PHOTOGRAMMETRIC SYSTEM CALIBRATION
Pasi Laurila, Lic.Tech.
Hannu Salmenperä, Assoc. Prof.
Tampere University of Technology
Department of Civil Engineering
Geodesy and Photogrammetry
P.O. Box 600
SF-33101 Tampere
Finland
Commission IV, Working Group 4
KEY WORDS: Accuracy, Large Scale Mapping, System Calibration, Test Field, Training.
ABSTRACT
System calibration means a process where the main purpose is to establish the overall accuracy of survey results. It is also
important to analyse those factors which are critical for the accuracy. System calibration integrates instrument calibration and
testing of survey methods, programs and personnel in one process. A photogrammetric survey for mapping or for digital terrain
model is a complex process, where the human factor, the stereo-operator, is in key position. Also the properties of ground objects
are quite often such that there are problems to identify and to locate them. Photogrammetric system calibration makes it possible to
handle these kinds of problems, which are beyond easy control under operational conditions. This report describes the experiences
of photogrammetric system calibration obtained in Tampere University of Technology, Finland. The reference used is the
calibrated stereomodel of Rusko, which is an accurately surveyed test field for large-scale photogrammetric mapping. It gives us
the means to establish the accuracy of photogrammetrically mapped coordinates in various point classes, to evaluate the capabilities
of stereo-operators and to train them. It has also been possible to find out the critical parts of the photogrammetric survey process.
They are the resolving power of the photography, the image deformations due to heat in the stereoplotter, the stereo-operator's
skills, and the characteristics of ground objects. Accordingly, the training of stereo-operators seems to be an important way to
improve mapping accuracy. Other important things are improvement of the resolving power of the camera/film combination,
image motion compensation, the careful handling of photographs throughout the process, and the control of image deformations in
the stereoplotter. In the future use of digital images will improve the survey process, for example, in respect of image
deformations.
instrument calibration and theoretical considerations only.
Here we propose a concept and methods of system calibration
The design of civil engineering projects requires data about to establish the accuracy of the final mapping products, and to
the terrain surface in the form of plans, maps, and digital analyse those factors which are critical for the accuracy.
elevation and terrain models. As planning methods are getting
more sophisticated and accurate than ever before, this sets new
requirements for the survey data. Especially geometric
accuracy must be higher than before. Today most of this survey
data is generated by photogrammetric methods. Modern Definition
cameras are accurate instruments, as are analytical and digital
stereoplotters, too. Their accuracy can be tested by instrument — Traditionally calibration is understood as instrument
calibration. But they are only part of the photogrammetric calibration which establishes the accuracy of the surveying
survey process. The rest involves for example the execution of ^ instruments. That is necessary but not enough for complicated
the flight missions, the photographic properties and survey processes. What is needed is system calibration. It
deformations of the films, the personal capabilities of the means a process where the purpose is to establish the overall
INTRODUCTION
PHOTOGRAMMETRIC SYSTEM CALIBRATION
stereo-operators, the accuracy of the control points, and the
properties and interpretability of terrain details. In many cases
the accuracy of stereophotogrammetry in mapping applications
has been overestimated. One can see several reasons for this.
First, the accuracy figures are mainly related to targeted or
otherwise well-defined points. Second, the personal
capabilities of the stereo-operators have not been considered
properly. And third, the difficulties to identify and to locate
terrain points and other features have not been taken into
account. These are matters which cannot be handled with
accuracy of the survey results. It integrates instrument
calibration and testing of survey programs, methods and
personnel in one process which follows the steps of the real
survey process. This is a very demanding task, especially
when one wishes to analyse the results thoroughly. In fact it is
not always possible to find out all the reasons and connections
which have an effect on the accuracy from the results of system
calibration alone. In this respect the conventional methods of
instrument calibration are important, too. The above definition
involves more than the well-known photogrammetric test field
486
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996
calil
the
accu
vari
neec
expe
(Nis
Cali
Testi
The s
whic]
or sh
perso
Espex
of sy:
to te
them.
The c
of th
contri
mark:
Opera
the c«
rando
mark:
meas