没有合适的资源?快使用搜索试试~ 我知道了~
从地面激光扫描数据中自动提取建筑物
5 下载量 92 浏览量
2021-03-03
02:23:35
上传
评论
收藏 984KB PDF 举报
温馨提示
试读
6页
从大量具有不同局部密度的点云中提取建筑物,尤其是在存在随机噪声点的情况下,仍然是一个巨大的挑战。 在本文中,我们提出了一种从地面激光扫描数据中提取建筑物的完整策略。 首先,提出了一种新颖的分割方法来促进建筑物提取的任务。 这些点基于法线和邻接关系进行分组。 其次,基于高斯图像的属性,从分割结果中识别出平面。 最后,根据形状,法线方向和拓扑关系等点云段的特征集合,从城市点云中提取建筑物。 实验结果表明,该方法可作为一种可靠的方法从地面激光扫描数据中提取建筑物。 同时,建筑物被分解成几个小块,为建筑物的重建打下了良好的基础。
资源推荐
资源详情
资源评论
Advances in Electrical and Computer Engineering Volume 13, Number 3, 2013
Automatic Building Extraction from Terrestrial
Laser Scanning Data
Wen HAO
1
, Yinghui WANG
1
, Xiaojuan NING
1
, Minghua ZHAO
1
, Jiulong ZHANG
1
, Zhenghao SHI
1
,
Xiaopeng ZHANG
2
1
Institute of Computer Science and Engineering, Xi'an University of technology, Xi'an, 710048, China
2
National Laboratory of Pattern Recognition, CAS Institute of Automation, Beijing, 100190, China
haowen@stu.xaut.edu.cn
1
Abstract—The extraction of building from the huge amount
of point clouds with different local densities, especially in the
presence of random noisy points, is still a formidable challenge.
In this paper, we present a complete strategy for building
extraction from terrestrial laser scanning data. First, a novel
segmentation method is proposed to facilitate the task of
building extraction. The points are grouped based on the
normals and the adjacency relationships. Second, the planar
surfaces are recognized from the segmentation results based on
the properties of the Gaussian image. Finally, the buildings are
extracted from the urban point clouds based on a collection of
characteristics of point cloud segments like shape, normal
direction and topological relationship. Experimental results
demonstrate that the proposed method can be used as a robust
way to extract buildings from terrestrial laser scanning data.
At the same time, the buildings are decomposed into several
patches which lay a good foundation for building
reconstruction.
Index Terms—building extraction, point cloud segmentation,
plane recognition, terrestrial laser scanning.
I. INTRODUCTION
Numerous practical applications are related to buildings,
such as virtual tourism, urban planning and environmental
monitoring. Therefore, automatic extraction of building
from laser scanner data becomes necessity due to the
growing demand for urban planning and virtual tourism,
coupled with the advance in 3D data acquisition technology.
In the last decade, extensive studies about building
extraction have been undertaken on LiDAR(Light Detection
and Ranging) data[1-4] or image data[5]. However, since
urban scenes need to be realistic not only from a bird’s point
of view, but also from a pedestrian’s point of view, the
extraction of building from TLS(Terrestrial Laser Scanning)
data becomes essential. Recent advances in sensing and
laser technologies make TLS become a common way to
acquire 3D data of complex urban scenes. Unfortunately,
although techniques for the acquisition of 3D urban point
clouds via TLS have constantly been improved, processing
these large 3D data sets in order to extract static entities
such as buildings and roads is still a formidable challenge.
This is due essentially to the difficulties of exploring
directly and automatically valuable spatial information from
the massive unstructured 3D data.
1
This work is supported in part by National Natural Science Foundation
of China under Grant No.61072151, No.61272284; and in part by Shaanxi
Educational Science Research Plan under Grant No.2010JK734; and in part
by Shaanxi Science Research Plan under Grant No.2011K06-35,Xi’an
Science Research Plan under Grant No.CX1252(3);and in part by Doctoral
Fund of Ministry of Education of China under Grant No. 20126118120022.
In this paper, a complete strategy for building extraction
from TLS data is proposed. The inspiration of our method
comes from the facts that a majority of buildings in
existence nowadays could be represented by planes. At the
same time, ground and grass can also be represented by
using one or a group of large planar surfaces. Due to the
facts mentioned above, we present a clustering algorithm for
extracting homogeneous segments in point clouds based on
the normals and the adjacency relationships of the points.
Then the planar surfaces are recognized based on the
Gaussian image. After recognizing the ground in advance,
the planar surfaces belonging to the buildings are extracted.
The whole process of our method is described as follows
with the flowchart in Fig. 1.
(1) Segmentation. A novel clustering method for reliable
and efficient segmentation of the urban point clouds is
proposed. The clustering method requires no prior clustering
number compared to the K-means clustering method.
(2) Plane recognition. Since most of the building
components in existence nowadays are planes, a novel
method is introduced to recognize the planar surfaces based
on the properties of the Gaussian image.
(3) Building extraction. This process tries to identify the
surfaces belonging to the buildings from the given 3D point
clouds. We first recognize the ground based on the position
and normal direction of the planar surfaces. Besides the
planes completely containing in the Oriented Bounding
Box(OBB) of the ground, the residual planes are considered
as the buildings.
The remainder of the paper is organized as follows. After
a brief review of point cloud segmentation techniques in
Section 2, Section 3 presents our segmentation method in
detail. The method of building extraction is proposed in
section 4 and experimental results are shown in section 5.
The limitations of our method and future research are
indicated in the last section.
II. R
ELATED WORK
Object extraction from TLS data has been a research
domain in recent years. Segmentation, the process which
partitions point clouds into regions with homogeneous
property, is an essential step which needs to be performed
prior to object extraction. Many methods are known in
literature for point clouds segmentation, which roughly fall
into the following categories:
Model-based: this strategy tries to fit primitive shapes
like planes, cylinders or spheres in the point cloud.
11
Digital Object Identifier 10.4316/AECE.2013.03002
1582-7445 © 2013 AECE
[Downloaded from www.aece.ro on Sunday, October 27, 2013 at 04:54:00 (UTC) by 221.11.20.101. Redistribution subject to AECE license or copyright. Online distribution is expressly prohibited.]
资源评论
weixin_38601364
- 粉丝: 6
- 资源: 949
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功