International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 5. Hakodate 1998
DIGITAL CLOSE RANGE PHOTOGRAMMETRY DATA ACQUISITION SYSTEM
WITH COMBINED GPS GEOREFERENCE AND ITS CALIBRATION
Esmond Mok, Bruce King, Qing Zhu, Zhilin Li, Kent Lam
Department of Land Surveying and Geo-Informatics
The Hong Kong Polytechnic University, HK
Commission V, Working Group 2
KEYWORDS: CCD, Close Range Photogrammetry, GPS, Self-Calibration
ABSTRACT
A low cost digital close range photogrammetric system is presented for the acquisition of 3D surface information for efficient large
scale and civil engineering mapping applications. Two GPS antennas are mounted together with two CCD cameras respectively on a
stable bar and are employed for direct determination of camera position and azimuth during imaging process. This paper discusses
the system calibration through a test field of control points. Some key issues about the integrated system such as the relationship
between the GPS antennas and cameras and the achievable accuracy of the system are examined. The use of GPS geo-referrencing of
the cameras, multi-station convergent imaging and simultaneous adjustment are stressed for improving the accuracy and reliability of
3D surface data acquisition.
1 INTRODUCTION
Since small-format CCD cameras first attracted photogram-
metrists’ attention in the mid eighties these cameras have
evolved high-resolution, large-area CCD cameras with digital
output. The increase in resolution has resulted in a demand for
greater accuracy and precision requirements for geo-referencing
the camera system. In order to reduce the number of necessary
control targets, position and attitude can be determined by
integrating other information such as GPS and INS data and
applying functional constraints. Obtaining complete, accurate,
and fast description of a 3-D object or environment may not be
possible without such integration.
GPS/INS systems has been successfully adopted in aerial
photogrammetry for direct measurement of the exterior
orientation parameters to decrease the amount of ground
controls (Schwarz, 1995). Similarly, more sophisticated systems,
and the methods for their integration have been developed for
digital close range imaging (Li, 1997, Schwarz, 1996). Such
research focuses on the areas of industrial and urban mapping
applications. The system presented here is intended for the
mapping of engineering sites, since civil engineering
construction is a major activity in Hong Kong, and at present
conventional mapping techniques are largely adopted despite its
intrinsic inconvenience and danger on a construction site. It
integrates the positioning technologies of GPS, photogrammetry
and image processing.
Using such an integrated system, the basic problem is just to
determine the relationship between GPS and camera. This
process is referred to as the system calibration. The adopted
approach is that of self-calibration as reviewed by Fraser (1997)
After briefly introducing the system configuration, a simple
self-calibration model and DLT algorithm (Karara, 1989) are
quoted directly. In the experiment analysis section, single-
Stereopair and multi-image processing and the imaging
configuration are addressed.
253
2 SYSTEM CONFIGURATION
2.1 Cameras
Two Apple QuickTake 100 digital cameras are employed.
These cameras take 8 hi-resolution (640 by 480 pixels) 24-bit
color images with a pixel size of 8.75 u m by 8.75 u m. Focus
range is from 1.3m to infinity. The principal distance is
approximately 8mm.
2.2 GPS system
Two Trimble 4600LS GPS receivers are used for camera
positioning and one Trimble 4000SSI GPS receiver is used at
base station for relative GPS positioning.
This configuration of Cameras/tripods/GPS antennas mounted
on a common stable bar is illustrated as in Figure 1.
( GPS )
C camera )
Camera
!
i
GPS ? d
à
camera
GPS/Camera/Mounting bar Configuration
9.
Figure 1 System configuration
3 SYSTEM CALIBRATION
Because it is attempted to directly use the GPS and calibrated
results for camera orientation and location in the future work,
for convenience the self-calibration method is adopted in our
experiment. The following issues need to be considered:
1. The design, location and observation of the controls;