What happens when you strap a stereoscopic camera onto a drone and transmit the video feed directly to your Oculus Rift? A pretty amazing experience, that’s what!
Several students from the Norwegian University of Science and Technology recently finished a term project dubbed Oculus FPV. In it, [Erik Hals], [Jacob Prescott], [Mats Svensson], and [Mads Wilthil] succeeded in combining virtual reality, a head mounted display, and a UAV for a great result.
Drones with cameras are the next big step in search and rescue, remote inspection, and many other use cases in other environments that are typically inaccessible for a human to poke around. What we really like about this project is they also mounted the stereoscopic cameras on a gimbal, allowing for full head movement — this means the pilot can “park” (read “hover”) his drone in remote locations, and then look around, without having to worry about performing complex aerial acrobatics to get the right camera angle.
This isn’t the first time we’ve seen something like this done, but at least the hardware has gotten quite a bit smaller!
[Thanks aRez!]
Nice response time
World Trade Federation!
“What happens when you strap a stereoscopic camera onto a drone and transmit the video feed directly to your Oculus Rift?” – I’m guessing you’ll feel really bad really fast :P My personal experience with the rift was that you have to be very careful or you get really really sick. It IS however an amazing project, one that will get better once the hardware comes up to par (the rift).
An ‘out of body experience’ simulator!
+1
it seems an easy project to me, isn’t it ?
this is exactly what I thought. Probably lots of small kinks that they had to sort out. Cool project regardless (although I’m not sure how useful this actually is)
Just some thoughts: wouldn’t it be better to remove stereoscopic view in favour of a higher resolution? Unless you fly your drone very close (like less than maybe 2.5m), you won’t have much of a stereoscopic effect anyway – the camera lenses are just too close, and if you widen the distance again, you either get motion sickness or everything will just look like a toy model again…
I would have to agree. Unfortunately the OR is the bell of the ball for those under 40 years of age that have not seen something like this every decade. I will be glad when the hype dies down, because if you look it has all of the red flags of vaporware. It is nothing new or special- just hype by a few in the industry. I would say to go look at arduino’s support forums for a loose idea of its ttl. I feel that they will sell a bunch of units but support will be completely lacking as with most of these type of things. The kind of support ya get from facebook seinfeld money lol.
Kudos to the students working on the project- it does have nice response time as others have said
“because if you look it has all of the red flags of vaporware”…what? They sold 50.000 DK1 kits, are shipping DK2 kits right now, got bought by facebook for $2Bn. How on earth is that vapour ware? OR is not the same as the earlier VR models, it really really isn’t.
Are you talking about Bill Brasky? I KNOW BILL BRASKY! Bill Brasky is a sonofabitch!
“He’d eat a homeless person if you dared him!”
“His poop is used as currency in Argentina.”
“He sweats Gatorade”
“He once breast-fed a flamingo back to health.”
“I once saw him scissor kick Angela Landsbury.”
“He sheds his skin once a year.”
“He makes brooms somewhere in Georgia.”
You’re ignorant. The rift is the culmination of high pixel density displays, high processing power, and accurate motion tracking capability. The Rift would not have been possible in the 90’s. Period. The display is what makes the difference.
In the future people are going to be using high-pixel-density displays right next to their eyeballs to simulate massive displays for non-interactive purposes. All the VR implications are moot, this is a 100+” screen (or whatever size ~120 degrees in FOV equates to) that costs ~$300 in practical terms.
You are ignorant BillBrasskey.
The rift is the culmination of high pixel density displays, high processing power, and accurate motion tracking capability. The Rift would not have been possible in the 90′s. Period. The display is what makes the difference.
In the future people are going to be using high-pixel-density displays right next to their eyeballs to simulate massive displays for non-interactive purposes. All the VR implications are moot, this is a 100+” screen (or whatever size ~120 degrees in FOV equates to) that costs ~$300 in practical terms.
