See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/255651498

Use of photogrammetry in 3D modeling and visualization of buildings

Article · January 2007

CITATIONS READS

56 6,902

2 authors:

Mesapam Shashi Kamal Jain

National Institute of Technology, Warangal Indian Institute of Technology Roorkee
37 PUBLICATIONS   146 CITATIONS    242 PUBLICATIONS   3,521 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

AI for Earth Observation View project

M.TECH IITK View project

All content following this page was uploaded by Kamal Jain on 11 October 2014.

The user has requested enhancement of the downloaded file.

 VOL. 2, NO. 2, APRIL 2007 ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2007 Asian Research Publishing Network (ARPN). All rights reserved.

www.arpnjournals.com

USE OF PHOTOGRAMMETRY IN 3D MODELING

AND VISUALIZATION OF BUILDINGS
M. Shashi1 and Kamal Jain1
1
Department of Civil Engineering, Indian Institute of Technology Roorkee, Roorkee-247667, India
E-mail: [email protected]

ABSTRACT
3D Models of architectural structures are very important in order to study, analyze and also to reconstruct and
document these structures. Visualization of these models allows the user to get the photo realistic impression of the
structures than graphic based object models. 3D visualization has many applications in the areas of architecture, civil
engineering, tourism etc. Normally, field surveys are more accurate than photogrammetric measurements; nevertheless,
they involve more personnel and consume a lot of time. Photogrammetry is a measurement technology that can be used for
the extraction of 3D points from images. The present paper highlights the project using photogrammetry for accurate 3D
modeling and visualization of structures.

Keywords: 3D modeling, visualization, photogrammetry.

1. INTRODUCTION 2. OPERATIONAL STEPS

In its restricted sense, the term photogrammetry In the frame of this work, the following activities
means the process of measuring images on a photograph. An were performed:
integrated digital Photogrammetry system is defined as
ƒ Photo acquisition;
hardware/software configuration that produces
ƒ Processing of photographs;
photogrammetric products from digital imagery using
ƒ 3D Model generation; and
manual and automatic techniques. Nowadays there is an
ƒ Texturing and Visualization.
increasing demand for full three dimensional data for
planning, architecture, environmental analysis, tourism etc.
2.1 Data acquisition
(Jose Luis Lerma, Antonio Garcia, 2004). To depict the real
The data acquired may be either hard copy
conditions of the object, and to measure 3D of all the
photographs taken with film cameras or digital
corners of the structure particularly the inaccessible points,
photographs taken with a digital camera. If hard copy
creation of accurate 3D models are very much necessary.
photographs are available, these photographs are to be
Recent years have seen a number of development in
scanned with high resolution scanner. Other way is to
geometric modeling of historical monuments in 3D
acquire digital data directly by means of digital CCD
especially their structural details and textures (Mila
cameras. Nowadays high resolution digital cameras are
Koeva, 2004). Close range Photogrammetry is a
available at reasonable prices which bypasses the step of
measurement technology that can be used for the extraction
scanning. The cameras are also classified into metric and
of 3D points from the images; further these points are useful
amateur cameras. Close range digital methods based on
for the accurate 3D modeling and visualization. Digital
images taken with amateur digital cameras are becoming
Photogrammetry derives all the appropriate measurements
more popular because of their economic aspects. As the
from the images itself rather than measurements directly
cost of developing the photographs is null, any number of
from the objects. Due to the digital data flow,
photographs can be taken and decided at the office to
photogrammetry has now become an efficient alternative
choose the best photographs required for the project.
to the classical building measurement and reconstruction
The camera used in the project is Kodak CX 7300
methods (Ulrike Herbig, Peter Waldhäusl., 1997).Sabry
digital camera. Before using the camera in the project, the
El-Hakim et al., 2005, discusses the various 3D model
camera needs to be calibrated, under required
generation techniques and shows the advantages of
photographical conditions. Recovering 3D structure from
photogrammetric techniques over CAD based techniques.
images a simpler problem when the cameras used are
For the present project, an experiment has been conducted
calibrated, that is, the mapping between image coordinates
to check the accuracy of the 3D model generated using the
and directions relative to camera is known. This mapping
photogrammetric technique. The complete orientation and
is determined by, among other parameters, the focal length
calibration is done automatically only based on the
and its pattern of radial distortion (Paul E. Debevec,
images. There were no markers, or any approximations
1996). The calibration of the Kodak CX7300 digital
used. The object selected is the Institute of Engineers
camera was performed using Photomodeler 5 self
building, located in the Indian Institute of Technology
calibration module, using several images of a plane test
Roorkee campus area. In the project, the camera used is
field supplied with the software. The results of calibration
Kodak CX7300 and the processing is done using
are shown in Figure-1.
photogrammetric software package Photomodeler 5.

