Title: Folie 1
1Automated Wildland Fire Detection with AWFS
Wildland Fire Detection Workshop FERIC Hinton,
Alberta June 6-8, 2006 Presented by Joachim
Dreibach Fire Watch AG
2Automated Wildland Fire Detection with AWFS
- Introduction Methods of detection
- Automated detection of smoke
- Parameters
- How it Works
- System Overview
- Extended and Additional Applications
- Summary
3Automated Wildland Fire Detection with AWFS
- 1. Methods of Detection
- Lookout Towers with human resources
- - Reliability, human factors ( thickness, sleep,
concentration.. ) - Aircraft surveillance
- - expensive, time frame between trials
- Video-based systems
- - handling, human factors , technology
- Satellite
- - very expensive, time factors
4Automated Wildland Fire Detection with AWFS
- Automated Smoke Detection Parameters
- The early and precisely detection of wildfire
with an automated - system is a complex task.
- It far exceeds the use of a simple video camera
which is taking - pictures and
- maybe a correlated standard image processing SW
is running. - Automated detection requires a reliable, precise
recognition of - Events caused by a fire.
- .
5Automated Wildland Fire Detection with AWFS
- Automated Smoke Detection Parameters
- As smoke is one of the first visible sign for a
starting fire, - it is an preferred parameter for the automated
detection. - Automated detection must largely eliminate
factors in the - observed areas, not correlated to a smoke cloud,
- Automated detection shall submit additional
information - like coordinates of fire, distance, information
about type - and size.
6Automated Wildland Fire Detection with AWFS
- Disturbing factors for automatic detection
- Moving Objects like
- Cars, Animals, People, Aircrafts, Clouds,
Shadows, Air turbulences, - Fixed Objects, but changing by influences like
- Trees - moving by wind, Sun reflections on
objects, - These factors demonstrate that an automated smoke
detection requires - more than a commonly used image processing
software to guarantee an - acceptable number of false alerts
7Automated Wildland Fire Detection with AWFS
- AWFS, How it works
- This SW algorithm is based on a research program
of the GFMC(UN) - The method and system are patented
- by DLR (EP 0984413)
- AWFS has been successfully applied for
- 5 years with more than 100 installations
- Detects smoke fast and efficiently up
- to a distance of 40 km (depending on sight
conditions). - Scanning Time 360 5-8min
- AWFS has the capability for day and night
detection.
Survived lightning strike 2006
8Automated Wildland Fire Detection with AWFS
- AWFS, how it works
- To provide highest quality AWFS, is based on a
Sensor - used for space missions and software algorithm
developed within a EU - research program for forest fires.
- The method involves an optical detector, rotating
platform, an image - processing unit and a transmitter for local
alerts. - Searching smoke in a spectrum around 620 nm using
a filter with a small - bandwidth (to increase contrast ratios and cut
off other colors not related - to smoke).
- For the exclusion of the disturbing parameters,
AWFS evaluates smoke - parameters by a series of independent features
like - Shape, Brightness Dynamics, Expansion, S
tructure - This reduces the numbers of false alerts but
ensures safe detection.
9AWFS, How It Works
- Each image is analyzed pixel by
- pixel to identify the portions of
- images corresponding to the smoke
- criteria of shape, contrast, dynamics,
- expansion and brightness.
- The portions identified on three
- images are then compared in order to
- analyze any move in the targeted
- portions.
- If the system construes it as smoke
- emanating from a fire, it sends an
- alarm signal to the supervision/control
- system.
