3D OBJECT RECONSTRUCTION
IN CLOSE RANGE PHOTOGRAMMETRIC PROBLEMS
E. Stylianidis, P. Patias
Aristotle University of Thessaloniki, Department of Cadastre Photogrammetry & Cartography
e-mail: sstylianidis@yahoo.com, patias@topo.auth.gr
Commission V, WG V/4
KEY WORDS: Reconstruction, Modelling, Close Range, Algorithms, Matching
ABSTRACT:
It is well known that Photogrammetry has as main task the reconstruction of objects that presented in the images. Though a variety
of tools and techniques for fundamental photogrammetric procedures such as image orientations, image matching, interest point
extraction, orthoimage production etc., exist in an automatic form, in 3D reconstruction issues the research for automation still
remains under question. In close range photogrammetric problems the main point is that objects in images are characterized by
abrupt changes in surface in contrast to aerial case problems. These discontinuities must be taken into account in object modeling
and surface reconstruction, if a reliable and accurate work is necessary. Moreover, rapid changes in object surface are wittiness of
breaklines existence and therefore an a-priori knowledge of these graphical features gives additional and useful information for the
status of the model. Consequently, this assumption signifies a critical constrain in object surface reconstruction. In this paper a
developed algorithmic framework is described which leads to the 3D object reconstruction under close range conditions. The whole
process consists of 4 steps in turn and they involve issues like image processing, Hough Transform and image matching.
1. INTRODUCTION
A large number of structural elements are presented on the
images, in architectural, archaeological and industrial close
range photogrammetric problems. In most of the cases, these
objects are constructed by line elements that usually
characterize the sharp changes in object's surface discontinuity.
Edges on the images are denoted by this surface abrupt change,
which actually denote breaklines in the real world, ie. in
object's anaglyph.
For lots of years the development of tools for the detection of
these crucial graphical entities is a continuous challenge, very
useful and interesting. Indeed, lots of research and project have
been implemented in 3D reconstruction for aerial case issues,
especially in 3D city model reconstruction, building extraction
from aerial and space images etc.
In close range Photogrammetry where geometric and physical
conditions are different; in the research area more work needed
to be done. In any case, the relevant research that has already
implemeted in close range problems (Streilein, 1994; Pollefeys
1999; Stylianidis, 2001) is useful for further research. A
complete tutorial on 3D modeling from digital images
nowadays exists free on the web, which is very useful for basic
and advance knowledge in 3D modeling issues (Pollefeys,
2000).
The paper presents a complete framework designed to meet the
need for breakline detection and 3D model reconstruction using
the minimum user intervention. The research still proceeds to
reach the optimum conditions for automatic breakline detection.
2. BREAKLINES IN CLOSE RANGE OBJECTS
2.1 Introduction
Breaklines define and control the behavior of a surface model in
terms of smoothness and continuity. As their name implies,
breaklines are linear features and they can be shaped in an open
or close form. Breaklines can describe and enforce a change in
the behavior of the surface.
N
Figure 1. Breaklines in the object model
The surface discontinuity can be defined as a set of points
where the object’s anaglyph presents rough changes of its form.
Commonly, they referred as jump edges or occlusion
-216-
nD)
lA d m jd SN Jeu me 1
