Has Anybody played with the Camera?

[Alienvin]

Has anyone put together the camera? If so, how does it work......? Are there plans....

[KVermilion]

Yes
Somewhat
Yes

[Bharat Nain]

Originally Posted by KVermilion

Yes
Somewhat
Yes

What results have you had? Does it detect the green light easily? How about when light conditions around change slightly? Any more details? Thanks a ton.

[DanDon]

The camera detected the green target very easily, and tracking is pretty fast. Kevin's code works awesomely (as always ). Light conditions dont seem to affect the tracking.

The pan-tilt assembly works great, once you've put it together.

And finally, don't try testing on the old controller unless you take all the separate modules and install them into a version of last year's default code.

Hope this helps somewhat,

[AndyB]

for us we tested it for the first time today. we mounted the green light panel on an overhead projector cart and ran it around to see if the camera around. at times the camera seemed a bit slow or sluggish, but it definitly did a much better job finding it. we have it working on day 2 lol, we didnt get it working on day 42 last year... the camera this year should bring for a promising competition. i am SO curious as to see what other teams have come up with...

[Kevin Watson]

I figured the menus would drop some jaws on the floor, but I haven't gotten any feedback (positive or negative). Are folks trying this feature out? Any problems?

-Kevin

[Rombus]

Originally Posted by Kevin Watson

I figured the menus would drop some jaws on the floor, but I haven't gotten any feedback (positive or negative). Are folks trying this feature out? Any problems?

-Kevin

They look cool, i would love to get my hands on them, problem is our camera isnet tracking. Im guessing its probably the fact our backup battery was near dead, but im not sure. I know i wasnet getting any feedback from the sytem, could it be the fact i was using a USB to Serial cable?

I wish the RC would finally swap over to USB (Hint hint IFI!)

[Eldarion]

Originally Posted by Rombus

They look cool, i would love to get my hands on them, problem is our camera isnet tracking. Im guessing its probably the fact our backup battery was near dead, but im not sure.

YES! The camera picture "brightness" varies tremendously with battery voltage, becoming a dim grey at low voltages. I would suggest building a standalone 7.2-volt regulator (LM317 chip will do nicely), and run the camera off of that. That way your test results will be consistent.

[Chris Orimoto]

Been playing with the camera for 3 days now...

So far we've found the search function to pickup the green light relatively quickly. As for tracking, the camera has an impressive rate so far. I somehow believe that the search/track functions can be bogged down by other code when more complex functions are added. However, the prospect of directing shots entirely by camera seems very promising so far.

As for the menus, kudos to Kevin. Quite beautifully done. So far, I've utlilized them to toy with different settings (without having to reset the entire program). However, the default settings seem to work just fine. My only concern is that the menus completely halts the current code, which could be problematic depending on other functions that should be executed. In the end though, I hope to find absolute settings (that can be loaded through the header file).

One question for everyone though...since the camera is interrupt driven, how does it affect other interrupting encoders? Does anyone know how many interrupting devices the new PIC can handle?

Thanks all

[SpaceOsc]

Originally Posted by Bharat Nain

What results have you had? Does it detect the green light easily? How about when light conditions around change slightly? Any more details? Thanks a ton.

http://www.ifirobotics.com/camera.shtml

the link contains most of what you need to get the cam up and running. ours started working last night and even with no formal calibration it was able to track fast. works great!

how fast it moves worries me some but ill let programming deal with it

[Kevin Watson]

Chris,

Originally Posted by Chris Orimoto

I somehow believe that the search/track functions can be bogged down by other code when more complex functions are added.

This shouldn't be the case because the same amount of camera and tracking code is guaranteed to execute every 26 milliseconds. It's very consistant.

Originally Posted by Chris Orimoto

My only concern is that the menus completely halts the current code, which could be problematic depending on other functions that should be executed.

Actually, it doesn't. As people get into and understand the code, they'll realize that the camera code implements a real (cooperative) multitasking system. Every 26 ms loop, a little piece of camera code and tracking code executes giving the appearance of multitasking. This is exactly how your single processor PC makes it appear as though many applications are running at once. Getting back to the menus, you'll notice the camera still tracks when you invoke the camera menu. Well this happens because I just add the camera_menu() task to the other three tasks that are running. When the camera_menu() code sets a flag indicating it's finished, I remove it from the task list. On the other hand, when you invoke the tracking menu, the tracking stops, giving the appearance that everything else has stopped. Actually, this is by design because while you're in the tracking menu, the user might invoke the interactive PWM sub-menu, which needs control of the servos. BTW, the "scheduler" is located in Process_Data_From_Master_uP() if you're interested in taking a peek.

Originally Posted by Chris Orimoto

In the end though, I hope to find absolute settings (that can be loaded through the header file).

Check out the streamlined version of the code, which gets its configuration from camera.h and tracking.h.

Originally Posted by Chris Orimoto

One question for everyone though...since the camera is interrupt driven, how does it affect other interrupting encoders?

It's not interrupt driven, it's message driven. The only interrupts that are firing off are about 120 serial port interrupts per second to handle the camera telemetry, which is easilly handled.

Originally Posted by Chris Orimoto

Does anyone know how many interrupting devices the new PIC can handle?

It has more to do with how much time you're spending in the interrupt service routines. As a data point, my ADC code will generate up to 12,800 interrupts per second at the highest sample rate.

-Kevin

[Chris Orimoto]

Thanks for clearing all that up, Kevin. With that info in mind, I think I know what to work on next

Back to the other posts (that I seemed to have skipped over)...

Originally Posted by Rombus

They look cool, i would love to get my hands on them, problem is our camera isnet tracking. Im guessing its probably the fact our backup battery was near dead, but im not sure. I know i wasnet getting any feedback from the sytem, could it be the fact i was using a USB to Serial cable?

We found that the servo movement and the camera power was dependent upon the backup battery... so you might want to change that before messing with the USB-to-Serial cable. In theory, a USB port should be able to emulate the data transfer of a Serial port, but that's just theory... it might or might not work

-Chris

[naor52]

wich software every body track after the green light with the MPLAB or in the CMUcam2GUI software?

[CarpeDiem]

from the manual i have tried to hook up the camera to the rc and nothing happens can anyone tell me what i am doing wrong, and where does the pwm cable hook into the cmucam from the ttl board on the rc?

[Rombus]

Originally Posted by Chris Orimoto

We found that the servo movement and the camera power was dependent upon the backup battery... so you might want to change that before messing with the USB-to-Serial cable. In theory, a USB port should be able to emulate the data transfer of a Serial port, but that's just theory... it might or might not work

Yeah, Hopefully they got it working today, I remember last year we had the same issue with the camera and the battery not being charged. I know that some stuff can be picky with a usb/serial interface, but im hoping the rc isnet one of them.

Thats the one problem with being a college student metor, i cant be up there every day to test this stuff, especially with gas prices the way they are, its about a 45 min drive for me to get to the school i mentor for
chances are they will have fixed this issue by the time im back up there