International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
Another imaginative application to make use of object-
relational features is real time tracking of deer location.
Concerned parties can be notified when deer are crossing
borders or other areas of interest. The total national park area or
the upper area 1100 m above sea level can be entered into a
database table:
create table AREA. OF. INTEREST (
area id NUMBER (2),
area name VARCHAR2 (50),
border MDSYS.SDO GEOMETRY
Deer migration can be followed — based on GPS collars - by
tracking its locations at a fixed interval (e.g. every few seconds),
and in real-time inserting the records into another table:
create table DEER. LOCATIONS (
deer id NUMBER (3),
time TIMESTAMP (2),
location MDSYS.SDO_GEOMETRY
y
In this table time is stored using the SQL type TIMESTAMP,
which is a high-precision time and date type that allows to store
fractions of a second (Qian, 2004). Whenever a GPS receiver
updates its location the current timestamp and location will
automatically be sent out and inserted into the above table:
insert into DEER_LOCATIONS values (3,
to timestamp (17-APR-2004 19:17:21.00^,
‘dd-mon-yyyy hh24:mi:ss.ff”),
mdsys.sdo_geometry (2001, 82032,
mdsys.sdo_point_type (4607260.47, 5425484.18, null),
null, null)
Y
The query inserts a new record for the location of deer with id 3
at the specified timestamp. It is possible to insert thousands of
such records in one second which makes it quite feasible for
real time tracking of numerous deer.
Now a database trigger can be implemented that is sparked off
whenever a new record is inserted into a specific table. In the
trigger body it can be checked if deer is still within the national
park resp. within another area of interest. This operation can be
achieved using the relate function. The query from the trigger
body to determine if deer is still within the park is:
select sdo_geom.relate (a.location, ‘ANYINTERACT",
b.border, 0.5)
from deer. locations a, area. of interest b
where a.deer id = 3 and b.area. name like 'nationalpark? and
a.time ^ (select max (time) from deer. locations where deer. id ^
3%
The result of the query is "TRUE' if the deer is still within
park's borders or ‘FALSE’ otherwise. The subquery returns the
most current timestamp for the deer with id 3. If a ‘FALSE
value is returned the trigger can e.g. open an HTTP connection
and send a message to a Web service outside the database with
the deer id, the location and the time the deer is crossing the
border. The Web service can then act accordingly, for example
sending out emails to the concerned institution or creating
dynamic maps with the current deer location.
Let us assume that the deer perambulate special areas once a
year we can create user-defined object types that represent such
events as spatio-temporal zones:
create or replace type ST_ZONE_TY as object (
name VARCHARQ2 (50),
from DATE,
tll DATE,
zone MDSYS.SDO_GEOMETRY):
A method can be added to the object type that validates if deer
location and stopover falls inside such a zone. By this means
histories and changes to real world objects can be captured and
managed inside the database.
4. RESUME AND CONCLUSION
Object-relational technology in DBMS helps to model and
analyze spatial-temporal events as it provides highly flexible
means for storing, manipulating and validating diverse and
complex data structures. As for applications in Bavarian Forest
National Park the real challenge is to understand the
relationships of the most complex processes at all - those of
nature - and find definite and describable entities that can be
part of a database model. Thus it is often useful to extend the
concept of objects from a conventional point of view to a more
general sense, so to speak to describe real world scenes (instead
of objects) that show on their part some correlation with a set of
commonly used objects. Object views leave existing (relational)
data models unchanged but can centrally merge miscellaneous
information of any database model and serve for making use of
object-relational features in spatio-temporal applications.
References from Books:
Loney, K., 2003. ORACLE?9i: Die umfassende Referenz.
Carl Hanser Verlag, München Wien, ISBN: 3-446-22170-0
References from other Literature:
Bauer, M., 2002. Walddynamik nach Borkenkäferbefall in den
Hochlagen des Bayerischen Waldes: Dissertation at Lehrstuhl
für Waldbau und Forsteinrichtung, Technische Universität
München
Geringer, D., 2003. Oracle Spatial Best Practices: An Oracle
Technical White Paper, Dezember 2003
Lee, G., 2003. Simple Strategies for Complex Data: Oracle 9i
Object-Relational Technology: An Oracle Technical White
Paper, October 2003
Nationalpark Bayerischer Wald, Bayerische Staatsforstverwal-
tung, 2001.
und Borkenkäferbefall: Academic series book 14, July 2001
Plabst, S., 2001. Entwicklung eines objektrelationalen
Datenmodells für ein kulturhistorisches | Geoinformations-
system: Thesis at Fachgebiet | Geoinformationssysteme,
Technische Universität München
Qian, L., 2004. Oracle Database 10g -- Developing Spatial
Applications Using Oracle Spatial and Map Viewer: An Oracle
Technical White Paper, February 2004
References from websites:
General Oracle technical resources: http://otn.oracle.com
Oracle Spatial resources: http://otn.oracle.com/products/spatial/
Waldentwicklung im Bergwald nach Windwurf
AI
KEY\
ABST
Path fi
Infrast
algorit
centerl
networ
networ
essenti
is sugg
solely
feature
finding
set of d
1.1 Gq
With t
(GIS) t
GIS en
applicat
Geogra|
have be
(1996),
(1999),
(2000).
The mc
network
correspc
segment
arcs so
centers
applicat
form a «
node-arc
traverse
network
Howeve|
ground.
photogra
kind of d
common
