Adaptive Partitioning of Urban Facades

晃晃

Adaptive Partitioning of Urban Facades

Chao-Hui Shen1 Shi-Sheng Huang1 Hongbo Fu2 Shi-Min Hu1

1TNList, Tsinghua University, Beijing 2City University of Hong Kong







Abstract

Automatically discovering high-level facade structures in unorga-

nized 3D point clouds of urban scenes is crucial for applications

like digitalization of real cities. However, this problem is chal-

lenging due to poor-quality input data, contaminated with severe

missing areas, noise and outliers. This work introduces the con-

cept of adaptive partitioning to automatically derive a flexible and

hierarchical representation of 3D urban facades. Our key obser-

vation is that urban facades are largely governed by concatenated

and/or interlaced grids. Hence, unlike previous automatic facade

analysis works which are typically restricted to globally rectilin-

ear grids, we propose to automatically partition the facade in an

adaptive manner, in which the splitting direction, the number and

location of splitting planes are all adaptively determined. Such an

adaptive partition operation is performed recursively to generate a

hierarchical representation of the facade. We show that the con-

cept of adaptive partitioning is also applicable to flexible and ro-

bust analysis of image facades. We evaluate our method on a dozen

of LiDAR scans of various complexity and styles, and the image

facades from the eTRIMS database and the Ecole Centrale Paris

database. A series of applications that benefit from our approach

are also demonstrated.

1 Introduction

With the recent advances in LiDAR scanning devices, the acquisi-

tion of 3D point clouds from urban buildings is getting more ef-

ficient and more convenient. However, the captured point cloud

often suffers from severe missing data, noise and outliers, making

the reconstruction of architectural models with faithful geometry

and topology from such data rather challenging. The key to this

problem is to explore and utilize the characteristics of urban scenes

as prior knowledge, especially repetition of building elements in

facades. To obtain such architectural characteristics, the state-of-

the-art works [Zheng et al. 2010; Nan et al. 2010] rely on user as-

sistance, which would be labor intensive for applications like digi-

talization of real cities.

There exist automatic solutions for discovering facade structures in

single- or multi-view facade images [M¨ uller et al. 2007; Xiao et al.

2009; Musialski et al. 2010]. A common assumption made in those

works is that facades are inherently governed by global rectilin-

ear structures. That is, a facade can be split into building blocks

(e.g., windows) by a single rectilinear grid. Although there indeed

exist many real facades satisfying such assumption, it is not gen-

eral enough to handle many other patterns like asymmetric patterns

(Figure 1), which are also ubiquitous in urban scenes.

This work aims for a more flexible representation of high-level fa-

cade structures, which is based on the key observation: the high-

level structure of a facade is largely governed by either a rectilinear

grid or a mixture of rectilinear grids by concatenation and/or in-

terlacing. For examples, Figure 2 shows facades with concatenated

grids and interlaced grids. Therefore, how to identify different grids

of repetitive elements and their relations is our core problem.





全文请下载附件：

tc

一帆风顺，二龙腾飞，三羊开泰，四季平安，五福临门，六六大顺，七星高照，八方来财，九九同心，十全十美。

晃晃

万事胜意，合家幸福，人强马壮，心想事成，顺理成章 步步高升。

菜刀吻电线

好，真棒！！

菜刀吻电线

凡系斑竹滴话要听；凡系朋友滴帖要顶

菜刀吻电线

路过……

菜刀吻电线

不错 非常经典  实用

菜刀吻电线

不错哦,顶一下......