10How It Works - Sensor Operations
11- The long way to develop a Smoke Detection
System. -
- Headed by the German Aerospace Center (DLR)
ground, satellite and aircraft based systems for
fire monitoring have been studied since 1992 - Development and tests of aircraft based payload
ABAS (IR and optical line cameras, onboard - processing)
- Development and operating the microsatellite
(BIRD) for fire recognition (launch in 2001 - Development and tests of main components of the
tower based FIRE-WATCH technology - 1996 initial contacts with German forest
authorities - 1997 first tests with different sensors
(optical and IR camera - development of the detection software
- 1997 EU tender by the forest authorities of
the state of Brandenburg - for a pilot project was released
- 1998 successful demonstration of FIRE-WATCH,
tender was awarded to DLR - 1999 - 2000 pilot project (parallel with
conventional monitoring) - 2001 - 2002 FIRE-WATCH became fully operational
- 2002 after a selection process of the entire
technology IQ wireless GmbH, - with a license for the German marked only,
was installing 120 Systems in Germany
12System Overview
13Key Parts of AWFS
Mobile Applications North Finder GPS Unit
Sensor Unit
Pan and tilt head with special sensor
CameraTower 1
IPU Image Processing Unit with lightning
protection
CameraTower 2
Telemetry Unit Point to Point, Satellite, ISDN
APU Autonomous Power Unit H2 500 watt fuel cell
14Core Software Modules
- Image Processing Software
- Complex analysis of typical smoke characteristics
(color, brightness, dynamics, structure and
expansion) - Normalization, filtering, matching, cluster
search algorithm and probability assessment - Automatic cloud reduction
- Programmable parameter settings for sensitivity
- Alerts in the event of smoke detection
- Values implemented based on a UN research program
- Control Office
- Displaying of up to 5 towers to one 19 Monitor
(connectable up to 64Sensors) - Communicates the alert information from the tower
- Evaluation of alarm data by special software
tools (zoom, filters, contrast, continuous fire
observations) - Evaluation of transmitted images by an operator
- Computer aided alarm handling
- Archiving of data and actions of operator into a
data base
15Control Office
16Control Office
17Alert
Alert 16km from sensor position
18Live sequence
19Event Monitoring
Continuously monitoring of an event while
continuing to survey area By requesting life
sequence images from selected area of
interest. Optional color images for monitoring an
area of interest and classifying the smoke to
manage the resources
Database
Archiving of all collected data, operation tasks
as well as system status information in databases
for long term documentation and training.
203. Extended and Additional Applications
- Mobile Units
- Comand Center, controlling several Offices
- Forecast fire propagation
- Meteorological data
- Ecological data
- With GPS unit locate ground assets on map
- Vehicles and personnel
- Communicate with Air and Ground Assets
- Send map and ground asset data to aircraft
- Advise Ground assets of impending air drop
21- 4. Summary
- FIRE-WATCH is a highly specialized tower based
system with a CCD-camera and sophisticated
software for the early recognition of forest
fires that has been tested and optimized for more
than 8 years. - FIRE-WATCH technology has been installed in
Germany since 2001 and has fulfilled the
expectations of the clients. In the moment 120
systems are operational. In many regions it has
completely replaced manual surveillance from
lookouts. - At conditions as prescribed in Germany,
FIRE-WATCH is less expensive (savings of 20)
and faster compared with the traditional
monitoring from lookouts by trained staff and
with aircraft monitoring. - Presently the system is equipped with additional
components to make it ready for other countries
with other conditions.
22ContactFire Watch AGChristian Raboud
CMOKehrstrasse 12CH-3904 NatersSwitzerlandPho
ne 41 27 922 0804Fax 41 41 922
0805info_at_fire-watch.ch
Location
Thank you for your attention. Dont hesitate to
request additional information.
23 Typical Installations
24Additional Application Fire Propagation C2I
Center
Implementation of a Fire Propagation
Software Meteorological data from a sensor site
like Temperature Wind speed Wind
direction Humidity Air pressure at ground
level GPS data Time
25Additional Application 3D - C2I Center
Implementation of AWFS Data onto a 3D Digital
map Information about Identification of
Resources Position of Resources Movement of
Resources Strategic Coordination of
Resources Identification of Fire Position
26System Components
27System Components
28System Components
Meteorological Sensor Temperature, Wind speed,
Wind direction Humidity, Precipitation
(Rain) Air pressure at ground level,
Illumination
29Business Partners
Deutsches Zentrum für Luft-und Raumfahrt e.V.
(German Aerospace Center)