基于车载激光点云的街景立面自动提取

doi:10.12082/dqxxkx.2018.170440

地球信息科学学报 ›› 2018, Vol. 20 ›› Issue (4): 480-488.doi: 10.12082/dqxxkx.2018.170440

• 全国激光雷达大会特约稿件 • 上一篇下一篇

基于车载激光点云的街景立面自动提取

耿雨馨(), 钟若飞^*(), 彭宝江

1. 首都师范大学资源环境与旅游学院,北京 100048
2. 首都师范大学北京成像技术高精尖创新中心,北京 100048
3. 首都师范大学三维数据获取与应用重点实验室,北京 100048

收稿日期:2017-09-21 修回日期:2018-01-11 出版日期:2018-04-20 发布日期:2018-04-20
通讯作者: 钟若飞 E-mail:378259897@qq.com;40427310@qq.com
作者简介:
作者简介：耿雨馨（1992-）,女,硕士生,主要从事点云数据处理相关研究。E-mail: 378259897@qq.com
基金资助:
国家自然科学基金项目(41371434)

Automatic Extraction of Vista Facades Based on Vehicle-borne Laser Point Cloud

GENG Yuxin(), ZHONG Ruofei^*(), PENG Baojiang

1. College of Resources Environment and Tourism, Capital Normal University, Beijing 100048, China
2. Beijing Advanced Innovation Center for Imaging Technology,Capital Normal University, Beijing 100048, China
3. Key Lab of 3D Information Acquisition and Application, Capital Normal University, Beijing 100048, China

Received:2017-09-21 Revised:2018-01-11 Online:2018-04-20 Published:2018-04-20
Contact: ZHONG Ruofei E-mail:378259897@qq.com;40427310@qq.com
Supported by:
National Natural Science Foundation of China, No.41371434.

摘要/Abstract

摘要：

街道景观图是城市规划设计和城市管理的重要参考依据,车载点云数据能够提供沿街建筑的三维点信息,精度高,覆盖范围广泛,为街景立面整治提供了新的解决方案。为此,本文提出一种适用于车载点云的街景立面的自动提取方法,提取立面点云的具体步骤为：对原始数据去噪滤波;选取非地面点构建规则格网并二值化,依据语义特征筛选出建筑物点云;用POS数据拟合直线段帮助选取参考向量与参考平面;计算点云到参考面的距离,按距离分类点云数据,并对前述步骤中未分类点另行提取,合并面点集得到以沿街建筑物立面为主的街景立面点云。为了验证这一方法的可行性和有效性,采用点云数据进行实验,实验结果表明本方法在一定程度上提高了数据处理效率,能得到较理想的结果。

关键词: 立面提取, 车载点云, 街景立面, POS轨迹, 法向量

Abstract:

The street landscape describes the view of buildings and other objects on both sides of the road and works as the window of a city′s overall image. It is quite vital for urban planning and design, which could be helpful reference for management of government. Vehicle-borne point cloud data, with high precision and wide coverage, can provide position information and shape characteristics of the buildings along streets. This makes it possible to provide a new solution for urban vista facades extraction. Based on facades management, we propose a novel approach for automatic extraction of vista facades from vehicle-borne laser scanning data. The detail introduction focuses on the extraction of facades. In the approach, we divided the ground points and non-ground points after de-noising raw data and separated buildings from non-ground points in order to extract vista facades. It works in following four steps: (1) denoise raw point cloud and remove surface feature points from the raw data in order to acquire points of objects on the ground; (2) construct regular grids for non-ground points with binarization processing and select points of building according to semantic features; (3) estimate the reference vectors via POS(Positioning and Orientation System) data and set those vectors as normal vectors of chosen reference planes; (4) compute the Euclidean distance between each point and each plane. Points are classified by the distances with the same plane, based on which we extract point cloud of vista facades. To verify the feasibility and effectiveness of this method, we used a large group of vehicle-borne laser point cloud to carry out a series of experiments, including separating buildings from ground in origin data,extracting facade points from building points and comparing the automatic extraction with manual selection and results of other methods. The results showed that the method could improve the efficiency of data processing to some extent and return good results. The superiority of it was also verified by experiments.

Key words: facade extraction, vehicle-borne point cloud, vista facades, POS data, normal vector

耿雨馨, 钟若飞, 彭宝江. 基于车载激光点云的街景立面自动提取[J]. 地球信息科学学报, 2018, 20(4): 480-488.DOI:10.12082/dqxxkx.2018.170440

GENG Yuxin,ZHONG Ruofei,PENG Baojiang. Automatic Extraction of Vista Facades Based on Vehicle-borne Laser Point Cloud[J]. Journal of Geo-information Science, 2018, 20(4): 480-488.DOI:10.12082/dqxxkx.2018.170440

