How can you stream 640x480 video with 1Mbit bandwidth (with Linux)?

N

-=Nu||=-

Gawd
Joined
Jan 5, 2002
Messages
974
Hi, so my question is what kind of software can I use to run on Linux to stream 640x480 video feed if I only have 1 megabit of bandwidth. The reason I'm asking is because I'm working on a project to build a UAV, and we decided it would be best to stream video back from the plane and do image processing on the ground, rather than having a heavy and powerful computer onboard the airplane (here's a thread with some more info about our UAV). We plan on using this for our wireless link, it has great range, but only ~1 megabit of throughput. We plan on having a 400MHz Gumstix on the airplane to do the encoding of the video, and it would then pass it back through the wireless link to the ground station. And in case you're wondering why we aren't using 802.1 a/b/g, it's because we need a max range of 7 miles.

So does anybody have any ideas for what we could use to do this? I thought about using a Slingbox AV to simply do the encoding, but that seems like overkill. Any comments and suggestions are welcome. Thanks.
 
im not entirely sure if the gumstix is up to the challenge... i was about ready to suggest converting it to mpeg2 or something... but it looks like the proc might be a little lacking

EDIT: I read the thread on the UAV challenge, and I ask : why its 30 fps required? even 2 fps seems overkill to me (read: why not 1 picture a second?
 
640x480 at 30FPS?? :eek:

I don't think it's possible..... you are talking about a 125KB/sec stream MAX(your connection)... I can't see that quality of video, ever getting that small.... both of my Canon's for instance record at almost 2000Mbps(double what your connection is capable of) at 640x480 @ 30FPS..... if you lower it to 15FPS, it might be possible.... then they record ataround 900Kbps.... which would work if your 1Mbit connection was working at full speed with no lag.
 
30fps may not be a requirement, but it would be nice, the reason I want a high frame rate is that the UAV will need to search a 2x2 mile grid in an hour and find a person, so the higher the frame rate, the faster the airplane can go. I spoke with the company that makes the Slingbox and they said that streaming 640x480 video with 1 megabit of bandwidth would not be a problem, so we may have to give that a try, although I'm not sure how much the person I spoke with actually knew what they were talking about.

EDIT: I guess I could look for higher bandwidth long range links, we will need a maximum range of seven miles, hopefully it will be line of sight. I would love to have a standard 802.11g connection, but I'm not sure if we can get that kind of range. Any ideas for the wireless link?
 
802.11g stuff will work over that distance the problems will be that you may need to point an antenna at it.

have you considered seperating the video and data links? maybe something like http://www.clubmac.com/clubmac/shop/detail.asp?dpno=7121895&store=clubmac&source=cwbfroogle with a 7dbi antenna

The hard part is the signal will not be fixed so you will need to work out what the best way tto orient the antenna will be which may cause a loss in sig every time you turn for the search grid

if i can get up the hill near here I will see what ranges i can get with my 2 meraki units which would certainly suit being chucked in your uav.

i see that you are no longer sticking the full computer system in the UAV
what changed your mind? the weight requirments?

btw got a public project page or are you keeping all your stuff secret

I am definatly going to have to go drive the 3000 odd kMs to this thing to check it out

/me reloads the laptop and gps systems ready for a 3000kM wardrive(each way).
 
tacticus said:
802.11g stuff will work over that distance the problems will be that you may need to point an antenna at it.

have you considered seperating the video and data links? maybe something like http://www.clubmac.com/clubmac/shop/detail.asp?dpno=7121895&store=clubmac&source=cwbfroogle with a 7dbi antenna

The hard part is the signal will not be fixed so you will need to work out what the best way tto orient the antenna will be which may cause a loss in sig every time you turn for the search grid

if i can get up the hill near here I will see what ranges i can get with my 2 meraki units which would certainly suit being chucked in your uav.

i see that you are no longer sticking the full computer system in the UAV
what changed your mind? the weight requirments?

btw got a public project page or are you keeping all your stuff secret

I am definatly going to have to go drive the 3000 odd kMs to this thing to check it out

/me reloads the laptop and gps systems ready for a 3000kM wardrive(each way).
Click to expand...

Pointing an antenna at the plane probably won't be an issue. The reason we decided not to have the image processing done on the UAV was because we could not find the perfect system to put on the plane, and by doing the image processing on the ground a lot of things are greatly simplified, and we essentially have unlimited processing power (just get a faster laptop). We're still working on our website, but I just got it up yesterday, here's the link: www.aessuav.org . We will probably be updating our site with updates somewhat frequently.

I would really like to use 802.11g, and I guess if we get a powerful enough card, and a massive yagi antenna, I think we might be able to get the kind of range that we need. Although for the actual link with the autopilot we will have a powerful 900MHz link with a range of up to 40 miles, just not much bandwidth (~112 kilobits/sec).
 
Ditch the data link for video and just use a dedicated video transmitter and receiver.
 
Eickst said:
Ditch the data link for video and just use a dedicated video transmitter and receiver.
Click to expand...

Do you know of any video links that can have the kind of range that we need and are relatively cheap (<$700)? The only reason I wanted to use a datalink for the the video was that I was able to find a datalink that had the kind of range that we needed.
 
-=Nu||=- said:
Do you know of any video links that can have the kind of range that we need and are relatively cheap (<$700)? The only reason I wanted to use a datalink for the the video was that I was able to find a datalink that had the kind of range that we needed.
Click to expand...