The stereo display is a necessity as well. In order to make this design work you need it to be small and light, and in order for that to work it needs to be very close to your eyes- closer than the focal distance. Hold your hand in front of your face, move it closer until the point at which you see double. That’s where the display would need to be. Imagine a rectangle that would fill your vision to 120 degrees at that point in space. It would be too large to strap to your face. The FACT is that displays with a pixel density high enough to not “screen door” into a grid that close up DID NOT EXIST until recent years. That’s why it’s happening for the first time just now, for real this time, no bullshit.
When I say the rift, substitute “HMD” or “Head Mounted Display”; the rift just happens to be the first. Practical and comfortable entertainment HMD’s will become ubiquitous as their overall manufacturing cost will be lower for the impressive experience they produce. Once the OR enters the public eye people will want a version that only serves as a personal display, not as a virtual reality device; and it’ll become common from there.
Gameboy Virtual Boy. And no it has not progressed from the RAPTOR or ONYX in any meaningful way. It will have the same pitfalls and shortcomings. Keep drinking the Kool Aid though. OR needs folks like you to multi-post and keep the hype alive ;)
You are right that there were early attempts to achieve this, back when the tech was not sufficiently advanced. You are not so right with the rest.
You may like to bolster your argument with some numbers – comparison of the display resolution, frames per second, motion tracking latency, and, inter alia, end-user affordability of the setup.
Then, only then, your naysaying may approach something similar to credibility.
That’s exactly what I was talking about. Virtual Boy? You’re being a shithead. If you did actual research you’d discover the display tech making this possible hasn’t even been around 5 years. The Retina display on an iPad – extremely high pixel density? That’s a critical difference, the teeth that will make the gear turn. Until now there was no adequate display technology available; and any and all comparisons with legacy text are irrelevant.
agtrier – You can’t just use the whole screen, each display is sitting inside the eye’s focal distance. The purpose is not to make stereoscopic 3D possible, that’s just a very convenient side-effect. The only way to up the resolution is to increase the pixel density for each eye. If you make a display large enough to sit at the focal distance of the eye, it’s too big to strap to your head, so you might as well keep going making it bigger so that you can sit farther away from it. That’s why you have big TV’s.
If you could make the LCD dense enough, you could just put one right in front of each eye and fake a display of any size. That’s what this is. That’s what’s different about it. All those previous head mounted displays had terrible resolution.
The reason shitty product after shitty product has existed is simple: if you could pull it off, it would be incredible. Engineers know that. Apparently a lot of them (and marketing, and businessmen) weren’t smart enough to step back and say “do we really have the elements to make this work and have it not also blow” so you have a lot of failed attempts. The people who started OR decided to go through with it because they stepped back and said “wait, we actually can do this now.”
They delayed the device once already specifically because they decided the display wasn’t good enough and better alternatives were available. They are also improving the head tracking with external cameras, but that’s almost secondary. WHen you put a 720p rift dev kit on it has light grid-lines. The 1080p version for consumers should mitigate that to the point of the product being acceptable.
Needs vibration reduction it seems. But something simple and light. The trick with HD’s is to hang them on rubber bands, so maybe something based on that concept
Come to think of it, microphone vibration/shock isolators also basically suspend them with rubber bands on all sides.
i’d just leave it here:
http://hackaday.com/2014/06/17/birdly-the-virtual-reality-simulator-guaranteed-to-wear-you-out/
and what is the latency of this setup ?
Yeah it must be pretty bad… latency is already a problem in wired solutions.
Latency could be compenstated somewhat by transmitting the orientation during recording. I image the end result would be much like how video stabilzation works where there is a larger black area and the video frame move around. ( http://www.youtube.com/watch?v=BgAdeuxkUyY ). At least it might reduce the sickness factor.
Yes, I would love to see some actual numbers on this.
In the video at around the 30 seconds mark or the GIF leading the writeup you can see the camera assembly follow his headmovement, the thing is a small drone that needs to be in visual range, radio moves at the speed of light, the cameras are old style 728×488 resolution, and oculus is specifically designed to reduce latency.
I’m waiting for DK2 before getting the transporter3d and 3dcamfpv hardware.