International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004
3.2 The Extensive Model
In order to provide better visual and analysis, MOE purchased
the parcel layers form the Department of Lands and Surveys
(DLS) with scale of 1:10000 and also, street layers with scale of
1:50000 from RJGC. This will allow building detailed
geospatial information around the school building in terms of:
available land parcels to purchase, populated areas, buildings
blocks, city blocks, municipality land use and zones, parks,
streets with all categories, public places such as mosques and
churches, commercial building such as shopping centers and
Banks...etc. It is important also to add population data to all
land use and city polygon features. Noting that, in Jordan, the
new population data based on the latest senses data will be
available in 2004, instead of projected and sample data
currently used. Furthermore, the statistical educational database
will be imported from the EMIS for all previous years, which
will be available by the end of year 2004. In order to achieve
seamless data, one has to resolve the associated problems with
these newly added geospatial data. The three major problems
are explained in the following subsections.
3.2.1 Projections Problems: The major problem emerged
as the land parcel data from the DLS are in the old projection
system that inherited lots of distortions based on the Cassini-
Soldner projection using Clark 1880 ellipsoid. The
transformation parameters (using 3D conformal) between this
system and the UTM system (WGS 84 ellipsoid) are well
defined. On the other hand all other vector data are based on the
newly standard Jordan Transverse Mercator projection using the
European Ellipsoid 19950, and also has known transformation
parameters with respect to the UTM system. The direct
transformation between the JTM and Cassini systems has
distortions ranges from 3 meters to 25 meters. The DLS worked
out some solutions to minimize that by dividing Jordan into
zones of 5 by 5 Km polygons, defining for each zone its own
transformation parameters. In any case, distortions are inherited
in the transformation of the geospatialdata from Cassini to JTM.
The Key solution for this problem from our prospective
however, is our required accuracy for the resultant merged
maps, which is within a map scale of 1:25000. Ikonos Geo
Imagery can be used provide a base for re-adjusting shifted
layer to be within the acceptable precision as resulting from our
preliminary testing.
3.3.2 Scale Problems: As mentioned earlier we have data
with various scales, e.g. 1:2500000, 1:50000, and 1:10000. The
easiest procedure followed is that when working with map on
the country level or the directorate level with map scale over
1:100000 for analysis that required illustrating or using large
scale map format. However, this is not applicable for small
scale maps as all data has to be seamlessly integrated or
adjusted, what is referred to interoperability and standards by
Heipke 2004. The key solution for this is based on the accuracy
requirements for these maps, as it doesn't have to be better than
5-10 m, which is equivalent to map scale of 1:25000, as
suggested by Fraser 2003. The role of Geolmagery to resolve
this problem proofed to be much faster than just speculating
which is true or accurate or not. Geolmagery acts as the base
template when merging these data and deciding which to keep
and/or ignore.
3.2.2 Land Parcels DLS-Key Problem: Each parcel has a
unique DLS key number, which theoretically can be used to
identify the parcel of interest. The main usage for our
applications is to build from these parcels the following:
e. Owned and rented schools parcels layers
e Street layers
e Public property parcels classified according to its
usage
e Scctors, blocks, city-blocks, sub-districts, districts,
directorates, villages, and governorates boundaries
built based on parcels boundaries.
However, there are lots of discrepancy in the built DLS-key,
which gave false search/identification results. One major
problem is the conflict between the new numbering scheme
followed by DLS the old records of the villages, blocks, and
sectors respectively. Also, some street parcels are falsely
assigned a DLS-key, instead of a null value.
Geolmagery helped a lot in this, as it also provide visual
information about the type of feature and validate the search.
3.3 Geolmagery Solution
The new technology of high-resolution satellite imagery has
demonstrated its metric potential for mapping using either
Ikonos or Quickbird imaging systems with potential map scales
as large as 1:10000, Fraser 2002 and Fraser et. all. 2001. The
author/s has/have the opportunity to use both in addition to
SPOT 10 m resolution. This is still in its early testing and
experimenting stage, doing feasibility study to justify the
investment of purchasing such imagery for major parts in
Jordan. Three types of imagery were tested, the Geo reference
imagery, Geo ortho kit with RPC file imagery for Amman area,
and a stereo Ikonos sample. Accuracy assessments were based
on GPS control points distribution in Amman and the vector
formats. The 2-D affine and polynomial functions models were
used for the geo reference imagery and the affine model given
in RPC terms were used for the ortho kit imagery, Fraser 2002
and Kaichang et. all. 2003. Software used for processing
imagery is the ENVI and SOCET Set software. GIS softwares
used are Geomedia and ArcGIS.
Figure (2):
Example in
using Ikonos
Imagery to
validate
streets
layers built
from parcels
vector layer.
The accuracy results showed that geo reference images of either
Ikonos is better than 5 meter, which is very suitable for our
application to assess in resolving the encountered problems.
Many literatures proved that running cost-benefits are
beneficial compared to savings in time and effort (Baccardo et.
all. 2003).
Figure (3X
using Ikonos
Imagery to
validate the
streets vector
layer, the
shifts in
street range
from 5 to 15
meters.
266
Example of
Intern
direct
43 C
school:
density
