摘要
在城市低空复杂环境中,无人机路径规划任务常依赖于三维建模与环境感知,导致系统部署复杂、作业成本较高。本文提出了一种基于城市道路图建模与Dijkstra算法的路径规划方法,将城市道路结构抽象为二维拓扑图模型,在不依赖高精度感知数据的前提下,实现路径长度最短的航线规划。试验选取深圳市典型城区作为测试区域,基于高德地图构建图模型,并应用Dijkstra算法进行路径搜索。通过与DFS算法枚举路径中随机抽取样本的对比,以及大疆无人机的实飞验证,结果表明所生成路径在长度与飞行可达性方面表现更优。该方法计算简洁、便于实现,适用于城市导航类软件系统,在实际作业中具有良好的工程适用性。未来将进一步拓展地图应用范围,并结合测绘与无人机任务软件,推动自动化航线生成在实际作业流程中的集成与应用。
In complex low-altitude urban environments,UAV path planning often relies on 3D modeling and environmental sensing,leading to increased system complexity and deployment costs.This paper presents a lightweight path planning method based on urban road network modeling and the Dijkstra algorithm.By abstracting intersections and road segments into a 2D topological graph and using path length as the edge weight,the algorithm computes the shortest route without requiring real-time perception or 3D reconstruction.A typical urban area in Shenzhen was selected for testing.Road data were extracted from the Gaode map platform to construct the graph,and planned routes were compared against randomly sampled paths generated by depth-first search(DFS).The resulting paths were also validated through real-world UAV flight experiments.The results demonstrate that the proposed method produces more efficient and executable routes.Due to its low computational cost and ease of integration,the method is well-suited for road-based navigation systems and frequent mapping tasks in urban environments.Future work will expand the geographic scope and integrate the method with UAV mission and mapping software to support fully automated route generation within operational workflows.
作者
曾理
ZENG Li(Shenzhen Polytechnic University,Shenzhen 518055,China)
出处
《测绘通报》
北大核心
2025年第S1期233-237,246,共6页
Bulletin of Surveying and Mapping
关键词
无人机
智慧城市
自主测绘
路径规划
图论
人工智能
UAV
smart city
autonomous mapping
path planning
graph theory
artificial intelligence
作者简介
曾理(1993-),女,博士,主要研究无人机应用技术。E-mail:zengli@szpu.edu.cn
