„OS 
er gently. On the 
elevation tends to 
vers demonstrated 
ance for artificial 
represent artificial 
ropriate for ALOS 
ited PRISM image 
es by selecting 
ition according to 
from the satellite 
  
RISM 
  
t information from 
I-100 which is an 
by STARLABO 
in 2000 (Chen ef 
'SI-100. 
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Item Specification Remarks 
Nadir | Forward | Backward 
Observation 0.52 to 0.77 um 
Band 
IFOV 3.61 urad Converted distance on the ground 
surface: approx. 2.5 meters at the 
nadir 
FOV 5.8 degrees. 2.63 degrees. Max extraction width 
Focal Length 1939mm 1939mm | 1939mm 
Scan Cycle 0.37 millisecond + 0.004 millisecond 
Pointing Angle | > + 1.5 degrees. | > + 1.36 degrees. 
MTF >02 Including MTF degradation along 
track by spacecraft flight 
S/N 270 
Gain Setting 4 steps It is possible to set per radiometer 
B/H - | 1.0 
AD Bit 8 
Data Rate « 960 Mbps (320 Mbps/telescope) Before compression 
Angle from + 23.8 degrees. (for forward and backward) Along track 
nadir 
Side angle -0.86 degree | -0.68 degree | 0.86 degree Perpendicular to track 
Table 1 PRISM Main Characteristics 
Item Specification Remarks 
Nadir Forward Backward 
Observation Band RGB RGB RGB 
IFOV 0.0067 degrees. Converted distance on the 
ground surface: approx. 10cm 
at the nadir 
FOV 61.5 degrees. Max extraction width 
Focal Length 60mm 
Scan Cycle 0.002 second 
Flight Height 600 m 
AD Bit 12 
Data Rate < 36 Mbps 
Angle from nadir + 21.5 degrees. ( for forward and backward ) Along track 
  
  
  
  
  
Table 2 SI-100 Main Characteristics 
This article deals with the extraction of height information from 
PRISM simulated imagery generated from SI-100 which is an 
airborne three-line-scanner developed by STARLABO 
Corporation jointly with University of Tokyo in 2000 (Chen et 
al., 2003). Table 2 lists the main characters of SI-100. 
As conversional binocular aerial frame images for DEM 
generation, image matching methods are the key technique in 
DEM generation with PRISM imagery. Since there are no 
epipolar lines in PRISM stereo images and one-dimensional 
matching is unfeasible, an improved matching method using 
feature points and grids of PRISM imagery is proposed. The 
method has been developed by STARLABO Corporation 
jointly with University of Tokyo and successfully applied in SI- 
100 photogrammetry system. This paper briefly reviews the 
concept of the method and gives the DEM results with PRISM 
simulated images from SI-100. It also reports the accuracy by 
comparing with aerial images. 
2. DESCRIPTION OF THE APPROACH 
Figure 2 shows the concept of the algorithm for DEM 
generation with PRISM simulated imagery. In the following 
subsections we will describe these steps. 
2.1 Preprocessing of the Image Data 
According to the design specification of PRISM, each line 
sensor of PRISM is composed of several CCD line units 
collecting image segments for one image line in order to satisfy 
the high sampling frequency of image data. Each CCD unit 
generate one image strip. Certain overlap pixels between two 
neighboring CCD units ensure a whole image line for forward, 
nadir and backward view of PRISM. One task of preprocessing 
of PRISM image data is to merge the image strips generated by 
CCD units into a whole strip for PRISM forward, nadir, and 
backward viewing sensor. 
The second task of preprocessing of PRISM image data is to 
extract the high precision orbit and attitude data from the 
recorded trailer data and interpolate these orientation data at