699 
A NEW TEXTURE MAPPING ALGORITHM FOR PHOTOREALISTIC 
RECONSTRUCTION OF 3D OBJECTS 
T. Hanusch 
Institute of Geodesy and Photogrammetry, Swiss Federal Institute of Technology (ETH) Zurich, ETH-Hoenggerberg, 
CH-8093, Zurich, Switzerland - thomas.hanusch@geod.baug.ethz.ch 
Commission V, WG V/4 
KEY WORDS: Texture, Photo-Realism, Colour, Rendering, Visualization, Resolution, Cultural Heritage, Virtual Reality 
ABSTRACT: 
In the last decade, new developments in remote sensing, photogrammetry and laser scanning generated new possibilities for 
acquisition, processing and presentation of different objects at various scales. Beside object modelling, photorealistic texturing is 
still a challenging task. This paper focuses on a new texture mapping technique, which extends the known 2.5D approaches of 
texture mapping to 3D objects. A new algorithm will be described, suitable to achieve high quality results by eliminating problems 
of current approaches e.g. resolution dependency, need of huge amount of memory and handling of colour and brightness 
differences. A new, vector based visibility analysis, using vector algebra instead of pixel based algorithms will be presented. In the 
context of texture mapping, the fusion of different texture sources (different images) with different illumination and exposure 
conditions as well as images with different resolutions is an important task. To achieve best results in terms of homogeneity and 
level of detail, a multi-image texturing approach was developed, to preserve the fine texture information of the object. To adjust the 
brightness levels of the image dataset, a global and a local colour adjustment was developed. The software “Blender” was used to 
generate high quality images, movies and animations. The result of the investigations and developments is a software system to 
texture 3D models photorealistically and automatically. All algorithms are discussed in detail. The performance is demonstrated 
using artificial and real world data sets. 
1. INTRODUCTION 
Textured models are established as a proper possibility to 
present the results of photogrammetric data acquisition and 
processing to experts as well as to non-experts. The 
expectations of the customers are increasing according to the 
development of the hardware. It started from simple 
visualizations of small 2.5D areas over increased datasets with 
artificial texture e.g. colour coded surface models to 
photorealistic textured 2.5D models. Now, the step towards 3D 
objects with millions of points, photorealistic texture and real 
time visualizations is in progress. This paper describes a whole 
workflow from the oriented images and the geometry 
information over the visibility analysis and the global and local 
brightness correction of the images up to the final visualization 
using open source software. Therefore, a number of algorithms 
were evaluated, modified, developed and combined. 
For texture mapping of 3D objects a visibility analysis has to be 
conducted. The approaches developed for 2.5D surfaces or for 
true-orthophoto generation (Rau, 2002), (Amhar, 1998), (Zhou, 
2003), (Biasion, 2004) have to be adapted to 3D data. The 
existing 3D algorithms were basically developed for computer 
sciences applications and have some disadvantages, e.g. high 
expectations concerning image overlaps, high number of 
images and short image acquisition times to avoid changes of 
the illumination conditions. These requirements are mostly 
problematic to fulfil, especially for photogrammetric data sets. 
Therefore we developed a new visibility algorithm. 
recognizes grey value difference of less than one percent, a 
colour adjustment is necessary to achieve satisfying results 
during multi image texture mapping. Different algorithms were 
presented in (Agathos, 2003) and (Bannai, 2004). In this paper a 
solution, combining three different processing steps, is 
presented. 
Finally, open source software was used to generate high 
resolution images, virtual reality models and movies of the 
processed data. 
2. DATA ACQUISTION 
Two datasets are used to show the performance of the 
developed algorithms. 
• A Khmer head, carved out of sandstone in the 12th or 
early 13th century, with a height of about 30cm. The 
geometry as well as the images were acquired in the 
Museum Riedberg, Zurich, Switzerland (Akca, 2007). 
• A part of the St. Gallen Globus, manufactured around 
1595 in St. Gallen, Switzerland, with a height of 
around 230cm, made of wood, fully painted with the 
earth and sky map of Mercator and diverse other 
ornaments. The used part is a support of the sphere 
with a size of 70 x 25 square centimetres. The dataset 
consists of images acquired using a standard still 
video camera. 
For the triangle to image assignment, existing approaches were 
slightly modified and adapted (Niem, 1995), (Rocchini, 1999). 
According to the high sensitivity of the human eye, which 
Dataset 1 was used to show the whole workflow beginning from 
raw-data to the final textured model. Dataset 2 was used to 
show the brightness correction algorithm in detail.