/en. 
and the parallel type in 
nal types are adopted 
4). 
€ 4 distances.The cross 
ompared with the box 
condition equations can 
oed from the cross type. 
improvement compared 
ent is that the initial 3D 
ints can be calculated 
ause control points are 
rmed under nearly the 
first case. Unknown 
he following equation, 
+ le(s AY} + a) 
z) + LCS] 
: (3) 
since it was possible to 
ons. Consequently, the 
ere of equal value(=1). 
developed types. 
veloped types 
Z 
mi = 4.05 M 
2.77 
+ 4.85 
+ 2.63 
  
  
  
  
  
early coincide with the 
d from equation 2). 
Id condition, the cross 
ing the control points. 
letic surveying in site, 5 
ht and 4 pipes with a 
pared (Figure 5). 
  
MIT CE ® O 
oleo ooo 
DIC Q 
I e 
00 
IC C 
MO 
(c) F type 
70000 
Et C 
  
  
  
  
enna 1996 
  
  
  
  
  
Figure 5 Control points linked with pipe 
4. DIGITAL PHOTOGRAMMETRIC SYSTEM FOR 
SITE RECORDING 
In order to acquire real-time images, a wireless CCD 
camera(FR CD-5C; 768 x 494) (Figure 6) and wide angle 
lens (f =4mm) were adopted. This wireless CCD camera 
with a transmitter mounted on top can take 
  
  
  
  
  
Figure 6 Wireless CCD camera and BS tuner 
high quality images from 300m away using 1.2 GHz micro 
wave and BS tuner. The weight with lens is 165g and the 
Size is 34.5 x 35.2x 52.0mm. The CD-5C becomes a 
very useful tool when the image is taken from tethered 
balloons, kites or model helicopter. 
For the estimation of CD-5C, camera calibration for CD-5C 
and XC-75 were previously performed by using the same 
‘test site as in the first experiment where stereo image for 
XC-75 was taken from 1.61m height and base-height ratio 
0.12 .Similarly, CD-5C were taken from 0.59m height and 
base-height ratio 0.33. 
Tables 4 and 5 show the calibration results with 9 control 
points and the root mean square error for check points . 
Table 5 RMS E for XC-75 and CD-5C 
  
XY Z 
  
XC-75 | +0.12m (+030) | +1.14 mm (+2.49) 
CD-5C | +0.21mm (+0.34*) | +0.90 mm (+1.03*) 
  
  
  
  
  
where, ( ) means computed values from equation (2) 
*; as pointing accuracy for CD-5C, 0.002 mm 
was adopted. 
Experiment results were expected to show better values 
than the theoretical values because these experiments 
used 9 control points and only circle targets. Results for 
XC-75 are better than theoretical values in expectation. 
Results for CD-5C are nearly equal to theoretical values 
contrary to expectation. The natural reason may be that 
the images using CD-5C were taken with a wide angle 
lens. 
However, it is concluded that a wireless CCD camera is 
still a useful tool for the purpose of real-time image 
acquisition from tethered balloons or kites. 
5. APPLICATION IN THE ARCHAEOLOGICAL 
SITES 
Tests were performed at an actual archaeological sites 
using the cross type, and a stereo image was taken from 
about 9 mheight using a CD-5C wireless camera(Figure 
7). 
Figure 8 shows a stereo image and Figure 9 shows the 
configuration of real-time photogrammetric system used 
in this test. 
Table 6 shows calibration results for a stereo image. 
Table 4 Calibration results with 9 control points 
  
Xo Yo Zo Q) 
K X0 yo (o ai a, — poo 
  
pixel pixel mm 
mm mm mm 
left image 12.482 81.983 1714.864 0° 28’ 09.4" 0° 55’ 22.5” 0° 09' 31.3" 248.597 241.499 16.183 100.638 0.006 1.00 
  
XC-75 
right image | 207.494 80.039 1718.701 -1 34 27.7 1 43 324 0 47 47.6 280.703 274.447 16.267 100.580 -0.002 1.00 
  
mm mm mm 
pixel pixel mm 
left. image 44.746 84.026 588.362 1° 07' 45.5" 2^ 09' 16.3' 1^ 04' 04.5" 285.049 220.798 4.143 131.020 0.322 5.00 
  
CD-5C 
  
  
  
right image | 236.397 82.212 583.328 2 00 548 1 13.127, 0 53 .449 . 281.3380 211.210 . 4121 131.139 0.152 7.00 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B5. Vienna 1996