What this does is remove a huge mess and headache of trying to mate a Spartan 6 FPGA with a BBB/Raspi in a small formfactor. The project also references a GitHub project called HardCV, which does lots or the heavy lifting with signal processing. The result: Run the higher-level CV application code on the BBB/ Raspi, and the disparity matching/stereopsis code is munched on the FPGA. I imagine that this could result in real-time (50 Hz or faster) 3D disparity maps with color tracking and such. I'd love to find somewhere that I could work on this for a masters degree, as soon as I've graduated (calling all researchers with lab openings! :) )
LOGi FPGA Development Board for Raspberry Pi - Beaglebone
LOGI, when FPGA meets community supported ARM platforms.
I've got one of those pixy-cmu-cams coming actually and I'll try and squeeze in some time to try it out. I suspect I'll end up sticking with the Odroid because it's got more flexibility (at the cost of having to do the opencv work myself).
I think we'll have to buy some test balloons and go give it a try but it is unfortunate that when 1 second on the clock equals 1 point, the 25 points / seconds for the balloon isn't worth the time it takes to find them. We also might want to suggest to Robert that they increase the point value so more people will go for them.
Yeah, they originally said the balloons would be 10 feet, and I didn't see how any man-portable aircraft would be able to survive popping them directly. Now they are smaller, there's a chance. But seeing them bobbing around in the air like that... I can't even imagine how that would be done if there's any wind at all.
Yes, I can't justify going for the balloons as an actual way to win the competition but I'm still going for it because where's the fun is you just play it safe.
I was too conservative last year (i.e. flew at too low a speed), I thought most people would end up in the drink and I'd be the last man standing but instead I found that the reasons for failure were hard to forsee and avoid. Best just to go max speed, no holds barred and get maximum points and hope that that's enough to make up for whatever goes wrong.
Comments
@Randy:
Another one I'm really excited for is the LOGi FPGA board here: https://www.kickstarter.com/projects/1575992013/logi-fpga-developme...
What this does is remove a huge mess and headache of trying to mate a Spartan 6 FPGA with a BBB/Raspi in a small formfactor. The project also references a GitHub project called HardCV, which does lots or the heavy lifting with signal processing. The result: Run the higher-level CV application code on the BBB/ Raspi, and the disparity matching/stereopsis code is munched on the FPGA. I imagine that this could result in real-time (50 Hz or faster) 3D disparity maps with color tracking and such. I'd love to find somewhere that I could work on this for a masters degree, as soon as I've graduated (calling all researchers with lab openings! :) )
@John, @Stephen,
I've got one of those pixy-cmu-cams coming actually and I'll try and squeeze in some time to try it out. I suspect I'll end up sticking with the Odroid because it's got more flexibility (at the cost of having to do the opencv work myself).
And, the more recent and cool version of it: https://www.kickstarter.com/projects/254449872/pixy-cmucam5-a-fast-...
These are made to sniff out red balloons... http://www.cs.cmu.edu/~./cmucam2/
... and there were bonus points for a German accent. :D
Would be better if there were 99 of them (younger folk may not understand)
I think we'll have to buy some test balloons and go give it a try but it is unfortunate that when 1 second on the clock equals 1 point, the 25 points / seconds for the balloon isn't worth the time it takes to find them. We also might want to suggest to Robert that they increase the point value so more people will go for them.
Yeah, they originally said the balloons would be 10 feet, and I didn't see how any man-portable aircraft would be able to survive popping them directly. Now they are smaller, there's a chance. But seeing them bobbing around in the air like that... I can't even imagine how that would be done if there's any wind at all.
Yes, I can't justify going for the balloons as an actual way to win the competition but I'm still going for it because where's the fun is you just play it safe.
I was too conservative last year (i.e. flew at too low a speed), I thought most people would end up in the drink and I'd be the last man standing but instead I found that the reasons for failure were hard to forsee and avoid. Best just to go max speed, no holds barred and get maximum points and hope that that's enough to make up for whatever goes wrong.
I think the balloons are a sucker bet, 50 points doesn't seem like enough to justify the risk of entanglement.
And like Randy said, hitting those things will require luck or tremendous tech (especially in the wind).
Any time spent aiming for them will likely be a loss.
-
1
-
2
of 2 Next