3D point cloud map building[J].IEEE Robotics and Automation Letters,2021,6(2):2272-2279. [18] GA/T 1381－2018 道路交通事故现场图绘制系统通用技术条件[S]. [19] GA/T 49－2019 道路交通事故现场图绘制[S].

无人机多源异构感知融合的交通事故现场三维重构方法

Three-dimensional reconstruction method for traffic accident scenes using unmanned aerial vehicle multi-source heterogeneous perception fusion

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 0

编辑推荐

Metrics

无人机多源异构感知融合的交通事故现场 三维重构方法

Three-dimensional reconstruction method for traffic accident scenes using unmanned aerial vehicle multi-source heterogeneous perception fusion

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 0

编辑推荐

Metrics

无人机多源异构感知融合的交通事故现场三维重构方法