Anonymous User
|
Posted: January 23, 2010 00:34 by Anonymous User
|
|
Hi Bruce, I have been meaning to get a fire model based system going on worldwind, mainly for community interaction rather than fire fighter personnel support - which in Australia actually means the same thing since fire fighters are drawn as volunteers from the community. I work for a remote sensing company - http://apogee.com.au/blog/. Our main interest is in supporting the fire prediction and management system with elevation data and fuel load information. I am an active worldwind developer and can help out on the visualization section of CPS. We have had an internal project called FireEye for a while - http://www.apogee.com.au/product-nextimage-fireye.php. We extensively use and contribute to open source to bring best value to our clients. We implemented the basic calculations for rate of spread in McArthur Model: http://www.firebreak.com.au/mcarthur_meter.html in Python and can easily translate to Java. For the modelling to be impressive the fire needs to spread from the ignition point using the available fuel,terrain,wind, humidity information. The cellular automata based spread linked below can help. Java based cellular automata to model forest fires: http://student.agh.edu.pl/~artiko/ForestFire/download.html I was considering implementing a Java based spread model on the Google AppEngine, since it consumes a fair amount of CPU resources and cellular automata based tasks are easily parallelized. So large simulation tasks can be run, even long term threat analysis done. I look forward to co-operating on CPS and bringing it to Australia. Cheers, Tisham (whatnick) http://whatnicklife.blogspot.com |
CPS In Australia
3 posts
Replies: 2 - Last Post: February 01, 2010 16:39
by: Anonymous User
by: Anonymous User
showing 1 - 3 of 3
Bruce Schubert
|
Posted: January 25, 2010 15:53 by Bruce Schubert
|
|
Hi Tisham, Hey, I know you from the WorldWind forums. Welcome. Yes, I look forward to working with you also. I can certainly use your help on the visualization, as that is not my specialty. My initial efforts are not focused on modeling the fire-spread, but instead they are focused on data mining of the model outputs in conjunction with CPS fire signatures. My hope is the knowledge discoveries from the data mining will identify where and when significant changes in fire behavior will occur (potential trigger points and fire runs). Thus, it appears there's the possibility of a synergistic relationship without too much overlap. For the US, I'm planning on leveraging the LANDFIRE data for fuel loading - http://www.landfire.gov/. How does that compare with the remote sensing data from your company? Standard GIS raster data products? What geographic regions are you supporting? Meanwhile, I'll checkout the links you provided. Now, if I can just get this iteration milestone completed!! Best regards, -- Bruce Emxsys.com |
|
Anonymous User
|
Posted: February 01, 2010 16:39 by Anonymous User
|
|
We can continue the discussion after the project has moved. I am building some system diagrams for spread models. I will look at JMonkeyEngine for vis work (real flames). Keep up the good work. Our data is mostly rasters - no national repository exists in Austrlia to the best of my knowledge. It will need to be pieced together from Landsat/Modis/RapidEye etc and field work. In your opinion what level of detail is needed for the fuel maps ? what recency ? 3 months - 6months - years ? The elevation data is from interferometric radar 5m posting. Cheers, Tisham. |
Replies: 2 - Last Post: February 01, 2010 16:39
by: Anonymous User
by: Anonymous User
