摘要
随着城市化的发展,中国的城市路网正以惊人的速度延伸和更新,如何高效、科学地管理庞大的路网成为难题。车载激光雷达系统具有野外作业量小、操作方便、大数据量采集等优点,可以用于恢复真实的道路场景,进一步实现街道级的道路变化检测。本文利用车载激光雷达扫描获取的点云数据,提出了一种基于数字高程模型和密度图的变化检测方法,用准确率、精准率和召回率作为检测的评价指标。以高速发展变化的上海陆家嘴地区的一个路段为例,验证了该方法检测的准确率为81.8%、精准率为93.1%、召回率为83.6%,变化检测效率得到大幅度提升。
The road network is expanding and updating at an amazing speed with the development of urbanization in China.Taking the advantages of less field work,convenient operation,and massive big data acquisition capability,the vehicle-borne LiDAR system are utilized to restore the real road scene,and further to realize street-level road change detection.A change detection method based on digital elevation model and density map in point cloud is proposed and a set of indicators for its accuracy,precision,and recall are established.A case study on the road segment experiencing great changes in Lujiazui of Shanghai is presented,which proves that the proposed method can efficiently recognize change locations and improve change detection efficiency.The accuracy,precision,and recall rate of the proposed method are 81.8%,93.1%,and 83.6%,respectively.
作者
赵婧文
ZHAO Jingwen(Shanghai Surveying and Mapping Institute,Shanghai 200092,China)
出处
《测绘与空间地理信息》
2021年第5期17-20,30,共5页
Geomatics & Spatial Information Technology
关键词
道路变化检测
车载激光雷达
数字高程模型
密度图
road change detection
vehicle-borne LiDAR
digital elevation model
density map
