×

You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Title
Close-range imaging, long-range vision


DESIGN AND IMPLEMENTATION OF A PHOTOGRAMMETRIC SYSTEM F OR
CLOSE RANGE INDUSTRIAL APPLICATIONS
B. Ergün’, M. O. Altan
ITU, Civil Engineering Faculty, 80626 Maslak Istanbul, Turkey — (bergun@ins.itu.edu.tr)
KEY WORDS: Close Range, Photogrammetry, Multi Sensor, Calibration, Application, Expert System, Machine Vision, Industry
ABSTRACT:
As for any photogrammetric application, the accuracy of the derived object data is mostly dependent on the accuracy of the camera
calibration. For many photogrammetric applications, use of a simple model for lens distortion in conjunction with co-linearity
equations is sufficient. However, the combination of very close ranges and the large distortions typically associated with the lenses
used on line scan cameras requires an extended lens model to account for variation of distortion within the object space. The fidelity
of the calibration model becomes particularly important where strin
surface measurement.
1. INTRODUCTION
Digital photogrammetric systems have been used to solve
various measurement problems in industrial applications for
many years, ever since high-resolution CCD cameras and
powerful computer technologies have been available. In close
range photogrammetric applications, working conditions of the
industrial platforms are usually difficult. In these conditions, a
surveying engineer has to find the best solution for the problem
especially in the experimental area. Industrial line-scan video
cameras have been widely adapted for close range
photogrammetry and machine vision applications. Although the
advantages of onboard storage of digital images, such as quick
data storage and portability, industrial cameras are replacing
small format sensor on cameras with possibility of the huge
image scales and PCI cards requirements.
For the expert systems, hardware and the software have been
worked together. This is the fundamental algorithm of the
expert systems. In this study, the equipments of expert system
have been tested with CDP 50 (Linear Variable Differential
Transducer) in an experiment of reinforced concrete slab
deformation.
Firstly, Industrial Basler A302fs camera twin and 16 mm
Cosmicar Pentax fixed focused lenses have been calibrated on
the test field. Afterwards, the experimental data capturing
system have been designed in order to capture the images at the
same time when the actuator press onto the reinforced concrete
slab. In the experiment time 19 stereo images have been
captured periodically. The first image pairs have been captured
when the actuator was zero force and the exterior orientation
parameters have been calculated. For deformation analysis
camera stations have been fixed full time in experiment in order
to use same orientation parameters for other image pairs.
2. EXPERIMENTAL ALGORITHM
Experimental process in deformation analysis has been designed
for civil engineering close range applications in this study. For
an application in industrial photogrammetry, twin industrial
cameras with IEEE 1394 standard, which are the essential part
of an expert system has been calibrated with 16 mm fix focused
lenses from Pentax on a test field. In order to determine the
deformations of a reinforced concrete slab loaded stepwise by
an actuator in the Earthquake Laboratory, these cameras are
located in the object diagonal with stereoscopic view of the slab
gent tolerances are set, for example construction experiments or
surface. The main orientation of the photogrammetric
extraction has been obtained from the first image set of the
experiment steps. The configuration of the signal points have
been designed which coordinate differences give us the
deformation directly. ^ Afterwards for orientation process,
coordinate differences between suitable signal points, which is
decided during the test, have been measured with a compass
with 0.001 mm precision. These differences have been used for
condition equations in the Bundle Block Adjustment. At the
end of the bundle block adjustment, the derived exterior
orientation parameters have been obtained with 0.01 mm
accuracy for the principle point and rotations have been
obtained with 0.001 radian accuracy. Afterwards the system has
been used for determining geometric properties of the slab. The
3D surface model of the slab has been obtained precisely and
the deformations of the slab have been documented. The
deformations on the slab have been calculated with
approximately 9 micron Rms in the photogrammetric restitution
for signal points in image scale. At last these test shows that
industrial expert systems can be used in such industrial
applications with success. At the end, the results of
photogrammetric and TML CDP 50 have been compared.
. Camera Calibration
Station Design
Data Capturing
Photogrammetric Process
Compare the Displacements
USD

Figure 1: Experiment Area
zIS.