Volunteers scour drone footage in search for woman who went missing on hike | Hacker News Reader

Volunteers scour drone footage in search for woman who went missing on hike

geekwire.com

77 pointscurtis79 comments

In the doc where the findings are documented, someone wrote this for video DJI0015:

    5:09 - 5:24	at 5:09 two vertical rocks pan onto the screen, which appear to
    be part of a small natural shelter, and are visible in the center of the screen
    until 5:24. There appears to be movement in this area/shelter, at first I thought
    it was the light hitting the rock, but the movement looks intentional and the spot
    moves away from the rock. waving motion being made from about 5:13 through 5:24 when
    the camera pans away from the spot.

I looked at it a couple of times now, and indeed it looks like some waving motion. What can this be? I have no idea what size the rocks are. Can this 'shelter' house a human?

I see it, but I think the motion looks more like something swaying in the wind.

I marked it in a screenshot here: https://i.imgur.com/w38bGrR.jpg

link to the scene here: https://youtu.be/gNZa8Ug-e40?t=309

I've replayed the video about a dozen times now, and I see absolutely nothing where you marked - nothing that moves, and nothing that looks like a shelter.

There is a small judder, when the camera pans, which I attribute to a compression artifact.

It's difficult to tell the camera scale from the images, but judging by the tufts of grass (and the balloon earlier in the video) that area is quite small - those rocks are perhaps the height of a human. I would think that a shelter would be very obvious at that scale.

<edit> I see the movement now, after kicking up the resolution on YouTube. The scale is more obvious with hi res as well. I'm definitely calling it a blade of grass.

The marked small dot definitely moves. Maybe just too small to be seen on a phone screen?

I can see it when I full-screened the video on a 1920 x 1080 display.

Yup, clearly visible on full-screen at 2650x1440; in windowed standard YouTube size I couldn't spot it either. Doesn't look like a compression artifact to me (though there is an artificial-looking jump at the exact moment the camera/drone rotates).

looks like a leaf in the wind to me or perhaps a small bird.

I see absolutely nothing

At 5:13-5:17 in the video you can see the thing moving in the the black void area of those rocks, where the parent marked the image. The problem is, the drone proceeds to produce another better top-down view as it moves over top of the thing in question - it has little to no consequential mass (something akin to a fat reed), it's plainly not a person. Full screen at max resolution helps a lot.

I see it too and agree that it looks like the wind. The movement is too fluttery to be human movement. Too much irregular back and forth.

Could see it clearly on 5k 27" monitor. That's something alive, not foliage. Could have been a small animal but it's hard to gauge the scale.

I definitely see something, but at a loss as to what the something is.

But it seems like it shouldn't be that hard to go back to that spot and investigate more?

I'm fascinated by how humans are amazing classifiers. But also so easily fooled, especially in groups, into seeing things that aren't there. We are probably best served being backed up by computer classifiers that may be less susceptible to bias like that.

Not suggesting this instance is a false positive or negative though.

Unfortunately computer classifiers are trained by datasets that are labeled by human vision. There's a theoretical limit to how much better the computers do compared to humans, which is limited by our human vision.

I watched it a couple of times; to me, it looks like an animal grazing. Don't know the geography there, so can't verify if this area has animals living in mountain.

If that's a human I'd really like to know why they would pick that spot, because the only reason I can think of is for hiding...

Readable quote for mobile:

> 5:09 - 5:24: at 5:09 two vertical rocks pan onto the screen, which appear to be part of a small natural shelter, and are visible in the center of the screen until 5:24. There appears to be movement in this area/shelter, at first I thought it was the light hitting the rock, but the movement looks intentional and the spot moves away from the rock. waving motion being made from about 5:13 through 5:24 when the camera pans away from the spot.

This reminds me of a similar search, maybe 10 years ago, for a tech employee and his family.

His car and his family were eventually found (from inspection by volunteers of aerial footage), on an old, abandoned logging road. Unfortunately, he himself had walked away from the car, seeking help, and did not survive.

Was it James Kim?

https://en.m.wikipedia.org/wiki/James_Kim

Oh, what, I grew up watching him on TechTV. I did not know he'd passed, let alone in this kind of way.

Ah jesus, that is terrible. He believed he was 4 miles away from a town called Galice. How could this be?

