4.3 Geometric correction
On the image processing system satellite images are registered to the
map image data. This may be done by using a split screen and
displaying the two images in each half. Both images can be zoomed or
roamed and by picking corresponding points in the two images the
coefficients of the tranformation formula may be calculated. The
final processing is done by the Warper of the Model 75 image
processor. |
4.4 Raster/vector convertion
As vectorizing always is performed on binary data, images must be
slized using a lower and upper threshold to produce binary data. Then
the data are converted to the compressed format (run length coding)
used by the scanner (Module IIS2LSC). The data is then vectorized
using standard SysScan software which produces boarder lines from
polygons and lines from linear features. (Module MONITOR)
4.5 Storing of data in Map Information System
1f data from the image processing system contains header information
about transformation to user coordinates, the data may be loaded into
a specific layer or sublayer in an existing vector file. (Module LGSW)
5. APPLICATIONS
The modules described above gives the user the opportunity to
integrate data from both systems. Image-maps with different overlays
from the map system are easily produced. Information from the map
system are used as ground truth when selecting training areas for the
classification process which simplifies and improves the
classification. Such classified areas are then used for updating the
map system.
Below is shown the result from a processing where an image from the
Thematic Mapper has been registered to the layer River in the mapping
system (figure 2). This transformed image has then been classified,
and one of these classes has been used for entering new data into the
map system as shown in figure 3.
- 413 -