图/表 13

表1

图1

图2

图3

图4

图5

图6

图7

图8

图9

表2

图10

表3

参考文献 23

[1]	王昱. 基于车载LiDAR数据和街景照片的街道美景度评价[D].南京:南京大学,2016.
	[ Wang Y.Whether the street is beautiful? A novel evaluation method based on mobile LiDAR data and street view[D]. Nanjing: Nanjing University, 2016. ]
[2]	万怡平,习晓环,温奇,等.地面点云数据快速重建建筑模型的分层方法[J].测绘工程,2015(5):47-51. doi: 10.3969/j.issn.1006-7949.2015.05.011
	[ Wan Y P, Xi X H, Wen Q, et al.Fast building reconstruction based on a layered Method using TLS data[J]. Engineering of Surveying & Mapping, 2015(5):47-51. ] doi: 10.3969/j.issn.1006-7949.2015.05.011
[3]	顾斌. 数字图像与激光点云配准及在建筑物三维建模中的应用[D].北京:中国矿业大学,2014.
	[ Gu B.Registration of digital images and laser points cloud and its application in 3D model of buildings[D]. Beijing: China University of Mining and Technology, 2014. ]
[4]	Arachchige N H, Perera S N, Maas H G. Automatic processing of mobile laser scanner point clouds for Building FAÇADE Detection[J]. ISPRS - international archives of the photogrammetry, remote sensing and spatial information sciences, 2012,XXXIX-B5:187-192. doi: 10.5194/isprsarchives-XXXIX-B5-187-2012
[5]	Monszpart A, Mellado N, Brostow G J, et al. RAPter: Rebuilding man-made scenes with regular arrangements of planes[C]. ACM Transactions on Graphics, 2015,34(4):103:1-12.
[6]	Lalonde J, Vandapel N, Huber D F, et al.Natural terrain classification using three-dimensional ladar data for ground robot Mobility[J]. Journal of Field Robotics, 2006,23(10):839-861. doi: 10.1002/rob.20134
[7]	江恒彪. 基于实体化的地面点云建模技术[J].测绘工程,2014,23(7):30-33.
	[ Jiang H B.A modeling technique of point cloud based on entity-classification[J]. Engineering of Surveying & Mapping. ]
[8]	杨必胜,董震,魏征,等.从车载激光扫描数据中提取复杂建筑物立面的方法[J].测绘学报,2013,42(3):411-417.
	[ Yang B S, Dong Z, Wei Z, et al.Extracting complex building facades from mobile laser scanning data[J]. Acta
	Geodaetica et Cartographica Sinica, 2013,42(3):411-417. ]
[9]	Vanegas C A, Aliaga D G, Benes B.Automatic extraction of manhattan-world building masses from 3D laser range scans[J]. IEEE Transactions on Visualization & Computer Graphics, 2012,18(10):1627-1637. doi: 10.1109/TVCG.2012.30 pmid: 22291152
[10]	Börcs A, Nagy B, Benedek C.Fast 3-D Urban Object Detection on Streaming Point Clouds[C]. European Conference on Computer Vision. Springer, Cham, 2014:628-639.
[11]	Zhang C, Wang L, Yang R.Semantic segmentation of urban scenes using dense depth maps[C]. European Conference on Computer Vision. Springer-Verlag, 2010:708-721.
[12]	彭晨,余柏蒗,吴宾,等.基于移动激光扫描点云特征图像和SVM的建筑物立面半自动提取方法[J].地球信息科学学报,2016,18(7):878-885. doi: 10.3724/SP.J.1047.2016.00878
	[ Peng C, Yu B L, Wu B, et al.A method for semiautomated segmentation of building facade from mobile laser scanning point cloud based on feature images and SVM[J]. Journal of Geo-information Science, 2016,18(7):878-885. ] doi: 10.3724/SP.J.1047.2016.00878
[13]	熊伟成,杨必胜,董震.面向车载激光扫描数据的道路目标精细化鲁棒提取[J].地球信息科学学报,2016,18(3):376-385. doi: 10.3724/SP.J.1047.2016.00376
	[ Xiong W C, Yang B C, Dong Z.Refining and robust extraction of roads from mobile laser scanning point clouds[J]. Journal of Geo-information Science, 2016,18(3):376-385. ] doi: 10.3724/SP.J.1047.2016.00376
[14]	周华伟. 地面三维激光扫描点云数据处理与模型构建[D].昆明:昆明理工大学,2011.
	[ Zhou H W.Data processing and model building of ground 3D laser scanning point cloud[D]. Kunming: Kunming University of Science and Technology, 2011. ]
[15]	吴芳,张宗贵,郭兆成,等.基于机载LiDAR点云滤波的矿区DEM构建方法[J].国土资源遥感,2015,27(1):62-67. doi: 10.6046/gtzyyg.2015.01.10
	[ Wu F, Zhang Z G, Guo Z C, et al.Method of deriving DEM in the mining area based on filtering of airborne LiDAR data[J]. Remote Sensing for Land and Resources, 2015,27(1):62-67. ] doi: 10.6046/gtzyyg.2015.01.10
[16]	Zhang K, Chen S C, Whitman D, et al.A progressive morphological filter for removing nonground measurements from airborne lidar data[J]. IEEE Transactions on Geoscience & Remote Sensing, 2003,41(4):872-882. doi: 10.1109/TGRS.2003.810682
[17]	谭贲,钟若飞,李芹.一种车载激光扫描数据的地物分类方法[J].遥感学报, 2012,16(1)50-66. doi: 10.11834/jrs.20120408
	[ Tan B, Zhong R F, Li Q.Objects classification with vehicle-borne laser scanning data[J]. Journal of Remote Sensing, 2012,16(1):50-66. ] doi: 10.11834/jrs.20120408
[18]	Coughlan J M, Yuille A L.Manhattan world: Compass direction from a single image by Bayesian inference[C]. The Proceedings of the Seventh IEEE International Conference on Computer Vision. IEEE, 1999:941-947.
[19]	周煜,刘勐,马正东,等.基于分层聚类和拓扑连接模型的点云自适应简化[J].农业机械学报,2016,47(12):416-423. doi: 10.6041/j.issn.1000-1298.2016.12.053
	[ Zhou Y, Liu M, Ma Z D, et al.Adaptive simplification for point cloud based on hierarchical clustering and topological connectivity model[J]. Transactions of The Chinese Society of Agricultural Machinery, 2016,47(12):416-423. ] doi: 10.6041/j.issn.1000-1298.2016.12.053
[20]	Fischler M A, Bolles R C.Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography[J]. Communications of the ACM, 1981,24(6):381-395. doi: 10.1145/358669.358692
[21]	许烨璋,王鑫森,郑德华,等.一种改进的RANSAC算法提取多模型圆弧特征点云[J].测绘工程,2015(1):28-32.
	[ Xu Y Z, Wang X S, Zheng D H, et al.An improved algorithm of RANSAC to extract feature point cloud of multi-model arc[J]. Engineering of Surveying & Mapping, 2015(1):28-32. ]
[22]	余洁,余峰,张晶,等.结合区域生长与道路基元的高分辨率遥感影像道路提取[J].武汉大学学报·信息科学版,2013,38(7):761-764.
	[ Yu J, Yu F, Zhang J, et al.High resolution remote sensing image road extraction combining region growing and road-unit[J]. Geomatics and Information Science of Wuhan University, 2013,38(7):761-764. ]
[23]	Fei Y, Fan Z, Wang R M, et al.A fast and efficient mesh segmentation method based on improved region growing[J]. China Society for Industrial and Applied Mathematics B (English version), 2014,29(4):468-480. doi: 10.1007/s11766-014-3240-0