I have gotten lost on several poorly-maintained roads in my time in the Canadian rockies. To think this could happen to me and the people I'm with makes me feel ridiculous and stupid.

What should a group of people do if they found themselves in such a situation?

Stay warm and hydrated?

I keep one of these (an older version) with me when backpacking/venturing somewhere in the wilderness without cellular coverage: https://www.amazon.com/Garmin-Lightweight-Handheld-Satellite...

Me too. Definitely gives me a good feeling that if things go really bad I can call for help. Hope I will never use it.

It's a really good idea to use it at least a few times in different conditions; you don't want the first time to be an actual emergency. I'm also a big fan of solo backpacking and carry one on trips.

I get 10 messages free each month so I often use some of them. It's the "SOS" button I don't want to use ever.

Or this if you don't want to pay for a subscription and/or only want to use it for search-and-rescue needs: https://www.amazon.com/ACR-PLB-375-ResQLink-Buoyant-Personal...

There's some interesting history behind distress beacons that broadcast on 406 MHz without any subscription. It's a good example of successful Cold War cooperation. https://en.wikipedia.org/wiki/International_Cospas-Sarsat_Pr...

Always stay with the vehicle. He set out to try to find help, and disorientation and delirium set in with hypothermia.

Kim did stay with the vehicle (and his family) for something like a week before going for help. I think this might have been the right decision at that point, because his youngest child was 7 months old and would not have been able to survive without food nearly as long as an adult would.

Kim's wife and children did stay with the car and were rescued some time later. But at the point that Kim left the car he had to be thinking that searchers weren't going to find them in time.

I've often wondered if Kim's fatal mistake wasn't leaving for help but rather leaving the road he was on and going cross country. By the time he made the decision to do that, his thinking might have been very seriously impaired indeed, though.

Interestingly he walked 16mi and died just 1mi away from their intended destination.

Can't imagine the horror of trekking through the wilderness, your family waiting in the car, to find help.

That's good advice if people are looking for you, or you are in an area where other people are likely to pass by and find you accidentally.

But what if enough time is passed that searches will have been called off?

This guy stayed with his vehicle for over two months, until he finally died from starvation [1].

At some point you have to recognize that help isn't coming. The question then is how to determine when it has reached that point. This should probably be part of planning for any trip where you might become lost and isolated. At the very least, find out how long they will keep an unsuccessful search going there.

[1] https://www.oregonlive.com/pacific-northwest-news/index.ssf/...

Very easy to say from a computer. Not easy to say 168 hours after no contact with the outside world.

Usually the advice is to stay where you are. This conserves energy and reduces the search area because most hikers leave for a specific destination known to friends. If you get lost and try to find your way out, you increase the search space.

Build a fire if you can, a smoky fire is better because it can be spotted from far away.

That’s usually good advice but that also doesn’t always work. https://www.google.com/amp/s/amp.theguardian.com/us-news/201...

She was 30 minutes away from a road and survived for 26 days but decided to stay put and wait for people to find her.

There are always exceptions to every rule-of-thumb.

In this case, sadly.

How sad this is.

Yet according to her diary she'd been wandering for two days prior before deciding to stop. The trick is to stop as soon as you realise you're lost.

In the first instance, it gives you time to overcome panic.

> Adam now knows that at least three K9 teams came to about 100 yards of the camp.

Speaking of search space...I've wondered if it would make sense to establish a virtual grid in wilderness areas, based on GPS coordinates. Then establish a convention that lost hikers should move to one of the intersections on the virtual grid and wait there.

An air search for an essentially a randomly placed point (the hiker) in a large area is very hit and miss. It is quite easy for even a trained observer to scan right over someone.

It's much easier for an observer to search a fixed location. The searchers could first specifically look at the grid intersections in a given sector of the search before then doing the general area scan, and then perhaps do a second check of the grid intersections before moving on to the next sector.

The idea is to have known places that will definitely be looked directly at by an attentive searcher during the search rather than just seen in passing in a sweep, that the lost person can find if they have a GPS (or if they have a map that has these places marked).

But if they are lost how can they find these points? If they have GPS and know where they are then they can find their way out. This only makes sense if you know you will be in trouble shortly but are not yet, I.e. you are very low on water, but most people don't realize how deep they are, then panic.

