4-3-1
^librai.
;,n 8 this
was
hiev abl e b<
•y the
n f 0n to °btai n
' drained
SBandi
ty the
time of
L ' is of
a zoom
camera
0n > the date for
: long.
Iev >ces Limited
'< Ms Daoxiang
ie experiments.
strabati Journal
p49-59.
sit Techniques,
ation without a
rammeti) and
-34.
heodolite-CCD
y and Remote
oects of
V
METHOD FOR ACCURATE CAMERA ORIENTATION FOR AUTOMOBILE PHOTOGRAMMETRIC SYSTEM.
V.A. Knyaz, S.Yu. Zheltov
State Research Institute of Aviation System
125319, 7, Victorenko str, Moscow
Russia
knyaz@fenix.niias.msk.su
KEY WORDS: Close-range photogrammetry, Camera orientation, Calibration.
ABSTRACT
Progress in close-range digital imaging over the recent years provides the basis for creating mobile mapping systems for spatial data
acquisition purposes. Applying close-range photogrammetric technique in system based on automobile for road mapping, road surface
control etc. looks rather attractive but has some problems. On the one side to obtain required precision in measurements one has to use
lenses with long focal length for suitable image scale. On the other side the possible basis of the stereosystem is restricted by vehicle
dimensions. This results in bad condition for relative orientation parameters determination and therefore in low precision of spatial
measurements.
The paper describes the method for accurate automated cameras relative orientation for photogrammetric system based on automobile
developed in State Research Institute of Aviation system. The method is based on bundle adjustment procedure, using special test scene
and additional information about reference points of the test scene. It uses original coded targets as reference points markers for
automatic and precise reference points determination. It allows to calculate image coordinate of reference points with subpixel accuracy,
using elliptic approximation for target image. The correspondence problem is solved automatically due to target code recognition.
The developed method was applied for cameras positioning on laboratory vehicle used for road scenes acquisition. Acquired stereopairs
was used for objects feature extraction such as road digital model and surrounding objects reconstruction. The description of hardware
of the mobile photogrammetric system and software for system calibration is presented. The results of cameras orientation are given in
comparison with standard relative orientation procedure. The results of system application, estimations of system precision along with
approaches to further accuracy enhancement are presented.
otogrammteric
p. 69-14.
1 INTRODUCTION
Fast progress in GIS-technologies creates the growing demand for
3D models of urban objects such a buildings, bridges, roads etc.
for filling appropriate layers of databases. Digital
photogrammetry approves to be an adequate technique for precise
and productive acquiring necessary information. It becomes
background for creating mobile mapping system based on various
platforms such as planes, balloons, automobiles [Klemm, 1997].
For automobile mapping photogrammetric system the developer
has to find compromise between needed image scale determined
by focal length of lenses used and requirements to stereosystem
precision depending on its configuration (basis of stereosystem,
convergence).
developed for accuracy enhancement are described along with
accuracy estimation and results of system application.
2 MOBILE PHOTOGRAMMETRIC SYSTEM
Mobile photogrammetric system is developed for acquiring
primary data for roads and surrounding areas GIS layers. It is
aimed at the solution of the next primary tasks:
• stereo images and stereo sequences registration of
environment data to create and update database of roads
• capturing data for urban cadaster and management land areas.
The system is based on ZIL-131 vehicle (Figure. 1.) which has
special systems for working in hard conditions such as bad roads
and winter.
Below an overview of automobile mapping photogrammetric
system created in State Research Institute of Aviation system is
presented. System is designed for primary data collecting about
roads and surrounding area. At the present stage of development
it is equipped only with photogrammetric subsystem, external
position system such as GPS being installed later. So the problem
of vehicle positioning has to be solved only by photogrammetric
methods. The methods for system calibration and orientation
The surveying system includes two CCD cameras, two PC
equipped with framegraber and special software developed for
capturing stereo images and stereo sequences. CCD cameras is
installed in the vehicle bumper, basis being 1.62 m. Zoomed
image of CCD is shown in right upper comer of Figure 1.
For capturing left and right images simultaneously cameras are
synchronized each other, left camera working in the internal
synchronization mode and its HD (horizontal drive) and VD