37
 VOL. 2, NO. 2, APRIL 2007 ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2007 Asian Research Publishing Network (ARPN). All rights reserved.

www.arpnjournals.com

Figure-2. Positions of camera during data acquisition.

2.2 Data processing

The photographs acquired were imported into
Figure-1. Camera information after the calibration. photogrammetric modeling software, Photomodeler.
Reference points were selected on the building like
PhotoModeler performed automatic camera corners which are easily identifiable and separable. The
calibration in 2 stages and the total error after calibration same points were identified on the building in the next
was 0.01 which is less than 0.02 (value indicating very photograph. Similarly all the salient points were identified
good calibration for digital camera). After calibration of the around the building and each point is referenced in all the
camera, it was ready to take photographs. For a ideal photographs. This process is known as orientation. Figure-
photogrammetric survey, the base line/distance ratio ie., 3 shows the reference point selected over the photographs.
the distance between two camera positions while taking
photographs to the distance between the camera to the
object ratio should be within reasonable limits and not be
too small and that should be between 1/15 and 1/20. If the
building façade is large then camera is kept at large
distance from the building (John Badekas, 1975). For best
results, 3X3 rules were given by Waldhausl and Ogleby,
1994, which are to be observed for photography with
amateur cameras.
Photographs are taken for the Institute of
Engineers building and totally 35 photographs were used Figure-3. Orientation of photographs.
for the project taken at different positions around the
building. The positions of the camera were carefully 2.3 3D Model generation
planned for the geometric accuracy of photogrammetric 3D model generation is typically done
reconstruction of the project and to identify all points on interactively to segment the points into separate objects
photographs required for determination of the frame of and also to edit the output. For large environments, since
reference and textural data of the object. Figure-2 shows the technique may require large number of images, model
the positions of camera maintained during data acquisition generation still necessitates significant human interaction,
of the building. regardless of the fact that 3D point coordinates were
computed automatically (Sabry El-Hakim et al., 2003). A
3D model without surfaces or textures is a wireframe
model. The wireframe model represents the building as a
quantity of vertices and edges. This representation is true
if one is interested in the general form of the building.
Figure-4 shows the 3D model of the building.

38
 VOL. 2, NO. 2, APRIL 2007 ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2007 Asian Research Publishing Network (ARPN). All rights reserved.

www.arpnjournals.com

(Virtual Reality Modeling Software) format. Figure-5

shows the 3D textured model of the building.

Figure-4. 3D wireframe model of the building.

2.4 Texturing and visualization Figure-5. Textured 3D model of building.

3D Wireframe model is just the line drawing and
gives the shape of the building. To get the realistic 3. ACCURACY ASSESSMENT OF THE MODEL
impression of the building, textures are added for each The quality of the model is assessed by the
side of the building. Textures give a new quality to a 3D residuals between the image coordinates marked on the
model making it more similar to real object and more images by the user and the image coordinates calculated
adequate for human perception and understanding. using calibration of the cameras. Table-1 shows the RMS
Textures can be assigned automatically by selecting an values of some points in the model which depicts the
image with the best visibility or manually by picking accuracy achieved in the project.
proper face on the desired image (Ildar, 1999). The final
output of textured object 3D model is shown in a VRML

Table-1. Accuracy assessment of some points in the model.

