Here's what needed IMO to do this reasonably well

[billyd]

You need two hi resolution cameras (at least 480 lines) and each with a 90 degree field of view. Set the cameras about 3-4 inches apart and have them divergent about 3-5 degrees. Set both of these cameras on a pan and tilt system. Now you can either use two transmitters (one for each camera) or one and alternate frames in bursts. That will require some additional electronics to sync them to the stereo goggles. On the ground you'll need either two receivers or one receiver and the aforementioned electronics to send the alternate frames to the two monitors in the video goggles. If you have two receivers, each receiver will send it's signal to the separate video monitors in the goggles.

Attached to the goggles will be the head tracker which will communicate to the pan and tilt through your regular Tx that you control your heli with.

Additionally, superimposed onto the images would be some flight info such as altitude from an onboard sensor and an artificial horizon.

Until you can build something like this you are wasting your time in FPV and will likely crash your model or kill someone without spotters or a buddy box setup. JMHO.

[Kepler]

Wow. Its just gone from about $600 to $1,500 to have a go at this. Thats depressing.

[billyd]

Well you can always do it with one camera and the off the shelf systems that you find. It's just really hard to see that well in practice. Depth perception and visual clarity are a huge problem. It's ok for high altitude flying in open space. Just don't expect to be flying over your neighbors house peeking in his windows (lol) and landing with the goggles on.

[jrohland]

You do NOT need 3D to takeoff, fly and land. One camera is perfectly usable and safe. Even if you did want 2 video feeds, there are not enough radio channels available to make it work.

Guys have been flying aircraft under goggles with one camera for years. It works fine. Any of you could drive your car easily and safely with one eye closed.

Depth perception is only good out to about 7 meters. Beyond that, you use other ques to judge distance.

jrohland

[billyd]

I'm sure there are people out there that could fly FPV with a pinhole camera, but the technology is out there right now to do a nice fully 3d system that would essentially have no shortcomings. It would be exactly like sitting in the cockpit without the g forces. The other huge byproduct of this system would be a huge field of view with no distortion.

I figure if you're going to do this you ought to do it well. The hobby we're in is not for people without cash. What's another $1000 bucks or so? That's two batteries on a 10s setup.

Like I said you don't need two Tx's and two Rx's (that's the brute force method). You could do it with one each and simply alternate frames from one camera and then the other and send the signal to one monitor and then the other in rapid sequence. We generally fly in broad daylight so light is plentiful to run high shutter speeds.

My point in posting this thread was not to flame FPV it was just to get the folks out there producing these products to stop playing around with low rent systems and produce something that will seriously kick butt. This is the system I want, unfortunately I couldn't build this myself because I don't have the electrical engineering background required.

[DarkHeli]

3d or stereoscopic vision (I think that's the term) has been proven to be ineffective for FPV use. Some guy tried it on a plane and said little difference between that and a 1 camera system.

[DarkHeli]

Originally Posted by jrohland

Depth perception is only good out to about 7 meters. Beyond that, you use other ques to judge distance.

That is the reason I believe he said it was ineffective.

[Windbreaker]

My dad played pool, drove and rode a motorcycle at breakneck speeds, and he could see only out of his right eye.

[billyd]

Your dad had feedback from his one eye to his brain regarding focal length which aids in depth perception.

Regarding "it's been tried". If it didn't make a difference it wasn't done well. It's not about 3d as much as it's about an undistorted field of view that nearly matches human sight.

And regarding the 7 meters, well 7 meters next to a tree has a pretty big impact on your model if you don't judge it right. And the last 7 meters of a landing are fairly important too.

I remain unswayed.

[Macsgrafs]

Originally Posted by billyd

Like I said you don't need two Tx's and two Rx's (that's the brute force method). You could do it with one each and simply alternate frames from one camera and then the other and send the signal to one monitor and then the other in rapid sequence. We generally fly in broad daylight so light is plentiful to run high shutter speeds.

My point in posting this thread was not to flame FPV it was just to get the folks out there producing these products to stop playing around with low rent systems and produce something that will seriously kick butt. This is the system I want, unfortunately I couldn't build this myself because I don't have the electrical engineering background required.

