DIGITISING TABLE CALIBRATION USING CLOSE RANGE 
PHOTOGRAMMETRY 
Kevin Jones and Basil Pathe 
School of Surveying 
Queensland University of Technology 
BRISBANE QUEENSLAND 4000 
Commission V 
ABSTRACT 
Throughout the surveying and mapping industry there is a proliferation of digitising tables being used to acquire data for Geographic Information 
Systems. In many work situations the digitising tables receive only a cursory calibration, if they receive one at all. Users generally rely on GIS 
system software to make the appropriate corrections to the data when fitting to control. To obtain the most reliable and accurate information, 
however, all phases of a system should be calibrated. The paper describes a method of calibration using close range photogrammetry via the 
bundle reduction technique and discusses the results of tests carried out. 
KEY WORDS: 
1. INTRODUCTION 
The question of digitiser calibration arose while the authors were 
developing a data acquisition system using manually digitised aerial 
photographs. As the method depended on digitised coordinates, it 
was essential that the accuracy of the equipment was verified. Of 
course, this is in keeping with the sound and practical expectations 
that every piece of equipment used for precise scientific measurement 
must be calibrated to provide full information on the accuracy 
attributes of the data. 
At the time, the School of Surveying had just acquired a Rollei 6006 
metric camera. As all the tools, including analytical plotter and 
software, were conveniently available, it was decided to use close 
range photogrammetry to calibrate, or verify, the calibration of the 
digitiser being used. 
This paper is a description of the project and its outcome. 
2.  DIGITISER CALIBRATION 
The digitiser being used was an ARISTOGRID table digitiser MODEL 
GRT108 with a measuring area of 915mm x 1220mm, and a quoted 
resolution/repeatability of 0.025mm (1/1000 inch). The measuring 
process uses the principle of absolute coordinates. The sensor can 
be lowered or raised from the digitising surface at any time without 
affecting measurement accuracy. We had never experienced any 
problems with this digitiser and in fact it had a ‘feel’ of reliability and 
accuracy. There were, however, other digitisers within the University 
which had exhibited a gradual creep in ‘size’ over a number of years. 
The procedure proposed was to manually digitise a reference grid, 
use close range photogrammetry to coordinate the reference grid, 
and then relate the digitised coordinates to the photogrammetric 
coordinates. At this stage the reference grid was more of a 
convenience as a set of targets and transfer medium than as a 
calibration grid. The anticipation was that the photogrammetric 
coordinates would agree with the reference grid coordinates. 
3. REFERENCE GRID 
A reference grid on film was purchased from a Queensland 
Government Department. It was a 1cm grid with an overall size of 1m 
x 1m which did not exactly coincide with the digitiser measuring area. 
It was decided to carry out the calibration of the digitiser at 100mm 
intervals even though it extended beyond the recommended 
measuring area along one side; a total of 121 points. The 100mm 
grid intersections were high-lighted with marking pen to avoid 
misidentification. 
217 
Calibration, Cartographic, Close-range, GIS/LIS, Photogrammetry 
4.  DIGITISER COORDINATES OF GRID INTERSECTIONS 
Two sets of digitised coordinates of the grid intersections were 
acquired for this project. For convenience they were labelled Dig 7 
and Dig 10. A description of the data sets follows below. 
4.1 Dig 7 Data Set 
The XY coordinates of the data set are the arithmetic mean of 7 
sequential sets of measurements over the grid. That is, after a grid 
intersection was measured the cursor was moved to the next 
intersection, and then the next, until all grid intersections were 
measured. Then the whole grid was measured again, for a total of 7 
times. The arithmetic mean of each grid intersection was determined, 
giving one set of data. 
The precision/repeatability of the measurements was 0.06mm (one 
sigma). 
4.2 Dig 10 Data Set 
The XY coordinates of this data set are the means of 10 consecutive 
measurements over each intersection of the grid. That is, each grid 
intersection was measured 10 times, with the cursor being moved off 
the intersection after each individual measurement. When 10 
measurements had been made on a particular grid intersection, the 
cursor was moved to the next intersection, and the procedure 
repeated. The whole grid was measured once. 
The precision/repeatability of the measurements was 0.04mm (one 
sigma). 
5. PHOTOGRAMMETRIC COORDINATES OF GRID 
INTERSECTIONS 
51 Outline 
The coordinates of the grid intersections were determined 
photogrammetrically by the bundle method. The reference grid was 
fixed to the digitising table with tape and three one metre long scales 
helped keep it flat as well as providing control for the project. The 
scales were positioned so as not to obscure any of the grid 
intersections. 
5.2 Photography 
Photography was taken with a Rollei 6006 metric camera with a 50mm 
lens using black and white film. Lighting was normal room lighting 
consisting of standard fluorescent lights. Reflections obscuring the