My name is Greg crutsinger and I was the
volunteer data analyst for the public safety drone teams recently on the Carr
fire I also helped on the Tubbs fire in Santa Rosa last October as part of this
digital first responder effort we used 360 drone panoramics as a rapid tool to
visualize fire damage for the city of Redding and for surrounding Shasta
County now the teams did an amazing job during this situation in the first 48
hours we went from an initial drone flight in the air to maps going public
to the residence of the evacuated neighborhoods that we were working in
While I think that 48 hours is fantastic I think we can do better
40 hours or four hours or even four minutes and I think that we can scale
this from Redding to Rio to Riyadh and here's how those of us in the drone
industry know that drones are really all about the data
but drone data is overwhelming in these situation for example we captured about
a hundred pianos and each of these panoramics was made of 23 photos this is
about 12 gigs of data to upload to the cloud for those of you with slow
internet at home you know that this can be near impossible under normal
circumstances let alone an emergency so during the car fire I thought in wracked
my brain that there must be a better way and I think I came up with one it's a
method that's so simple I overlooked it for two fires now and I think many of us
in the drone industry have overlooked it for the past couple of years but knew
all about it it's cartography with drone telemetry so
let me show you how here's a log of a drone flight you can see that there's
lots of messy information about where I flew but by segmenting the flight log
into where I start and stop my video camera on a dream I now have a very
clean line essentially it turns the drone into a three-dimensional
highlighter in the sky pretend you're out there with the drone
team and reading so here's what it looked like then here's what it looks
like now while you're on the ground the team can simply launch a drone point the
camera straight down and manually fly a trace over critical infrastructure
trying to keep that in the center of the blyve field of view you can fly for
about 25 minutes per battery so you can get several linear kilometres of tracing
per battery this could be traces of burn lines where the fire came through it
could be bulldozer lines that were pushed through by the firefighters
it could be evacuation routes that you need immediately really it's any line
that you want to draw floodwaters as they rise up you could fly it repeatedly
across time I can further segment least tracer lines by altitude remember the
drone flies in 3d so when I fly at a hundred meters over a road that means
the road is open it's clear if I drop down to 50 meters and continue to fly
that means that segment of the road is damaged or I can fly different flight
patterns straight lines over intact power lines zigzags over damaged power
lines where I can fly shapes a square would be for a Supply Drop now remember
in addition to video triggering we also have photos when I snap a photo with the
drone a pin can be dropped if the pins that I care about not the photos
initially these pins could be dropped over all the burn homes in a
neighborhood and we would know within a single flight every burned home in a
neighborhood just by dropping pins you get the picture it's like Morse code in
the sky so you might be asking me okay I get it what's the point the crucial
point here is that all of this information is stored in very small one
or two megabyte text or CSV box the whole map of an entire burn city could
be in a couple of megabytes the equivalent file size of one half of one
drone photo and data the logs can then be rapidly process into KML files or
other types of files that can be exported to matte box ArcGIS or Google
Earth essentially we are turning the flight logs into a standard language
that most mapping software that's used by public
agencies in the city and town you live in during an emergency would understand
the telemetry can then be synced in real-time remotely from any drone fine
with a phone or tablet that can be get to just the most basic mobile service in
the drone industry we are super excited about data 3d models machine learning
I've been a reality passenger drones even but this method it's not that at
all it's really going back to the basics and recognizing for a good chunk of the
world and in emergencies right here in California life-saving information may
come down to a single megabyte text file from a $500 drone
flown manually over a disaster zone so here's the call to action if you work or
know someone in emergency services if you're a government agency or an NGO
know that this method exists and can be used today and improve upon it take it
and make it your own if you're interested in training go to scholar
farms comm slash fire and enter your name and email and sign up for the
newsletter as I release trainings on these different topics for the drone
software companies a tiny amount of developer work on existing drone apps
that work with low-cost DJI drones will let anyone map a rural village today or
your destroyed city tomorrow without internet or imagery if required but know
this will probably not make you very much money finally let me just say that
my heart goes out to those who have lost their homes and lives both in the car
fire and the recent fires in Greece in the next emergency when I get called out
to the scene I'll be using the same method for data analytics and I want you
to know that you came – my name is Greg crutsinger you can find me at scholar
Florence comm thank you