地物类别	分布特征	形状结构	点云数目
建筑物	与地面垂直高,一般不低于10 m	面状,平面或非平面	多,竖直方向点密度大,水平方向点密度小
树	木本上部呈类球状,下部呈竖直杆状,较高,约3、4 m;灌木呈类球状,较矮,一般3 m以内;群体沿道路呈带状分布	类球状或类球状与杆状组合	较多,局部密集
路灯	竖直方向呈杆状延伸,同街道路灯高度一致较高,一般不低于6 m	细长杆状	少
标志牌	由矩形或圆形的规则平面与细杆组成	杆状和面状组合	少,竖直方向局部密集
车辆	属于较低地物,一般不超过3 m	带圆角的块状	较少,点密度比较均匀

提取方法	总数量	点云总数目/个	误题数目/个	误提点数目/个	正确率/%	完整率/%
人工提取建筑物	12	3 239 703（噪点约 70 万）	-	-	-	-
本文提取建筑物	13	2 460 954	1	78 749	96.899	92.308
人工提取立面	16	456 853（噪点约 10 万）	-	-	-	-
本文提取立面	17	343 531	1	13 322	96.267	93.750

提取方法	提取点云数目/个	多提点数目/个	漏提点数目/个	准确率/%	时间消耗/s
人工提取	653 929	-	-	-	-
本文方法	632 066	922	22 785	96.517	346.051
区域生长	656 265	38 449	36 113	94.478	514.227
文献[12]方法	604 799	2034	51 164	92.176	987.539

基于车载激光点云的街景立面自动提取

Automatic Extraction of Vista Facades Based on Vehicle-borne Laser Point Cloud

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 13

参考文献 23

相关文章 2

Metrics

本文评价

推荐阅读 0

[1]	陶鹏杰, 宋孟肖, 段延松. 利用高精度DLG进行机载LiDAR点云精度评定研究[J]. 地球信息科学学报, 2019, 21(4): 608-614.
[2]	彭晨, 余柏蒗, 吴宾, 吴健平. 基于移动激光扫描点云特征图像和SVM的建筑物立面半自动提取方法[J]. 地球信息科学学报, 2016, 18(7): 878-885.