Full text: Proceedings, XXth congress (Part 3)

3D RECONSTRUCTION OF THE SOLID OF ROTATION LACKING IN TEXTURE 
: : a. * Ta ê 7 2 
Jianging Zhang *^', Jun Tao*, Zuxun Zhang * 
? Wuhan University, College of Remote Sensing and Information Engineering, 
#129 Luoyu Road, Wuhan, Hubei, P.R.China 
Jqzhang@supresoft.com.cn, martintao@etang.com, zxzhang@supresoft.com.cn 
KEY WORDS: Reconstruction, Three-dimensional, Image, Texture, Camera, Platform, Adjustment, Photogrammetry 
ABSTRACT: 
The paper proposed an approach to reconstruct the solid of rotation lacking in texture from two-dimensional images. The approach 
only requires a digital camera, an ordinary slide projector and a rotated platform. The camera and prôjector are required to be 
calibrated in advance. The rotated platform is functioned as a control ground and the target object is put on it. The projector projects 
the slide points onto the solid of rotation so that the texture is supplied to the surface of it. The digital camera takes sequential images 
with these points when the platform rotates continually. For each space feature point projected on the surface of the solid of rotation, 
there are two corresponding 2D points existing. One is an image point in one of the image sequences and another is a point within 
the slide points. Gotten the two 2D coordinates of the points, the 3D coordinates of the space points on the surface of the solid of 
rotation can be calculated by the collinear equations. The sequential images are taken from different directions of the solid of 
rotation. Using the correspondence of whole adjustment and the inherent structure characteristic of the solid of rotation, the 3D 
coordinates of all space points projected on the whole surface of the solid of rotation can be computed out entirely. The 3D model of 
this solid of rotation is acquired by connecting all neighbour space points. This approach is flexible and practical, which are 
confirmed by the results of the reconstructing experiments. 
1. INTRODUCTION 
1.1 3D Reconstruction 
3D reconstruction is referred as a process of the recovery of 
three-dimensional model from the two-dimensional digital 
camera images. During the image formation process of the 
camera, explicit 3D information about the scene or objects in 
the scene is lost. Therefore, 3D model or depth information has 
to be inferred implicitly from the 2D intensity images. The key 
of the traditional method lies in the matching of the 
corresponding features in the images. When there is no feature 
or lack of feature in the objects, or when the features cannot be 
matched correctly at all, the main problems appear in the 
process of the 3D reconstruction 
The paper provides a method to resolve the problems above. It 
is an ordinary slide projector that can supply any feature what 
you want to the target objects. These features are easy to be 
controlled and to be extracted out, which has paved the 
convenient path for the matching of them. According to the 
demand and real condition, the features can be changed or 
adjusted. 
The paper explains clearly how to use the ordinary slide 
projector and how to apply it to the applications of the 3D 
reconstruction, too. 
1.2 Definition of the Solid of Rotation 
In the real world, there exist many things being the solid of 
rotation or being component of several solids of rotation, such 
as: Ball, Cup, flowerpot, jug, pencil and cap, even dome. So the 
solid of rotation is a common thing and it is necessary to 
research its 3D reconstruction. 
  
General definition of the solid of rotation is that the main body 
of the target objects can be formed by a curve rotating around a 
fixed axis. 
The paper mainly researches the 3D reconstruction of the main 
body. Namely, the target object is only a simple solid of 
rotation and it has no other component with it. This kind of 
object has an obvious characteristic that each vertical section of 
the fixed axis is a circle. 
1.3 Overview of the Paper 
In the last few years the interest in three-dimensional object 
reconstruction from two-dimensional images has dramatically 
increased in the field of photogrammetry. The main difficulties 
lie in the texture feature extraction and the model acquisition 
with time-consuming and expensive process, especially the 
reconstructed object lacking in texture. An approach is 
proposed in this paper, which avoids most of the problems 
mentioned above. The target object needed reconstructing is 
given emphasis on the simple solid of rotation. 
When three positions of the camera, the slide projector and the 
rotating platform are adjusted to the suitable place, the solid of 
rotation is put on the centre of the rotating platform. The 
projector projects a feature slide onto the solid of rotation so 
that the texture is supplied to the surface of it. According to the 
different demand and condition, the feature slide can be 
designed differently. Then the digital camera takes sequential 
images with projected feature slide when the platform rotates 
by an angle continually. 
* jgzhang@supresoft.com.en; Tel: 86-27-87561116; Fax: 86-27-87561011 
162 
   
  
  
  
  
  
  
  
   
  
   
  
   
  
   
  
  
  
  
  
  
  
  
  
  
   
   
  
   
    
  
  
  
  
   
   
   
   
  
  
  
  
  
  
  
  
  
   
   
  
   
    
  
  
  
  
  
  
  
   
  
  
   
   
  
  
  
International 
For each spac 
solid of rotatic 
One is an ir 
another is a p 
method, the 2 
can be worke 
projected is d 
point within tl 
the 3D coord 
solid of rotatic 
The sequentia 
solid of rotatic 
and the inhere 
the 3D coordi 
surface of the 
The 3D mo 
connecting al 
reconstruction 
2. "M 
2.4 Backgro 
2.1.1 Thec 
The digital ca 
the late comp 
requires the i 
camera. There 
important prec 
Direct Linear 
used in close-i 
initial value o 
existed camer 
The restricting 
out using the c 
The decompo 
extrinsic para 
camera calib 
addressed. S 
accomplished 
of the digital « 
Zhang, and Y« 
If the camere 
measuring-car 
the intrinsic p 
the known dat 
camera are Stil 
212. Thec 
In this appro 
makes the sar 
slide projecto 
intrinsic and 
computation p 
The algorithm 
and collinear 
algorithm is a 
First, the im: 
carefully and 
computed by 
parameters of 
  
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.