检索规则说明:AND代表“并且”;OR代表“或者”;NOT代表“不包含”;(注意必须大写,运算符两边需空一格)
检 索 范 例 :范例一: (K=图书馆学 OR K=情报学) AND A=范并思 范例二:J=计算机应用与软件 AND (U=C++ OR U=Basic) NOT M=Visual
作 者:方莉娜[1,2,3] 王爽 赵志远 付化胜[4] 陈崇成 FANG Lina;WANG Shuang;ZHAO Zhiyuan;FU Huasheng;CHEN Chongcheng(National Engineering Research Centre of Geospatial Information Technology, Fuzhou University, Fuzhou 350002, China;Key Laboratory of Spatial Data Mining and Information Sharing of Ministry of Education, Fuzhou University, Fuzhou 350002, China;Academy of Digital China, Fuzhou University, Fuzhou 350002, China;Fujian Provincial Investigation, Design & Research Institute of Water Conservancy & Hydropower, Fuzhou 350002, China)
机构地区:[1]福州大学地理空间信息技术国家地方联合工程研究中心,福建福州350002 [2]空间数据挖掘与信息共享教育部重点实验室,福建福州350002 [3]福州大学数字中国研究院,福建福州350002 [4]福建省水利水电勘测设计研究院,福建福州350002
出 处:《测绘学报》2021年第9期1251-1265,共15页Acta Geodaetica et Cartographica Sinica
基 金:国家自然科学基金(42071446);福建省对外合作项目(2020l0007)。
摘 要:交通标线是重要的交通安全设施,其位置、属性和拓扑关系精细刻画道路交通结构,是智能交通、高精地图、位置与导航等应用的基础数据。本文提出一种融合空间上下文信息的车载激光点云标线分类图注意力模型(graph attention network with spatial context information,GAT_SCNet)。该模型利用图结构建立标线及其邻接对象的出现和依存关系,基于标线几何、拓扑、空间结构关系构建注意力机制进行节点特征动态更新,通过对节点分类实现标线的精细分类。基于分类后标线,设计不同方案实现对分类后标线提取标线矢量化数据。试验采用4份不同车载激光扫描系统获取的城市与高速场景数据验证本文方法的有效性,试验结果中9类标线分类的准确率分别为100.00%、93.77%、100.00%、100.00%、100.00%、96.73%、97.96%、100.00%、98.39%,召回率分别为100.00%、96.36%、100.00%、100.00%、100.00%、97.26%、85.72%、100.00%、94.16%。结果表明,本文方法能实现道路场景中全尺寸、多类型标线对象的精确识别,并对形状相似标线(如虚线、斑马线和停止线)的区分具有较强稳健性。Road markings are important traffic safety facilities.Its location,attribute,and topological relationship finely describe road traffic structure,and it is the basic data for applications such as intelligent traffic,high-precision maps,location,and navigation.This paper proposes a graph attention network with spatial context information(GAT_SCNet)to classify the road markings from mobile LiDAR point clouds.GAT_SCNet explores the graph structure to establish the appearance and dependence information among road markings.Meanwhile,GAT_SCNet incorporates the multi-head attention mechanism into the node propagation step,which computes the hidden states of each node based on the geometric,topological,and spatial structure relationships of the neighboring nodes.Finally,road markings classification is realized by the classification of nodes.Then,some schemes are designed for road markings vectorization.Four test datasets consisting of urban and highway scenes by different mobile laser scanning systems are used to evaluate the validities of the proposed method.Four accuracy evaluation metrics precision and recall of 9 types of road markings on the selected test datasets achieve(100.00%,93.77%,100.00%,100.00%,100.00%,96.73%,97.96%,100.00%,98.39%)and(100.00%,96.36%,100.00%,10.000%,100.00%,97.26%,85.72%,100.00%,94.16%),respectively.Accuracy evaluations and comparative studies prove that the proposed method has the capability of classifying multi-type road markings simultaneously and distinguishing similar road markings like dashed markings,zebra crossings,and stop lines in complex urban scenes.
关 键 词:车载激光点云 标线分类 图结构 注意力机制 标线矢量化
分 类 号:P307[天文地球—地球物理学]
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在链接到云南高校图书馆文献保障联盟下载...
云南高校图书馆联盟文献共享服务平台 版权所有©
您的IP:216.73.216.63