The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
1312
• The step of image processing isn’t time consuming
• The large covered area with each satellite image will
cause the elimination of block adjustment and mosaic
production
• Needing to less control point and having orbital
information that can be used as exterior orientation
parameter or primary values
• With access to high resolution satellite images, these
images can be replaced or complemented for aerial
images for mapping
• The facility of obtaining stereo images with using
tilted mirror
Of course beside these advantages, satellite images also have
some defects;
• Receiving stereo images with regards to problems
such as existence of clouds or snow and sometime
overlapped image are related to different season that
cause problems in 3D model production and object
extraction
• Resolution of satellite images and their information
contents are less than resolution of large scale aerial
images and their information contents
The second problem maybe is most problematic issue in
satellite imagery. This cause satellite images not to be replaced
by aerial images but used as complement for aerial images.
Although, it is maybe possible to ignore this problem because
of the capability of receiving repetitious and periodic satellite
images, fast image processing, large covered area with each
satellite image and the advantages of access to far and restricted
areas (Holland and Marshal, 2006). With regard to the
advantages of satellite images and also access to the IRS-P5
stereo images, in this paper, geometric accuracy and
information contents of these images related to the objects
containing in 1:25000 scale maps are evaluated.
2. IRS-P5 STEREO IMAGES FOR REVISION OF MAPS
AT 1:25000
2.1. IRS_P5 mission
IRS-P5 is a spacecraft of ISRO (Indian Space Research
Organization), Bangalore, India. The objectives of the IRS-P5
mission are directed at geo-engineering (mapping) applications,
calling for high-resolution panchromatic imagery with high
pointing accuracies. The spacecraft features two high-resolution
panchromatic cameras that may be used for in-flight stereo
imaging. Hence, IRS-P5 is also referred to as CartoSat-1. In this
mission, the high resolution of the data (2.5 GSD) is being
traded at the expense of multispectral capability and smaller
area coverage, with swath width of 30 km. The data products
are intended to be used in DTM (Digital Terrain Model)/DEM
(Digital Elevation Model) generation in such applications as
cadastral mapping and updating, land use as well as other GIS
applications. PAN-F (Panchromatic Forward-pointing Camera),
featuring a fixed forward tilt of 26*PAN-(Panchromatic Aft-
pointing Camera), it is fixed at an aft tilt of -5*. Each camera
provides spectral range of 0.5 - 0.85 pm, a spatial resolution of
2.5 m, a swath width of 30 km, and data quantization of 10 bits.
Stereo imagery is acquired with a small time difference (about
50 s) due the forward and backward look angles of the two
cameras. The major change in imaging conditions during this
time period is due to rotation of Earth. An algorithm for Earth
rotation compensation is being used to eliminate the delayed
observations of the two cameras (1RS user's Hand book).
2.2. The project of 1:25000 scale maps production
The primary study of the project of 1:25000 scale maps
production started in 1987. Production of these maps started in
1991 and National Cartography Center could produce % 90 of
these maps until July 2007. As mentioned before, because of
flying problems other of areas related to frontiers.
Contour
Interval
Accuracy
Accuracy
Resolution
in height
0.3xC.I.
Planimetrie
(in 0.3 mm)
Planimetrie
(in 0.1 mm)
10m
3 m
7.5 m
2.5 m
Table 1. Features of 1:25000 scale maps
3. PRACTICAL EVALUATION
3.1 The area of case study
For IRS P5 stereo images evaluation, the stereo images of
Shiraz are selected. For geometric accuracy evaluation, 113 full
points are extracted from 1:2000 scale maps. With regard to the
geometric accuracy of 1:2000 scale maps, it is necessary to
mention that geometric accuracy of the full points is better than
60 cm (contour interval of 1:2000 scale maps is 2 meter).
In header file of IRS P5 stereo images, we have 90 RPC
(Rational Polynomial Coefficients) that these RPC calculated
based on time and position of imagery (ERDAS IMAGINE 9.1
Online documentation). By using these RPC, geometric
correction without any control points is possible. But by using a
number of control points, we find out significant improvement
in result. Access to these RPC in ERDAS IMAGINE 9.1 is
possible that we used this software for evaluation of 1RS P5
stereo images geometric correction. We used Geomatica V8.1
for geometric correction evaluation by using polynomial
equations and rational functions model without using RPC.
Figure 2. Distribution of control and check point in IRS_p5
image of Shiraz
3.2 The mathematical models using in this investigation
A rational function model and a 2D Polynomial equation model
applied on the images used in this project. Brief descriptions of
these models are given in the followings: