1946 IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 11, NO. 11, NOVEMBER 2014
On-the-Fly Extraction of Polyhedral Buildings
From Airborne LiDAR Data
Lijun Xu, Senior Member, IEEE, Deming Kong, and Xiaolu Li, Member, IEEE
Abstract—This letter presents an on-the-fly method for extract-
ing polyhedral buildings from airborne light detection and ranging
(LiDAR) data. By using the gridding method, the planimetric po-
sition and elevation of laser footprints (normally treated as points)
in the obtained scan line are mapped into a data sequence. Then,
discrete stationary wavelet transform is applied to analyze the
elevation variation in the sequence. Buildings in the scan line can
be obtained from the detail wavelet coefficients of the sequence.
Moreover, to improve precision of the extraction, the gradients
of grid points in the geometric planes of building roofs along the
direction of the scan line are calculated and remedied by using
the corresponding gradients acquired from the adjacent scan lines.
With the proposed on-the-fly method, polyhedral buildings in the
scan area can be accurately extracted from laser points along
the scan lines during the scanning process. The new method is
validated by using a set of real airborne LiDAR data.
Index Terms—Building extraction, light detection and ranging
(LiDAR), on-the-fly data processing, remote sensing.
I. INTRODUCTION
C
URRENTLY, airborne light detection and ranging
(LiDAR) has been widely used in high-precision topog-
raphy, terrain investigation, and Earth observation. Compared
with traditional aerial photogrammetry, airborne LiDAR is less
affected by natural factors, such as weather and illumination
conditions, because the elevation values of ground and objects
on the ground can be directly obtained using the laser beam
emitted by a LiDAR instrument [1].
In the geographic information system, buildings are an indis-
pensable component. Due to their palpable artificial features,
buildings are often used as key objects in urban planning,
disaster prevention and control, national defense, and many
other applications of remote sensing. Consequently, the re-
search on rapid and high-precision extraction of buildings from
airborne LiDAR data has drawn wide attention [2]. Many
effective building extraction methods have been proposed, such
as the methods based on point cloud segmentation [3], range
Manuscript received September 29, 2013; revised January 6, 2014 and
March 18, 2014; accepted March 21, 2014. Date of publication April 22, 2014;
date of current version May 22, 2014. This work was supported in part by
the National Basic Research Program of China (973 Program) under Grant
2009CB72400106 and in part by the National Natural Science Foundation of
China under Grant 61121003, Grant 61201316, and Grant 61225006.
L. Xu is with the Ministry of Education’s Key Laboratory of Precsion
Opto-Mechatronics Technology, School of Instrument Science and Opto-
Electronic Engineering, Beihang University, Beijing 100191, China (e-mail:
lijunxu@buaa.edu.cn).
D. Kong and X. Li are with the State Key Laboratory of Inertial Science and
Technology, School of Instrument Science and Opto-Electronic Engineering,
Beihang University, Beijing 100191, China.
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/LGRS.2014.2314458
image processing [4], polygon generation [5], and 2-D wavelet
transform [6]. A common assumption in these studies is that
buildings have a polyhedral form [7], i.e., buildings only have
planar roofs. By using the aforementioned methods, polyhedral
buildings can be detected and extracted from airborne LiDAR
data. However, in real applications, there is a common short-
coming: all the aforementioned methods are more suitable for
processing LiDAR data in the schistose form (i.e., point cloud
or range image generated from the point cloud) than for directly
processing LiDAR data in each raw scan line obtained in the
scanning process. Since the point cloud cannot be obtained
until the whole flight over the target area is finished, building
extraction by using the aforementioned methods is generally
considered as a post-treatment process instead of an on-the-fly
process.
Over time, the line-scan frequency of airborne LiDAR in-
struments has increased significantly, up to 200 Hz for Riegl
LMS-Q680i, for instance. Nearly real-time extraction of build-
ings can be achieved if rapid processing of scan lines during
the LiDAR scanning process is realized. Methods have been
proposed. According to Axelsson [8], break points located in
the elevation waveforms of scan lines, i.e., the grid points
whose second derivatives of elevation values are nonzero, are
selected to extract the boundaries of the buildings. We found
in real applications that the elevation differences between the
leafy crowns of tall trees and the ground are quite similar with
the differences between roofs and the ground. Therefore, if
considering only the break points in the scan lines, buildings
cannot be accurately differentiated from tall trees.
The objective of this letter is to develop an on-the-fly method
to accurately extract polyhedral buildings from airborne LiDAR
data. Discrete stationary wavelet transform [9] is introduced
to detect buildings from the LiDAR data along the scan lines.
From the obtained detail wavelet coefficients, both the break
points and the gradient information of grid points along the
direction of the scan line can be acquired. By comparing the
gradient information of grid points in the scan line with the cor-
responding information of grid points in the adjacent scan lines,
buildings in the scan line can be accurately differentiated from
trees. Consequently, polyhedral buildings in the scan area can
be extracted nearly in real time during the scanning process of
airborne LiDAR.
II. M
ETHODOLOGY
A. Elevation Sequence Generation
In most applications, the airborne LiDAR instruments are
operated in parallel line-scan mode. Ideally, the obtained laser
1545-598X © 2014
IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
资源评论
weixin_38593644
- 粉丝: 4
- 资源: 914
最新资源
- 自定义参照引用后保存显示主键或显示为空问题处理
- 我国1950-2023年外汇及黄金储备汇总+趋势变化图
- YOLOX,YOLOV5,YOLOV8,YOLOV9 针对 OpenVINO 的 C++ 推理，支持 float32、float16 和 int8 .zip
- 设置NCC单据参照字段多选（参照多选）
- 已安装xcb、X11库的交叉编译器（x86-64-aarch64-linux-gnu）
- 包含约100万条由BELLE项目生成的中文指令数据
- BIP集成NC65预算
- 包含约50万条由BELLE项目生成的中文指令数据
- 完整的交叉编译好支持xcb的qt库（qt5.15.2、arm64、xcb、no-opengl）
- 包含约40万条由BELLE项目生成的个性化角色对话数据，包含角色介绍
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