Now this is how I intend to fly FPV 2 x cams with a fast switch so that I only need 1 video downlink. I agree that the cheap crappy cameras & links should be ditched, there is quality out there if you know where to find it.

Ross

[Windbreaker]

An airline here ran into a difficult case when a one-eyed pilot applied for a job. The FAA cleared him to be a pilot. But of course airlines don't want to hire pilots who don't have two working eyes.

http://archives.starbulletin.com/97/...ss/story3.html


Bruce Pied, flying commercial aircraft for the past 10 years, lost sight in his left eye at 18 after a disease attacked his optic nerve. He experienced initial loss of depth perception but later learned to adjust.


"Ask anybody on my son's baseball team if I can catch a baseball," said Pied, who's been coaching the team for six years.


Dr. Russell Stodd, a Maui eye surgeon who flies a single-engine plane, sees no handicap in the loss of an eye. Depth perception derived from two eyes -- stereopsis -- can be relearned with one, he said.


"Stereopsis is very important for someone who works up close, like a jeweler or an eye surgeon," he said. "In terms of looking far away, which is what an airline pilot is doing, the ability to see with one eye serves just as well as two."

[billyd]

Once again a single human eye is far superior than a single camera for depth perception. The eye is connected to the brain on many levels, including the stress on the eye muscles to move the eye into focus. In individuals with one eye, the brain can quickly learn to compensate for lack of stereo vision to determine depth using the eye muscles as feedback for distance to an object.

A single camera has no way of giving depth information to a person outside of the general speed of objects at different distances. Add this to image distortion from wide angle lenses, or lack of field of view from standard angle lenses and you end up with a situation where close quarters maneuvering is extremely difficult.

Two cameras in a stereo arrangement, implemented correctly, would not only yield a true 3d image to the brain, it would also generate an UNDISTORTED field of view of approximately 160 degrees. Depth and situational awareness as well as simply clarity would be infinitely superior to one camera. Again if people have tried this and were disatisfied with the results, the problem would likely lie with the implementation of their system, rather than the idea itself. I guarantee you that if you used hi def cameras in a true steroscopic arrangement the result would be astonishing. Another big part of the system would be the stereo display (goggles) itself.

[Windbreaker]

I believe that if you eqiup it with a decent camera and train for it, it can be done.

Lindbergh learned to land the Spirit of St. Louis without a windshield. He could only see out the side windows. The forward view was provided by a periscope.

When I was learning how to fly, my flight instructor taught me to land at night, relying strictly on the runway lights and kept the aircraft's landing light turned off.

Since I was trained that way, it became easy for me to do it. The landing light became a safety item so that other aircraft could see me. But I didn't need it to make a smooth touchdown.

Here's a guy who flies with a single camera and does it from inside his house:

http://www.fpvvideo.com/fpv_helicopter.htm

There's even video of him succesfully flying and landing with his setup. It's a standard definition camera. Nothing too fancy.

[merlin703]

I think there is a misunderstanding here and perhaps the initial post wasn't too clear in some aspects since it kind of reads like you HAVE to spend $1,500 to get started with FPV. Of course that is not the case. BillyD was just trying to "push" a new set up that all new FPV equipment should start with (companies making them). Yes, "3D" with FPV would be really nice and the technology is make it happen is already there but you don't HAVE TO invest that much for JUST PFV. Perhaps, you can adventure yourself on "3D" once you have already started and become a "pro" of how to fly, set up, and mod your FPV stuff.

You can get started with FPV for under $500 for sure. I have actually seen a guy fly FPV with his equipment UNDER $250. I didn't believe it until I saw it.... and the quality was GOOD
One camera, on patch antenna and the cheapest tx/rx set from range video... he had 25 inches tv on the back of his SUV to make it FPV.

