Camera Success Stories, do they exist?

[Leav]

Hi everyone

I have a bit of experience with the camera but iv'e never gotten it to really work...

I have been asked to pull that issue forward this year and i'm stepping up to the plate.

I wanted to know if there are teams out there that had the camera working great, and if it's possible to maybe give us (and everybody else here!) some pointers regarding common pitfalls...

I was thinking along the lines of what lighting conditions you need or other stuff - just about any problem you overcame in your quest for your place in the CMUcam Hall Of Fame...

-Leav

[Tottanka]

With the correct min/max red/green/blue, i believe that you cant rack any colr, in all conditions.

Teams that had th camera working?

well there are many of them...

i know of 111, 217 in 2006's aim high

2007: 340, 1902, 111 and ofcourse, 910 =]

( there are many many more teams tat succeded)

i know we have, and therefore everyone can.

By the way, are you gonna be cback with 1578 this year?

[JBotAlan]

Note that vapor lights (don't know what they are actually called...pretty sure that's not it) flicker rapidly between red and green, and *will* make tracking impossible. Shop lights are bad news. Fluorescent does the same, I think. This is why we haven't had the cmucam do much for the past few years. I wish we had incandescent lighting in our shop...that's on my Christmas list. That, and a mechanically finished crab steer bot. Wow, I am weird.

Yes, it is possible to use the CMUcam with success. Will we do it? No. I'm not fighting with it--I'm using FALCON (FALCON home page). If you decide to plunge into this, realize that it is extremely beta right now. It does work, but I've put close to 20 hours into this system so far and it still doesn't work perfectly (troubleshooting a serial comms issue at the moment).

I am anti-CMUcam only because I've seen even the best of teams ditch it because it misbehaves.

Sorry if that comes across as anti-GP; I don't mean it that way. All I'm doing is stating fact.

JBot

[Leav]

Thanks for the help guys. that's valuable information for us.

Jbot, I will be PMing you...

Tottanka, No i'm not mentoring 1578, and as far as I know they will not be returning to FIRST (yeah very sad I know... we had great acheivements and had alot of potential).

Currently I'm mentoring team 2630 as much as I can while remembering that I still have to go to the Technion (university) sometimes...

see you at the Technion off season event tommorow!

-Leav

[David Fort]

I think the % of teams being really successful with the camera is (in 2007) pretty low - maybe 5 or 10 %, but here are some tips that helped our team:
1. Read the instructions.
2. Understand that the Java program configures the camera differently than Kevin Watson's example MicroChip code.
3. Take pictures of lots of stuff and understand what the camera really sees in various lighting conditions.
4. Make sure your target is in the linear range of the camera (i.e. not saturating).
5. Don't assume that the "numbers" they give you at a competition will work for you. Their camera configuration might be different than yours.

We were able to hang a few rings in Rack and Roll following that advice, unfortunately half the times we did things like crash into someone on the way there.