Camera Issues -- So close yet so far.

[team66t-money]

Well my hat is off to our world champions.. congrats. But this year in our journey i have found problems. as some of you know our team used vision, for auto. one of the few teams who actually got it working. But for some reason FIRST did not ban green from the game our anywhere around it, i think that because they did not do this they pretty much ripped our chance of making it farther in the finals in our division and maybe further in the championship. in the first match of our division finals the announcer for curie had green hair and pretty much an all green shirt on our camera was programmed to find all shades of the vision green because the lights going on and off changes but anyways my team asked him politely to move until autonomous mode was over, he did move but not far enough the camera picked him up drove forward turned to him and pretty much tried attacking him, we had so much power behind our robot that it slammed into the side and finally our manipulator flipped up over the side and snapped the steel cable for our arm then the ref wanted to give us a penalty for slamming other robots when they come after us first. i know that there is nothing that can be done but, next year if FIRST uses the camera our in the future they need to ban that certain color from around the field because it could really hurt another team our it just might get some one hurt. so please go to first and let them know something must be done not to let this happen again. though there was a lot learned by using the vision but it come to be more of a disadvantage than anything, beside the national award that we received i think that FIRST also needs to reward the more advanced robots more points for all of the hard work i don't think what is available is enough. Once again i am sorry to not use a lot of punc. i don't like for some reason using it while i am typing but thank you and great season and see you guys next year.

[Eugenia Gabrielov]

I won't comment on your pronunciation to much extent, though it was hard to follow.

I find it interesting that you mention the green. There has been quite the good amount of discussion about it in the past, and it seems that people came to a general consensus that green was ok. I wish you wrent' forced to expereince this, but please try to look at it in the sense that now that you understand your robot's ability to put a good deal of power into its actions and that you programmed it accurately, that in the future you may need to deal with adjusting this kind of programming on the fly.

If that announcer hadn't been standing there, something else green would have. I wish you luck in the future, and I hope you won't look at this as a cause for you not continuing to a certain level. There are many more factors than a match with an issue in the effect of winning competition.

Sincerely,
Genia

[PURPLE!]

Our program was so close... then the lighting on the field came up. Wow... talk about inconsistent. By the time the robot found the green it would go up then cast a shadow on it because of the lighting changing the color. Okay, after a few rounds of loosing the chance to cap we get the numbers and then we find out that shadowed yellow also picks up the tan in the volunteer's shirts... by the time we figured all of this out we had no real time to actually DO the capping...

[Wetzel]

Originally Posted by PURPLE!

Our program was so close... then the lighting on the field came up. Wow... talk about inconsistent. By the time the robot found the green it would go up then cast a shadow on it because of the lighting changing the color. Okay, after a few rounds of loosing the chance to cap we get the numbers and then we find out that shadowed yellow also picks up the tan in the volunteer's shirts... by the time we figured all of this out we had no real time to actually DO the capping...

You guys came the closest when the only other team in Archimedes tried to cap the center goal and your vision tetras collided. That was heartbreaking!

I commend you for your efforts with the vision system.

Wetzel

[Quatitos]

Originally Posted by team66t-money

in the first match of our division finals the announcer for curie had green hair and pretty much an all green shirt on our camera was programmed to find all shades of the vision green because the lights going on and off changes but anyways my team asked him politely to move until autonomous mode was over, he did move but not far enough the camera picked him up drove forward turned to him and pretty much tried attacking him, we had so much power behind our robot that it slammed into the side and finally our manipulator flipped up over the side and snapped the steel cable for our arm then the ref wanted to give us a penalty for slamming other robots when they come after us first.

I got to personally witness this event from about 10 feet away and when I saw 66 slam into the wall I knew that something must have gone wrong. After the match I heard from one of our mentors what had happened and I felt really bad that this had to happen to you guys during the finals. You had an awesome bot and a great autonomous. Its too bad you had to have it happen then but the only thing you can do with the situation is to try and learn from it and get even better, which I am positive that you guys will pull off.

[Dave Flowerday]

Originally Posted by team66t-money

But for some reason FIRST did not ban green from the game our anywhere around it, i think that because they did not do this they pretty much ripped our chance of making it farther in the finals in our division and maybe further in the championship.

I'm sorry to hear that the setup around the field caused you trouble, but the situation you encountered was an issue that could have been anticipated and solved by your team. Even if the announcer had not been wearing green, there could have easily been green being worn by people in the stands, or by other team members standing near the field, or even other robots painted green.