RMS
X Y Z
Id X Y Z Residual
Precision Precision Precision
(pixels)
1 3.1029 1.1521 8.9341 0.0022 0.0015 0.0024 1.3309
3 3.0325 1.1207 8.7921 0.0020 0.0013 0.0022 1.1926
4 3.0289 1.2502 8.7937 0.0019 0.0013 0.0022 1.0779
11 2.9963 1.3990 8.9492 0.0020 0.0013 0.0021 1.4736
12 3.1058 1.3995 9.0993 0.0023 0.0014 0.0023 1.6679

In Table-1, Id is the point identification number REFERENCES

for some points used during referencing, XYZ values are
model coordinates calculated during processing and XYZ [1] Jose Luis Lerma, Antonio Garcia. 2004. 3D city
precision values shows the marking accuracy, and the last modelling and visualization of historic centers, Int
column, the RMS residual value in pixels, achieved in the workshop on vision techniques applied to the
project. As per the guidelines of the photomodeler, rehabilitation of city centres, 25-27 oct, Lisbon,
(Photomodeler Pro 5 Help) for a good photogrammetric Postugal
project, RMS residual value should be less than 3. The
RMS value achieved is less than 1.5 for all the points, [2] Ildar V. Valiev. 1999. 3D Reconstruction of
which is a good value for a project. architectural objects from photos. The 9th
International conference on computer graphics and
4. CONCLUSIONS vision, Moscow, Russia, Aug 26-Sep 1.
The above model demonstrates the use of
[3] Mila Koeva. 2004. 3D realistic modeling and
Photogrammetry in accurate 3D modeling and
visualization of buildings in urban areas, Int sym on
visualization. The precision obtained in the 3D models
corresponds perfectly which is required for any further Modern technologies, technologies, education and
work like reconstruction etc. Due to the availability of professional practice in geodesy and related fields,
digital cameras in the market at reasonable costs, sofia, 04-05 nov
Photogrammetry offers the best alternative technique to [4] Mostafa Madani. 2001. Importance of digital
any other CAD based techniques used. Photogrammetry for a complete GIS. 5th Global
spatial data infrastructure conference, Cartagena,
Columbia, May 21-25.

39
 VOL. 2, NO. 2, APRIL 2007 ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
©2006-2007 Asian Research Publishing Network (ARPN). All rights reserved.

www.arpnjournals.com

[5] Paul E. Debevec, Camillo J. Taylor, Jitendra Malik. visualization of complex architectures (3D-Arch
1996. Modeling and rendering architecture from 2005), August 22-24, Venice-Mestre, Italy.
photographs: A hybrid geometry-and image based
[9] Ulrike Herbig, Peter Waldhäusl. 1997. Architectural
approach. In SIGGRAPH '96.
photogrammetry information system, October 1-3,
[6] Photomodeler Pro 5 Help, EOS Systems Inc. Goteborg, Sweden, ISPRS - International Archives
of Photogrammetry and Remote Sensing, Volume
[7] Sabry El-Hakim et al. 2003. Effective 3D Modeling
XXXII, Part 5C1B
of Heritage sites, 4th Int Conf on 3D Digital
Imaging and Modeling, Banff, Canada, October 6- [10] Waldhausl P, Ogleby C. 1994. 3X3 Rules for
10, 302-309. simple photogrammetric documentation of
architecture. In: J. G. Fryer (Editor), Int. archives of
[8] Sabry El-Hakim et al. 2005. 3D reconstruction of
Photogrammetry and remote sensing, Vol XXX,
complex architectures from multiple data, ISPRS
Part5.
Int. Workshop on 3D virtual reconstruction and

40

View publication stats