The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
374 
• Acquisition of Approximate Values 
+ Approximate values for trajectory of platform are calculated 
by post processing using GPS and IMU. 
+ Approximate values for exterior orientation parameters are 
given from direct-georeferencing system. 
+ Nominal values for interior orientation parameters such as 
focal length, principal points and so on are adopted as initial 
values. 
+ Coarse point clouds 3D data are generated using the 
trajectory values and exterior orientation parameters. 
• Camera Calibration 
+ Bundle adjustment with self-calibration is performed using 
coarse point clouds 3D data. 
+ Trajectory of platform is recalculated by GPS/IMU and 
calibration results. 
• Generating DSM 
+ All point clouds data are corrected using recalculated 
trajectory of platform. 
+ DSM is generated. 
3. CAMERA CALIBRATION 
The authors have been concentrating on developing a close 
range measurement system for consumer grade digital cameras 
using triplet images (Chikatsu & et al., 2006). The 
measurement system was adopted into digital aerial 
photogrammetry in this paper since triplet images have 
following characteristics. 
+ Triplet images have advantages in generating stereo pairs. 
+ Triplet images have flexibility for multiple images. 
+ Triplet images have ability to increase geometric restriction. 
On the other hand, lens distortion is the most important 
parameter among interior orientation parameters, and many 
distortion models were proposed (Fryer and Brown, 1986). In 
order to apply consumer grade digital cameras to digital 
photogrammetry, it was concluded that lens distortion was 
corrected sufficiently using radial polynomial 5th degree model 
(Chikatsu, 2007). Therefore, radial polynomial 5th degree was 
adopted in this paper, and accuracy performances for the 
simultaneous adjustment were investigated regarding following 
calibrations models. 
3.1 Calibration using Pseudo GCPs 
Pseudo GCP is interesting points on the image and Pseudo 
GCPs have following advantage. 
- It doesn’t require for Pseudo GCPs as initial value to have 
accurate coordinate. 
- Pseudo GCPs can be extracted as a feature point on the image. 
-3D coordinates for Pseudo GCPs are computed from point 
clouds data. 
Therefore, camera calibration using pseudo GCPs was 
considered as primary stage for the simultaneous adjustment. 
Figure 1 shows concept of generating pseudo GCP, and X,Y,Z 
coordinate for the Pseudo GCPs are computed by following 
procedures. 
+ Interesting points such as edge of road paint are extracted as a 
feature point on the image, and the image coordinates are 
acquired. 
+ Initial ground coordinates of the interesting points are 
calculated using the image coordinates, initial interior 
parameters and exterior orientation parameters which are 
obtained from direct-georeferencing. 
+ Point clouds data (X, Y, Z) around the interesting point are 
transformed to image coordinates (u, v) by collinearity 
equation (1). 
+ Positional relationship between the pseudo GCP and the point 
clouds on the image are computed using each image 
coordinate. 
+ Heights of the interesting points (the pseudo GCPs) were 
computed from weighted interpolation equation (2). 
+ Two-dimensional coordinates of the interesting points (the 
pseudo GCPs) were computed from equation (3). 
Figure 1. Concept of pseudo GCP 
f a n { 
0 
1 
M, 
(r-r.) 
[■z-z,) 
Chx 1 
f a 2 } 
(*-*„] 
k-*o) 
,+ a 32 l 
]+ a 2 2 ( 
fr-r.) 
(Y-r,) 
,+ £72, 
0 o 
N N 
1 1 
N N 
aj 
* 
1 
* 
© 
Er-r.) 
l+ aJ 
N 
1 
N 
O 
where u, v: image coordinates 
/: focal length 
ay: rotation matrix 
X, Y, Z: ground coordinates of the point clouds 
X 0 , Y 0 , Z 0 : exposure station