[Phil 33]

I was watching that match from the side of the field. It was very unfortunate. However, I have to ask you one question: when it was clear something was wrong with your autonomous mode, why didn't your human player step off the pad and disable your robot?

next year if FIRST uses the camera our in the future they need to ban that certain color from around the field because it could really hurt another team our it just might get some one hurt. so please go to first and let them know something must be done not to let this happen again

I don't think it's fair to blame FIRST for this, nor do I think FIRST should have banned the color green around the field. They gave teams a very nice "out" this year by allowing them to disable their robots by having a human player step off the pad. Again, what happened in that match was very unfortunate, and I'm not trying to sound ungracious or unprofessional, but I don't think FIRST did anything wrong by not banning the color green around the field.

[Kit Gerhart]

Originally Posted by Phil 33

I was watching that match from the side of the field. It was very unfortunate. However, I have to ask you one question: when it was clear something was wrong with your autonomous mode, why didn't your human player step off the pad and disable your robot?

My guess is that the HP didn't step off the pad because they didn't want a 30 point penalty.

[Mike]

Although I agree with you to an extent, it was just another obstacle that you had to overcome.

[Biff]

As a television engineer I have worked with cameras for over 25 years. Glib comments about "it's just another item you have to work around" don't help. Our attempts at getting anything useful out of our camera system were utter failures.
I watched 66 at GLR I was impressed.
From reading the forums here they are one of only about 4 teams that had a camera system that worked at all. That's four teams out of about 1000. Other comments in the forums here lead me to believe that the numbers for calibrations were at best inconsistent.
Looking at the playing fields with the skills I have learned shading cameras lead me to conclude the only way to get a good set of calibration numbers was to put your robot out there and do it yourself. This was not accounted for.
In the rules, and question and answer system mentions about programming for interference from colored objects. It was also stated that deliberate color schemes and clothing that sole intent was to confuse the vision system were against the rules.
So bottom line hat's off to 66 and the other teams that went down the vision system road.
In my opinion the match should have been replayed as the announcers clothing, although maybe not deliberate, clearly interfered with a vision system.
Like last year First clearly had high hopes for some kind of robot vision. The teams that had the best luck with vision last year clearly limited what they used the IR system for.
Maybe in the future there can be time on the field set aside for Teams to do the calibrations for future vision systems. My suggestion would be lunch on Thursday, First come first serve.(Pun intended) After all whining with out offering a solution is not a GP thing to do.

[Steve W]

I was the announcer on Curie field. There was only one time that I was approached, it was during the playoffs. When I came over to your team to see how we could remedy the situation I was told that it was not me but a field reset crew member that had a green collar on his shirt. As the collar was the same color as the green tetra, I asked him to tuck it inside his T-shirt which he did immediately. The fact was that you did not even come out to play that game. I was also told by another member of your team later that you were locking onto the volunteer passes that they were wearing.

It seems that there was a little confusion on your team as to why your vision was not working. I do laud you for giving it a go but please don't attack others when you are not sure of why you had problems. If you ask anyone there, they would tell you that I would do whatever necessary to help teams and not be a hindrance. I WANT to see teams succeed.

[Dave Flowerday]

Originally Posted by Biff

Glib comments about "it's just another item you have to work around" don't help.

It's not a glib comment - other teams (mine included) anticipated this problem and successfully solved it. Perhaps I missed it, but I didn't see any questions on this forum of how to solve it or I would have explained what we do. Part of engineering is anticipating possible problems and incorporating solutions into the design. One suggestion was already made (point the camera at the ground). Our solution was to have the camera look only at the exact locations that tetras should have been found and use the virtual window function of the camera to further restrict it's view. We did not have a single problem with our vision system at the Championship. We successfully tracked the tetras and goals in every single match we played using the calibration values that FIRST provided. Since our software knew which spots the tetras were located in, we could intelligently decide which one to attempt to pick up, and in the cases where we knew that we couldn't pick up either of them (due to the robot's design and not having enough time) we instead would drive over to the autoloader so that it was ready to go when driver control started.

Other comments in the forums here lead me to believe that the numbers for calibrations were at best inconsistent. Looking at the playing fields with the skills I have learned shading cameras lead me to conclude the only way to get a good set of calibration numbers was to put your robot out there and do it yourself.

We used the provided numbers at the regionals and Championship and they worked perfectly. I remember hearing after Week 1 regionals that some teams had trouble with the numbers, but I haven't really heard anything like that since. Did other teams have trouble with the numbers in Atlanta?

