ISPRS Commission III, Vol.34, Part 3A ,,Photogrammetric Computer Vision“, Graz, 2002
ASSESSMENT OF TWO CHEAP CLOSE-RANGE
FEATURE EXTRACTION SYSTEMS
a
Ahmed Elaksher , Mohammed Elchezali, Ashraf Saved, and Yasser Eimanaditli”
a all ; ; ; ;
School of Civil Engineering, Purdue University, West Lafayette, IN 47906-1284, USA elaksher@ecn purdue.edu
b ud :
Faculty of Engineering, Cairo University, Giza, Egypt
ABSTRACT
The use of non-metric cameras in photogrammetric applications is considered under very strict constraints due to their instability and
lack of fiducial coordinate system. Architectural building documentation, monuments registration, and monitoring structure
deformations are very essential close-range photogrammetric applications that require high accuracy and quick data acquisition.
Using metric cameras, in such situations, is quite uneconomic and non-metric cameras are in favor. In order to accelerate the
processing time of analyzing non-metric cameras digital techniques are preferred.
The aim of this research is to investigate the use of two inexpensive techniques for object reconstruction using digital images
produced by non-metric cameras. The first technique employs an inexpensive 35mm camera and a cheap scanner, while a low-cost
digital camera is used in the second technique. Both techniques are thoroughly evaluated and the RMS errors are investigated.
Results show that the 6-paramter transformation model is the best model to handle geometric errors introduced by scanners. The
object reconstruction process results show that sub millimeter accuracy, in object coordinates, can be achieved if systematic errors
are considered.
1. INTRODUCTION
The basic task of many photogrammetric systems is to derive
object space coordinates from 2D images. Analog, semi-
analytical, and analytical techniques have been employed for
a long period of time in photogrammetry to extract ground
coordinates of objects from hardcopy images. In recent years,
digital techniques are implemented in photogrammetric
applications. The advantages of using digital techniques are:
the ease and speed of data acquisition, the inherent on-line
and real-time capabilities, and the high degree of automation.
Acquiring digital images is done either by scanning hardcopy
images or by capturing the photographs directly in digital
format using digital sensors.
The aim of this paper is to investigate the process of
capturing ground features digitally through one of the
following schemes: scanning hardcopy images produced by
non-metric 35mm cameras or acquiring digital images
directly using non-metric digital cameras. Each technique is
evaluated and analyzed using a number of mathematical
models that relate image space coordinates with ground
space coordinates. The implemented mathematical models
were adapted to handle the systematic errors produced by
non-metric cameras.
2. PREVIOUS WORK
In (Boron, 1996) the accuracy of the UMAX 1200 scanner is
investigated; the correction method he proposed reduces the
scanning errors from +5 pixels to £0.15 pixels. The
correction is executed in two stages. First order corrections
are found for each point in the scanner plate first then the
second order corrections for each run are determined. In
(Bolte et. al, 1996) both the geometric and radiometric
properties of the scanners were studied. The RM -1 scanner
was used and it was found that its accuracy is equivalent to
the analytical plotter.
In (Karras and Mavrommati, 2001) the effects of the radial
distortions in the 35mm cameras is studied. A number of
approaches, ranging from the utilization of linear features to
the rectification of regular grids, were used. It was shown
that ignoring the radial lens distortion increases the RMS
errors dramatically. In (Cruz et. al, 2000) the inner
orientation of non-metric cameras was investigated. The
35mm camera images were scanned at 600 dpi and 1200 dpi.
A comparison between the 6-parmeters and 4parameters
coordinate transformation models showed that the former
transformation model is better than the later.
In (Seedahmed and Schenk, 1998) a bundle adjustment with
self-calibration scheme is presented for calibrating a high
accuracy CCD digital camera. An extended version of the
collinearity equations was implemented with corrections for
the symmetric distortion, the decentering distortion, the
image plane unflatness, and the in-plane image distortion.
The results showed the necessity to correct systematic errors.
In (Zolfaghari and Malian, 2000) non-metric cameras are
used to record architectural and historical buildings. The
work shows the effectiveness of using non-metric cameras
for capturing this type of features.
Section 2 presents the calibration process of flatbed scanners.
Section 3 summarize the mathematical models used to
transfer image coordinates to ground coordinates. The object
reconstruction process is presented in section 4. Conclusions
are discussed in sections.
2. GEOMETRIC CALIBRATION OF FLATBED
SCANNERS
During the scanning process, the positions of the scanned
features are corrupted causing the distances between them to
change. For cheap scanners, the distortions increase due to
the bad functioning of the mechanical, optical, and electronic