矢量道路拓扑追踪匹配算法

发布时间：2018-04-13 04:24

本文选题：道路追踪 + 拓扑结构　；参考：《中国图象图形学报》2017年05期

【摘要】：目的基于道路形状特征的匹配算法在匹配性能上比较稳定,但当遇到道路交叉口等复杂路况时容易出现误匹配,且实时性上有一定缺陷,而矢量道路良好的拓扑结构,为此提出一种利用矢量道路拓扑关系进行追踪匹配的算法。方法算法利用结点、路段和路口这3种对象来对矢量道路进行表达,建立各个对象之间的拓扑关系,并将匹配过程划分为4个不同的状态,根据各个状态实施相应的匹配方法。首先,进行初始化、追踪、路口和搜索4个状态的定义和划分,确定各个状态之间的转换关系;进一步,设计道路中的结点、路段和路口3种对象的数据结构,建立点、线之间的空间拓扑关系;其次,根据4个状态的具体任务和实际特点,对进入该状态的行驶轨迹进行相应地分析处理和匹配计算;最后,根据追踪的结果进行匹配分析,完成对车辆行驶轨迹的误差修正。结果采用GPS-RTK采集的北京市西五环及密云地区的矢量道路数据对实地跑车的惯性导航轨迹进行拓扑追踪匹配仿真实验,完成拓扑追踪匹配算法的路口距离阈值选取,并与传统基于道路形状特征的匹配算法在匹配效果和实时性进行性能对比测试,其性能指标为匹配准确率和匹配时间。当矢量道路拓扑追踪算法的路口距离阈值取20 m时,匹配准确率达到了最高值93.5%。在匹配性能对比上,拓扑追踪算法相较于其他两种算法也有一定优势,在相同道路段中匹配准确率达到了90.2%,匹配速度也提高了4 8倍。结论采用矢量道路数据的拓扑信息对车辆轨迹进行追踪匹配的方法,能够用于卫星信号"盲区"或者信号干扰等特殊环境和场合的组合系统辅助导航,弥补传统基于卫星的组合导航在自主性、抗干扰性的不足。同时,算法针对复杂路况的匹配结果也较为理想,能够满足组合导航匹配工作的要求。
[Abstract]:Objective the matching algorithm based on road shape features is stable in matching performance, but it is prone to mismatch when complex road conditions such as road intersection are encountered, and there are some defects in real-time performance, and the vector road has a good topological structure.In this paper, an algorithm for tracking and matching using vector road topology is proposed.Methods the algorithm uses the node, section and intersection to express the vector road, establishes the topological relationship between each object, and divides the matching process into four different states, and implements the corresponding matching method according to each state.First of all, the definition and division of the four states of initialization, tracking, intersection and search are carried out, and the conversion relations between the states are determined. Furthermore, the data structures of the three objects in the road, road sections and intersections are designed, and the points are established.Secondly, according to the specific tasks and the actual characteristics of the four states, the corresponding analysis and matching calculation of the driving path into the state are carried out. Finally, according to the results of the tracking, the matching analysis is carried out.Complete the error correction of the vehicle track.Results the vector road data collected by GPS-RTK in the west fifth ring ring of Beijing and Miyun area were used to simulate the inertial navigation track of the field sports car, and the threshold of intersection distance was selected by the topology tracking matching algorithm.Compared with the traditional matching algorithm based on road shape feature, the performance of the algorithm is compared with the matching effect and real-time performance. The performance index is matching accuracy and matching time.When the threshold of intersection distance of the vector road topology tracking algorithm is 20 m, the matching accuracy reaches the highest value of 93.5um.Compared with the other two algorithms, the topology tracking algorithm has some advantages, and the matching accuracy reaches 90.2 in the same section of the road, and the matching speed is increased by 48 times.Conclusion using the topological information of vector road data to track and match the vehicle track can be used in the navigation of satellite signal "blind zone" or special environment and situation such as signal interference.To make up for the traditional satellite-based integrated navigation in autonomy, anti-jamming deficiencies.At the same time, the matching results of the algorithm for complex road conditions are also ideal, which can meet the requirements of integrated navigation matching.
【作者单位】：信息工程大学地理空间信息学院;地理信息工程国家重点实验室;郑州铁路局郑州车站;69027部队;空军航空大学;
【基金】：国家自然科学基金项目(41271450,41471336) “十二五”国家科技支撑计划(2012BAK12B02) 地理信息工程国家重点实验室2015年开放基金项目(SKLGIE2015-M-4-4)~~
【分类号】：U495

【参考文献】