It was also stated that deliberate color schemes and clothing that sole intent was to confuse the vision system were against the rules.
... In my opinion the match should have been replayed as the announcers clothing, although maybe not deliberate, clearly interfered with a vision system.

Are you suggesting that the announcer was deliberately trying to confuse the camera with his clothing? I'm sure you're not, but FIRST made it clear that they would not go so far as to restrict people from wearing green or anything like that.

[Jack Jones]

Originally Posted by team66t-money

…. in the first match of our division finals the announcer for curie had green hair and pretty much an all green shirt on our camera was programmed to find all shades of the vision green because the lights going on and off changes but anyways my team asked him politely to move until autonomous mode was over, he did move but not far enough the camera picked him up drove forward turned to him and pretty much tried attacking him, …

That’s not the way I heard or saw it. I was the ref on scorer’s side Blue end. After 66 slammed the wall late in the Qs, we removed anything green from the table behind me - that’s where the robot aimed, right at me, not at Steve. It did no good. The robot came my way again. We then thought it was the green shirt on the auto load attendant and asked him to vacate next time during autonomous. A short time later the head ref told me that 66 had feedback showing that their video targeted my badge (maybe in concert with my hanging Thunder Chicken). That made perfect sense; it always works to blame the ref.

The odds of capping the center goal with a vision tetra were slim to none. As evidenced by that fact that nobody pulled it off. If we re-played every match that went without success, we’d still be trying to finish up the regionals.

[Biff]

Originally Posted by Dave Flowerday

It's not a glib comment - other teams (mine included) anticipated this problem and successfully solved it. Perhaps I missed it, but I didn't see any questions on this forum of how to solve it or I would have explained what we do. Part of engineering is anticipating possible problems and incorporating solutions into the design. One suggestion was already made (point the camera at the ground). Our solution was to have the camera look only at the exact locations that tetras should have been found and use the virtual window function of the camera to further restrict it's view. We did not have a single problem with our vision system at the Championship. We successfully tracked the tetras and goals in every single match we played using the calibration values that FIRST provided. Since our software knew which spots the tetras were located in, we could intelligently decide which one to attempt to pick up, and in the cases where we knew that we couldn't pick up either of them (due to the robot's design and not having enough time) we instead would drive over to the autoloader so that it was ready to go when driver control started.We used the provided numbers at the regionals and Championship and they worked perfectly. I remember hearing after Week 1 regionals that some teams had trouble with the numbers, but I haven't really heard anything like that since. Did other teams have trouble with the numbers in Atlanta?
Are you suggesting that the announcer was deliberately trying to confuse the camera with his clothing? I'm sure you're not, but FIRST made it clear that they would not go so far as to restrict people from wearing green or anything like that.

Dave, you hit he nail on the head with your description of how to use the vision system to get reliable results. By limiting where you are looking first and then tracking to a color in only those locations, it makes sense to me why the calibration numbers worked for your system. If your team would be gracious enough to share the code or a white paper more teams could come to an understanding of how to do it correctly. I was not suggesting the announcer was deliberately trying to cause interference, I was only pointing out that First did provide a rule to discourage intentional vision system confusion. Further reading in this thread indicates 66s' problems would have been solved if they had coded the way that your team did. Thanks for the insight Tom Cooper

[Goldeye]

To everyone who has made a comment about how teams should have simply 'focused' their view on given areas, please realize how difficult a task this really is.

Even under ideal conditions, configuring the camera system was an uncanny feat. In many cases, testing conditions were extremely far from the conditions on a FIRST field, and often extremely detrimental to the camera's function. The java utility could not find any suitable exposure values in my school, due to poor lighting. Still, we used substitute colors that the camera managed to see, and attempted to continue working with those.

There were many problems occuring with many parts of the camera routines. In addition to the troublesome values, controlling the robot at a speed appropriate to the camera readings was as difficult, robot-dependant task. Endless amounts of tweaks were needed for that. As problems came up, we solved them, for example, requiring a minimum size of any object the camera found. Still, by the end of the 6 weeks, we couldn't get the bot happily tracking anything.

Between all these problems and the need to get the robot tracking on near-ideal conditions, there's almost no time to create solutions for changing environmental conditions off of the field. I congratulate teams such as 111 for their ability and time to create such a thorough, reliable code, but for those making rash comments making it seem like an easy thing to do... If FIRST plans to provide such a tool, they should take every reasonable step to avoid conflicting with it. Last year with the IR, and now with the vision, they're not doing their part to allow the teams to focus on the game challenge rather than the world around it.