You won't get that kind of range and data out of that device. I can almost guarantee it. Don't rely on their specs for your project, do your own testing.

Maxstream has some cheaper radio modules than what that company was selling.

What kind of range do you want?

The problem with receiving video from an airborne moving object is that it rotates and may block out part of the signal with its own airframe. If the antenna rotates 90 degrees you may receive no signal at all.

I think you will find that a combination of transmitters/receivers will work best. You need a diversity receiver, one with two oppositely polarized antennas. Black-widow av has one that's relatively cheap, but I can't speak for it's quality as I have never used it.

If you want reliable video transmission of over 5 miles you had better be prepared to raise your budget. If this was a stationary object it would be no problem, but an aerial is tough due to antenna size limits, antenna movement, etc.

Stick to 1.2ghz for video, or 2.4. 1.2 will require a license, but it will have less multipathing and will penetrate more objects like trees, etc. 2.4 is license free but has less than half of the range of 1.2 at same transmit power. 1.2 requires larger antennas than 2.4

It's all about tradeoffs.

Tell me what your goals are, and tell me what your REQUIREMENTS are. They shouldn't be the same.
 
Eickst said:
You won't get that kind of range and data out of that device. I can almost guarantee it. Don't rely on their specs for your project, do your own testing.

Maxstream has some cheaper radio modules than what that company was selling.

What kind of range do you want?

The problem with receiving video from an airborne moving object is that it rotates and may block out part of the signal with its own airframe. If the antenna rotates 90 degrees you may receive no signal at all.

I think you will find that a combination of transmitters/receivers will work best. You need a diversity receiver, one with two oppositely polarized antennas. Black-widow av has one that's relatively cheap, but I can't speak for it's quality as I have never used it.

If you want reliable video transmission of over 5 miles you had better be prepared to raise your budget. If this was a stationary object it would be no problem, but an aerial is tough due to antenna size limits, antenna movement, etc.

Stick to 1.2ghz for video, or 2.4. 1.2 will require a license, but it will have less multipathing and will penetrate more objects like trees, etc. 2.4 is license free but has less than half of the range of 1.2 at same transmit power. 1.2 requires larger antennas than 2.4

It's all about tradeoffs.

Tell me what your goals are, and tell me what your REQUIREMENTS are. They shouldn't be the same.
Click to expand...

Thanks for the help, here are the requirements:
  • Weighs less than 5 lbs
  • Range of 7 miles line of sight
  • Physical dimensions less than 3x6x10 inches
  • Voltage requirement <= 14.7V
  • Max current comsumption of 2A

Our goal is to be able to transmit the video back from the UAV over a distance that will be less than seven miles most of the time, although according to the competition rules it can be a maximum distance of ~7miles. We would be able to design an antenna mount that keeps the antenna pointed at the UAV, which would probably help a great deal with highly directional antennas. A resolution of 520 or at least 480 lines would be nice. As for the airframe blocking the signal, I hope that won't be an issue since the airframe is made from balsa, of course there will be some metal such as the engine and electronics, but I don't think this will be an issue. The Black Widow system looks decent, I'll look into it.
 
Less than 5 lbs? Sweet jesus how big is this UAV? Most of my planes with equipment weigh less than you're willing to allocate to just a transmitter.

You guys are building your own UAV, that alone belies some kind of technical competence.

I suggest you simply build your own video transmitter. Use an Airwave 2.4ghz module and an RF amp. Use two receivers and build your own diversity box or just buy one from Blackwidow AV.

You might try contacting them, they might be able to do one for you that fits your requirements or at least send you in the right direction to a proper vendor. It's been so long isnce I actually bought anything, I just build most of my stuff now.
 
Eickst said:
Less than 5 lbs? Sweet jesus how big is this UAV? Most of my planes with equipment weigh less than you're willing to allocate to just a transmitter.

You guys are building your own UAV, that alone belies some kind of technical competence.

I suggest you simply build your own video transmitter. Use an Airwave 2.4ghz module and an RF amp. Use two receivers and build your own diversity box or just buy one from Blackwidow AV.

You might try contacting them, they might be able to do one for you that fits your requirements or at least send you in the right direction to a proper vendor. It's been so long isnce I actually bought anything, I just build most of my stuff now.
Click to expand...

There's some pictures of our plane on our website (www.aessuav.org or www.flickr.com/photos/aessuav/), we're using a Sig Kadet Senior (80 inch wingspan) for our aircraft. We are on a somewhat tight schedule, so I would prefer to buy a transmitter/receiver, although I think if we throw a decent RF amp we might be alright.
 
Buying a commercial product is going to cost you an arm and a leg.

Can you build an RF amp on your own? If you can, just buy a TX/RX kit, and build an amp for the TX on the aircraft.

I would suggest something in the 3-5W range to start...maybe you want to go bigger for whatever reason, but a 5W linear amp will pull close to 1 amp at 12VDC, in addition to the 300-400mA that the transmitter will take, so it's under your current limitations.

For size, the biggest element of all of this is the heatsink. The RF amp is basically all heatsink, so I would use something like aluminum for weight contstraints, and maybe place it in the airflow of the airframe for cooling purposes. That allows you to keep the size down to a minimum.

If you absolutely have to have a commercial product, I really can't recommend anything to you as I have never used a commercial RF amp. Sorry. Email the guys at blackwidow and see what they have to offer if anything. They deal with more commercial products than I ever will.
 
You must log in or register to reply here.
Back
Top