Hello, I am student working through an internship related to wildfire management.
I am trying to create an image processing program to detect fire, map fire, and send that data as a KML (google maps polygon) of the fire to a ground station.
I am running into issues of not finding enough images and data to train a robust model.
Is there an archive or some website that I can use to get this data.
Also, my idea is to be able to label the fire characteristic such as height, angle, fire head, flank, etc and predict where the fire will go in various time increments (5, 30, 60, etc minutes).
I did a quick search and found these two articles:
“From public fire dataset and Internet, we collect 7000 images for training and 4494 images for test, and then run experiments with the comparison of four baseline methods including deep neural network, support vector machine based on scale-invariant feature transform feature, stack auto-encoder and deep belief network.”
You’re going to need academic access to get at the original article.
This article directly links to eleven (11) datasets.
The references on these articles should get you into the various academic work that has already been done.
My first thought for ‘original’ content is YouTube (which you would have to scrape, but would get you a lot of aerial photography) and the Alert Wildfire camera network: http://www.alertwildfire.org/ (which would get you a lot of ‘non fire’ images for comparison). Then Instagram, with an honorable mention to https://www.instagram.com/slvsteve/
The JAIC built a process last year using hours and hours of MQ9 flight videos of fires in California. The end result was exactly what you are trying to do minus the identifying the head, flanks etc.
It outputs a KML off the video footage to an end user on the ground. They were using a plugin for ATAK that while playing the live video would annotate your map live.
I will see if some of the video clips of the end product are shareable to the public.