Before you set out on a trip it's best practice to read the map for the area and identify some natural or man-made features that bound the area you're going in. Then if you're lost you can use your compass to navigate to one off these "hand rails" and from there you know that you're somewhere on that line.

>> Build a fire if you can, a smoky fire is better because it can be spotted from far away.

Car tires burn very well. Lots of black smoke. And there are parts enough in any car to start a fire (wires+battery etc).

Don’t underestimate distance based on what you recall or see on a map, especially in the wilderness.

There’s times when people think it’s ‘only a few miles’ of a hike, but sometimes that can take a whole day based on the terrain, and you don’t want to find yourself still stuck in the woods when night falls with inadequate supplies.

I learned the hard way that the only reliable estimate of distance in the bush is hours.

Prepare properly. I bring a map and compass anytime I go out for a hike, even if it's a local trail that I know well. If I am going outside of cell coverage I bring a 2m radio (that requires your ham license but it is very easy to get).

Tell someone where you are going and give them a check in time.

Always bring a decent knife (within your local laws), a fire implement, and a decent medical kit.

"I bring a map and compass anytime I go out for a hike"

It's also a good idea to learn how to actually use a map and a compass. Doing so effectively is not necessarily the most intuitive thing in the world, and does take some skill.

Seriously don't take an ordinary passenger car up a dirt logging road in the Cascades, Oregon coast mountains, or Olympic peninsula mountains in mid Winter unless you are intimately familiar with the entire road, and have driven it many times before. There are huge dead areas of zero cellular coverage where your only option would be long range vehicle mounted vhf-uhf radio, or iridium satellite phone.

I have a high clearance 4x4 with 31x12.5" AT tires and still wouldn't try to take a shortcut like that.

Not so much because I'm terribly scared of getting stuck, in first gear, but putting on tire chains in a muddy slushy pacific NW winter is not really my idea of fun or a time saver.

You should never travel on rural roads without a verifiably working satellite radio (or personal locator beacon) in the worst case scenario. For communication, I travel with my main cell phone, a backup cell phone / wireless hotspot with a completely different provider (T-Mobile + Verizon, lately), a satellite communicator (several links posted in this thread that will work just fine), flares and high-intensity LED road lights that have constant + SOS flashing patterns, two high-intensity flashlights, and a high-powered green laser pen for alternative forms of signaling.

Additionally, you should obviously carry plenty of food, potable water, basic medical supplies including a round of antibiotics, tire inflation/repair devices, a multipurpose sharp knife, two long-lasting external batteries for your USB devices that can also jumpstart a car (super cheap on Amazon, surprisingly), and potentially a firearm if you're into that kind of thing and trained to use it. Standard prepper checklists will work here.

It sounds like a lot, but even semi-rural travel 15-20 miles off marked and standard roads are dangerous as hell, as this story and others clearly can show. You will almost certainly never need all of your equipment, but if you travel even occasionally, you will absolutely need some of it - and you won't be able to predict which piece of gear you do need. It's about being prepared for a wide variety of events, not one severe event.

Nature doesn't cooperate.

EDIT: Forgot, pack a basic set of tools: Screwdriver with various bits, socket wrench (ideally with an extension), car jack, pliers, length of wire, length of rope, aim-and-flame extension lighter (vastly preferred to a handheld cigarette lighter and not any more expensive; get a flexible shaft if you want to be fancy and shell out an extra $2), tire iron (add a breaker bar, probably), and a stack of kindling, which can simply be a bunch of old or free newspapers.

Really great advice here. Sometimes might feel a bit stupid having all that equipment for a relatively short journey, but so worth it when you do need it.

It doesn't even take up much room, even if you travel with considerable water supplies (which takes up the most room and also is the most important, both for obvious reasons).

Almost everything else on the list is space efficient and/or packs in specific areas in your car made to store things like tools, spare tires, etc. People have a lot of room in their vehicles that go to waste simply because they don't read user manuals.

I would include a folding saw, I believe Silky to be the best and most reliable. Much better than a a small ax.

"It sounds like a lot"

Yes it does. So many people travel without the basics.

I remember reading some years back about a couple being found dead of thirst in the Sahara desert, in to which they went on a day trip... without bringing water.

That... is impressive. But sadly not that surprising.

I heard Joe Rogan (not a frequent listener, but figure he deserves credit) say this once, paraphrased: Look at a 7-11 counter full of cheap pocket knives, $1 lighters, flashlights, and other stupid items. There comes a time in many peoples' lives that they would trade their entire wealth for access to a single one of those tools in a time of dire need.