BillyD... by the way... put some underware on man... there are kids in this forum

[billyd]

I totally understand that you can "get by" with a single camera and learn to deal with it. It's just that we have the potential for a far superior system and it seems like no one wants to bother. The electronics required for a 3d setup would seemingly be trivial in this day and age and the increase in cost would be in the hundreds of dollars. In this hobby, that's peanuts. I don't see the point in doing things half way, unless there are significant obstacles to doing it all out. In this case, it just appears to me that no one's gotten around to doing it. I'm just trying to get the wheels turning.

And Merlin, regarding the underwear, I find them very constrictive.

[billyd]

ps Another huge problem with FPV is specific to helis. Helis can fly backwards, sideways any direction. Generally speaking planes can't, they typically follow their nose.(Except for 3d style). So an FPV system for a heli, in order to be safe, should have some method of seeing all around you. Single camera FPV systems with fixed mounting create the illusion of more room than exists in reality due to distortion and limited FOV (Objects are closer than they appear). You might fly over a spot and think you are 50 feet from a tree (behind or off to the side) right as you are hitting it.

[javaheadjames]

To the best of my knowledge, NTSC video supports interlacing, and I'll bet some talented person could make one transmitter send alternating stereo perspectives in the time frames for the alternate (even or odd) fields. In other words interlacing the two perspectives (each at 29.9 fps at half the vertical resolution of NTSC, or 240 of 480 lines) into one video stream. Encode the video with the left view in the odd field, the right view in the even field. You'd need a receiver that would essentially reverse the multiplexing in the same way. I think its equally possible in PAL. With two cameras interlacing a single video feed, I think you could use specifically colored filters in your cameras and simply display the filtered double-image on a TV and look at it with color 3D glasses.

I have some questions: Can you space the camera pair less than your interocular length? Can the the camera pair scale down without distorting the 3D effect? Would not you get a sufficiently 3D experience looking through the two eyes of a little mouse to perceive the depth cues?

If you could space the cameras more closely to each other, you could enhance the perspective with periscopes with smaller mirrors.

One could place mirrors (like penta-prism) within the field of view of the camera that are angled and sized to make a split-view of front, side, down, panoramic views by using curved mirrors. You could have a series of linear and curved mirrors that create a single stereo composition of the front, down, and back, panoramic, horizontal and vertical peripheral views, on which you could even superimpose flight data. However, you might be upside-down or backwards.

Now all that is left to be desired is IR or LIDAR. I wouldn't want to encourage any kind of brinkmanship or anything.

Greater insight, anyone ?

You propeller heads are something else.

[xspare]

Originally Posted by javaheadjames

...Now all that is left to be desired is IR or LIDAR. I wouldn't want to encourage any kind of brinkmanship or anything.

Hmm, maybe a synthetic aperture imaging radar?

[HeliEye]

I've been toying with the idea of stereoscopic FPV for some time, like mentioned above the brute force method is likely to yield the best results.. but is more expensive.

Ideally 2 identical cams, spaced for best results on the heli, 2x TX's (i'm using 5.8GHz) 2x RX's, one on CH1 & one on CH 7.. Now for the tricky part, using lcd glasses, you would need two pairs, modified to build one Stereo pair ;-) I think it can be done ok if your an electronic engineer (I am btw) assuming this has been achieved?, both LCD's screens will be running at full frame rate, this is good. the other way of chopping the video for left & right is likely to induce flickering and be annoying, with an apparent half frame rate as 2 into 1 doesn't go, unless you can double the frame rate to compensate?

Another possible problem, the view your camera's are seeing..3D wise, is likely to be right at one set distance, but possibly out at shorter distances? our eyes turn in very slightly to compensate, not quite sure what will happen with fixed mount camera's?

Having spent £1000 on heli stuff in the last few weeks, it maybe a while (when the wife cools down) before I can afford to try this out ;-)

Just a few thoughts

Steve

[Weithy]

The further apart the cameras are, the better the sense of depth will be.
You can expirement by using two handycams as eyepieces.
Military range finder scopes had a very wide view, and made the battle field look like a model at arms length.
In the photo, the men are preparing to shoot a nitro heli just over the horizon, and the man on the right is wearing a silly hat.
Acually he is looking though wide binoculars that can give range by converging two images.

If anyone works out how to feed seperate video signals to each side of video glasses, I would like to give it a go.