Drones & Firefighting
I wanted to (quickly) share some experiences with the use of drones in firefighting, as my UAS and goals differ a bit from most departments we've seen that are starting to use drones.
I am a career firefighter that recently relocated to a small rural community in Oregon, surrounded by thousands of acres of agriculture. While we do get structure fires, the majority of our fires are wildland, usually involving wheat fields. In many situations, multi-rotors make a lot of sense (see: Branford fire uses DJI Phantom on quarry fire). I also have a multi-rotor that stays near my turnout gear and I've gotten in the habit of grabbing it if I'm heading to a structure fire. Thus far, I have yet to respond to a call where it would be useful. The main reason I started building my own fixed-wing drones is to increase flight time and allow for active and ongoing monitoring of large, non-structure fires. My main goal for this project is to create a resource that is invaluable for planning a fire attack, and is therefore useful enough to justify taking a firefighter "out of the fight" in order to operate the drone throughout the event.
We are still working on our 333 and COA application, which I have been neglecting way more than I should have during this entire 6 months. Because we don't have the authorization to use these during a fire (And I'm not talking about USFS or DOI managed forest fires), I have been very reluctant to use them during an event. Most of my testing is taking place after the fact and during my off time, and very soon we will begin using them during our weekly training drills.
In the mean time, I've been using a few to map and assess the damage a day or two after the event. I have been speaking to NFIRS and firehouse.com about using drone-created imagery for our fire reports and investigations.
I do have one of the fixed wings set up with video and some low quality "night vision" and 2 5.8ghz receivers on the ground, but the quality is underwhelming. Once I am able to get the gimbal tracking dialed in for ROI monitoring in a loiter, I will look into upgrading the live feed quality and hopefully try to find some funding for a thermal camera.
Thank you for looking. If there is some interest in progress and use, I'll try to keep the blog updated.
Wheat fire mapped 2 days later with a Phantom2
NIR and NDVI mapped 1 day later with a Pixhawk powered fixed wing, processed by agribotix.
Comments
Mario,
I guess my post sounded too much like a business entrepreneur Hana
I'm a researcher working on vision based navigation for UAVs and in always looking for civilian or humanitarian applications which pose interesting challenges.
So in you're view the missing piece right note is on the image processing side. I also imagined so.
It would be interesting if UAVs could eventually replace hidrant planes, where I guess there's the greatest risk for humans.
Thanks for your input!
Use by firefighters in Europe has been far far far and I mean far more successful than the Police role
If you're looking at something to market towards the firefighting community for operations, I honestly think everything needed is "already there", except the regulations. Active monitoring and the ability to track crews and progress and transmit that data to anywhere in the fireground, day or night, is very possible and can be done without a human operator for most of the event(not legally, though). Beyond that, most of my ideas involve activities that will likely be FAA restricted for awhile.
If you're asking because you're interested in developing a product for the firefighting community, I would say that you should focus on using drone-collected data to correlate with seasonal trends and current weather data to rapidly and accurately predict fire behavior. This information isn't as useful in what I'm currently doing, where a fire might last for 5 hours, but would be invaluable on large burns that take weeks to get under control.
Large, multi-organization fires is what I have been working towards since I first decided to commit to this idea. I'm happy that there are plenty of attainable goals and profits while working towards that goal, but I would like to see more people on the software side working towards the same purpose. There is a massive gap between hobby-grade drones and platforms like Precision Hawk, both in functionality and price. The functionality side can be filled by some DIY tenacity, but the data processing and management is something I have no ambitions of tackling.
Find a way to reliably pinpoint the area of origin of a fire, to assess the likelihood of ignition from non-human factors, to predict fire behavior and guard the safety of personnel on the ground, and to gauge the regrowth potential or recommend a crop that will thrive and condition an afterburn soil. If you do that, I think you've got a product that you can sell to 4 different agencies all working the same ground. I know I'll buy it, if I'm one of the agencies around when the day comes.
I would like to ask, since you are in (I think) unique position of having both experience with UAVs and firefighting, what do you think would be interesting in terms of applications in this scenario, which would be best solved autonomously instead of requiring a human operator?
Also, here is a quick video of the fire in the 2nd and 3rd pictures that one of our firefighters took as they got to the top of the hill in a brush truck.
There is a road running mostly north-south right in the middle of the burned section. Most of the work on the west side was done from inside the brush trucks and tenders. Most of the work on the east side, especially the darker section in the north east, was done on foot.