7 
3. Historical Account of Field Calibration 
3.1. ONE-DIMENSIONAL TEST OBJECTS 
The classic method of field calibration is based on photographing an array 
of distant targets. There have to be at least three. At the point defined by 
the exterior projection center one measures the angles to the targets. Their 
image co-ordinates on the photograph are also measured and a resection is done 
to determine the interior orientation. This method has been used, among others 
by E. S. Sewell [24]. These calibrations are based on measurements in the pic 
ture diagonals. The two diagonals are photographed and the calibrations are 
computed independently. The procedure is not very close to operational con 
ditions. The location of the exterior projection center must be assumed to be 
known. This is a great disadvantage for close-up cameras, which require targets 
at close distances. See chapter 7. 
3.2. TWO-DIMENSIONAL TEST OBJECTS 
Hubeny [19] has described a method to determine principal point and camera 
constant of close-up cameras. A glass scale with accurate millimeter divisions 
is photographed in two positions. The scale is parallel to the image plane and 
the two positions are obtained by translating the scale a certain distance per 
pendicular to itself. This is controlled by autocollimation and necessary for the 
method to be reliable. The image points are few and located in a line. The 
translation in the photography direction makes the test object two-dimensional. 
The procedure can be repeated with the line in different directions on the pic 
ture. 
The so-called grid method by Hallert [11] is very well suited to determine 
the radial distortion and the accuracy of the image co-ordinates of close-up 
cameras under operational conditions. In this method all points of the test ob 
ject are in one plane. To determine the camera constant it is suggested, that 
the test object or the camera be translated an accurately determined distance 
Ah. This causes a variation in the image scale which is used to determine the 
camera constant. The grid method has been applied to many imaging systems, 
e.g. roentgenography, see Hollender [18]. 
The radial distortion of aerial cameras can be controlled under operational 
conditions by the so-called ice-surface-method developed by Ekelund [8] or