.svg)
In this blog post, we aim to provide you an unbiased opinion of how drone search and rescue works. It seems like the internet has two groups of people: those who think drones are useless in search and those who think every problem can be solved with the latest technology i.e. drones. The reality is more in the middle. There are certain situations/environments where drones really excel in and there are some environments where drones face a lot of obstacles.
We’ve done some research on search and rescue forums and here list out what people are saying about it.
There are two ways drones are helpful in these environments: 1) it minimizes the safety risk rescuers face just to check an area that might be clear. 2) It minimizes the time it takes to check an area safely. Sure, you can get your trained crew down on a cliff but it will take valuable time to get rescuers down there safely just to check the area.
.png)
.png)
Although easy to scan, large grassy areas take a long time to scan. As noted by a Reddit user:
“Last year we were searching in tall grass, it would have taken a team a couple of hours to clear a section since we had to walk no more than 6 feet apart from each other. A drone could do it in minutes. Also one of the drones spotted a cougar in the area, something I would rather not do as a ground searcher.” - sergei1980
As a rule of thumb, the lighter the subject’s clothing is, the easier it is to spot the subject with a thermal camera. Clothing such as shorts and t-shirts, often worn during the summers, can be identified by thermal cameras more easily.
For winter clothing, it depends what type of winter gear the subject is wearing. Tight clothing is very visible in thermal cameras while loose clothing, such as ponchos, are much harder to see.
Another factor is time the subject has been out in the cold. After a certain time, the winter clothing could become the same temperature as the outside temperature.
We might unwittingly think of drone scanning as being equally as comprehensive as scanning an area by hand with boots on the ground but we are mistaken. Think of scanning an area with a drone as ~70% as comprehensive as scanning that same area with boots on the ground. Many times that ~70% quality of a scan is good enough to find the missing subject, but sometimes it is not. Your team could find itself in a situation where it calls an area as clear because the drone footage made it look that way but in reality, you are getting a false negative because of the lower quality scan that happened via the drone.
Does that mean you shouldn’t bother with aerial searches and just do everything manually? Absolutely not. We simply want you to think of a drone survey as a percentage of the quality you’d get with boots on the ground over the same area.
Some examples: A drone scan over a large grassy field might be 10x faster than scanning that area on foot and the quality of a drone scan could perhaps be thought as 90% as good as surveying by hand. I would feel comfortable calling this area as initially as clear and move on to the next. Had the area been a wooded forest, I’d be very apprehensive in calling this area as clear from the drone footage alone as the quality of the drone footage is far from what we’d get from boots on the ground. The drone footage in this situation could be thought of as more of a backup or a preliminary footage just to see if we get lucky but shouldn’t be thought of as a way to clear an area.
This is our opinion, not a fact, of different search and rescue environments and how good a drone scan is relative to a manual scan.
DJI leads the way with recommended drones with the Matrice 30T and Mavic Thermal being some of the most popular models. Something to note is that DJI is a Chinese company and if your funding comes from a US government agency, you could have some restrictions as to where you’re allowed to buy from. These regulations change all the time and vary by agency, so double check before buying if you’re using government funding to make your drone SAR purchase. A much more expensive American alternative is the Skydio X10.
There are several software tools out there that take your drone footage and analyze it to help you identify abnormalities you might have missed yourself. Most of these software tools take the footage and look for color disparities in the footage pixel by pixel. In other words, they look for colors (like a red t-shirt belonging to a subject) that don’t belong to the overall background.
The general consensus is that the tools work OK. Its effectiveness is heavily dependent on the color scheme of the environment against what colors the subject is wearing. A green grassy environment vs. a subject wearing a red t-shirt is ideal for this software while a subject wearing a black jacket in a wooded forest might not be as easy to spot by the software.
The two biggest tools are ADIAT from TexSAR, which is a free open-source tool, and Loc8. The latter is paid but has better reviews and has more features than TexSAR. We recommend you start out with ADIAT from TexSAR since it’s free and consider upgrading to Loc8 from there.