Spend a few hundred dollars, pack it away, and hope you wasted your money. It's the best possible outcome. But should you need anything in the kit, that $300-500 you spent will have the highest return on investment on anything you bought in your lifetime, even if you were an early adopter of Bitcoin or Google.

Awesome advice dude. If this saves one life its worth all the effort.

The ideal scenario is that it saves none. :)

Totally agree, however. And thank you very much.

Wait by your car, start a fire. He would have lived if he did.

They literally did just that for a week. Didn't help.

The car was still found before the person.

Well yea, car size > person size. In this case, making a fire had no effect on reducing the time to find them.

I wonder if a fake phone mast, similar to a stingray, would be useful for locating people lost in the wilderness. If the person had a powered on phone, and the area was located quickly enough, the device could at minimum tell you if you are hotter or colder. Use multiple and you could even triangulate!

If you put one on a helicopter or drone you could sweep a large area quite quickly. That could be especially useful in situations where a person wasn't planning to be in the wilderness and thus would be more likely to have a phone on them but not another GPS-enabled communicator.

Is really amazing to see the efforts we will go to in order save someone. Millions of dollars and massive volunteer effort is spent every year on SAR.

Historically it wasn't unusual for people to just never return. You'd never know what happened.

Puts human development in the 20th century into perspective for me.

A lot of lost people make their situation worse by wondering around and making it harder to be found.

> His car and his family were eventually found (from inspection by volunteers of aerial footage)

I don't believe there was any aerial footage. If I recall correctly, he was found by a local person who owned a helicopter and followed some hunches to find his car tracks at one point then quickly located them by following the roads.

It’s been a while since I’ve gone backpacking, but especially if I went solo I’d bring a locator beacon these days. https://www.rei.com/learn/expert-advice/personal-locator-bea...

even a fsr/pmr radio would help

I always carry one when tramping or hunting, especially solo, but the caveat is you have to be able to activate them - as most fatalities in the New Zealand bush are caused by falling (then drowning, then hypothermia), they're less helpful when the person is severely incapacitated.

Sadly, if this person has been missing 30 days, probably the best bet of finding them is a cadaver dog - our SAR volunteers use them.

Right there’s a limit, there’s also heat exhaustion and dehydration which affect ones ability to think. It looks like the terrain where Sam is a fall would immediately fatal due to the lack of trees and jagged rocks, your pack would increase your roll even.

Seems like it'd be smart to have a version with a "dead-man timer". Something like every X hours a beeper goes off and you have an hour to actively interact with the beacon (button, pin code, etc)

You could do something like that with an InReach device and it's satellite messaging functionality. Have a script (or person) expecte hourly preset message from you (it can send emails, sms, and even Facebook or Twitter messages) You could also use the tracking functionality as a double check that they didn't just forget to send the message by looking if their speed has changed inexplicably during recent intervals.

It's not as ideal as it does ultimately require a human to raise the alarm.

The false alarm rate would be way too high to be use useful just from people losing them. Maybe if you built it into a bracelet or anklet that's hard to take off, and there was a huge penalty for a false alarms?

Also if something happens to incapacitate you for several hours out in the wilderness, you'll mostly likely be dead before help arrives--even with a beacon.

A locator that can be remotely activated (periodically wakes up to check for an activation signal) could be pretty useful, but I think it would mostly be used to locate bodies. Requiring solo hikers to carry something like that might be good alternative to parks requiring rescue insurance.

I found myself looking at the video wondering if it would be feasible to sell bandanas with high contrast QR code on it, basically a digital SOS that image recognition software could easily identify in a fly by

more backstory https://medium.com/@iandustinscofield/lost-in-the-mountains-...

Where is the AI that would scour these videos instead of humans?

Busy eating captchas.

I wonder what the training data would look like to accomplish this task? Hide a bunch of objects and people in the wilderness and take drone photos?

It would be interested to put a thermal camera on the drone and fly at night.

AI is good for convenience and expediency, but when it comes to thorough searching of a video for another human being, nothing yet beats humans.

I'm reminded of that time when someone's child (?) got lost at a concert and there was a forum trying to find them. Don't recall the name or how it ended, though.

Reply on news.ycombinator.com