Field lighting interference with camera tracking

[aldenm]

My team has been having some issues with the field lighting interfering with the camera. We have an LED array around our camera that is being used to track the vision targets on the backboards. At the Autodesk Oregon Regional, we were shooting very accurately on the practice field without the field lighting, but whenever we played a match, the robot would only shoot a couple times and make only 1 or 2 baskets. I'm curious if any other teams are having this problem and how they are solving it. I'm on electrical on my team, so I don't know much about programming. I have heard that this has been mainly a programming issue, but I am curious if there are any types of LEDs that are effective, or any other tips with camera placement that could resolve this problem. Also, if anyone has programming tips, I can pass those on to our team's programmers.

[pfreivald]

Originally Posted by aldenm

My team has been having some issues with the field lighting interfering with the camera. We have an LED array around our camera that is being used to track the vision targets on the backboards. At the Autodesk Oregon Regional, we were shooting very accurately on the practice field without the field lighting, but whenever we played a match, the robot would only shoot a couple times and make only 1 or 2 baskets. I'm curious if any other teams are having this problem and how they are solving it. I'm on electrical on my team, so I don't know much about programming. I have heard that this has been mainly a programming issue, but I am curious if there are any types of LEDs that are effective, or any other tips with camera placement that could resolve this problem. Also, if anyone has programming tips, I can pass those on to our team's programmers.

Did you configure your camera using the tips in the vision white paper? Shining a bright light (we used an eWatt bulb at about 18") into your camera and then resetting the brightness makes the color of the ultra bright green LEDs that many people are using really pop. Without doing so, you get too much saturation.

[Al Skierkiewicz]

Did you take advantage of the calibration time on the real field to calibrate your camera with the field lighting?

[topgun]

When was the calibration time on the real field at Lake Superior Regional? We never heard about that.

[Brandon_L]

Originally Posted by topgun

When was the calibration time on the real field at Lake Superior Regional? We never heard about that.

Same at Chestnut Hill District.

[Deetman]

Originally Posted by Brandon_L

Same at Chestnut Hill District.

Seeing as the practice field in the pits was in the same room with the same lighting as the competition field, I doubt there was much of a calibration difference.

[billbo911]

You also may want to consider locking in the White Balance to the field lighting environment. If you leave the camera in "Auto", it will continually modify the image and ultimately loose the ability to detect the green from the LEDs.

[ChrisH]

Typically camera calibration will be at lunch on Practice day. Sometimes there will be a session right after the last practice matches. But the AV crew will turn off the lights when they leave and they usually want to leave right smartly after the last scheduled match.

Bottom line: ask the FTA or Field Supervisor when it will be for your event

[Al Skierkiewicz]

It was during lunch on practice day as announced during the driver's meeting Thursday morning.

[theawesome1730]

Team Driven 1730 had that problem. We went to the camera calibration, but between the our second and third matches, the lighting had changed ever so slightly, but enough to throw our camera off big time. Right now we are working on making a drivers console that has easy toggles and what not for changing settings while competing.

[Jared Russell]

Chestnut Hill had the interesting problem of having the video screen directly behind the Red Alliance baskets. As a result, we could quite easily see the screen through the smoked polycarbonate backboards (at least at certain angles). Whenever the camera panned over the MOE section of the bleachers, I prayed that we wouldn't start picking up their green jerseys

[fsgond]

We attempted to fix this problem this year. We have yet to be at a regional, but we have had great success with it in our shop. Mainly we filter out the visible spectrum. We have a large IR floodlight and a camera in a blacked out box with an IR filter on the front. The retro reflective tape reflects the IR very well, and all we see on the camera is the 4 glowing backboards. This made coding very easy, since there was very little noise to filter out. The only problem that we did have was the sun. We have scoped out the regionals that we are going to and the sun should not be a problem, so I think we are golden.

[neshera]

My understanding of retro-reflectivity (the material of the target rectangles) is that you only see them light up when the light source and the camera/eye are coaxial. This was our observation using infrared lighting on the robot, and observing sunlight coming in at a particular angle in the shop. We used Infrared because we thought there was not likely to be another infrared source behind the baskets. We tried several colors of LED lights around the robot, but ended up sticking with the IR.
This may be too late to help you this season; my apologies.
But - can you narrow your cameras field of vision, so it only picks up light being reflected back to your camera directly?

[Madison]

I was very happy to see that, at the Alamo Regional, the banks of lights on each of the field that were just above the baskets were turned off before qualification matches began on Friday morning. This was a welcome surprise and, even though we didn't end up using our automated distancing much during the event, it was nice to know that our team's hard work spent developing that capability wouldn't be washed away by something as silly as this.

[topgun]

Originally Posted by Al Skierkiewicz

It was during lunch on practice day as announced during the driver's meeting Thursday morning.

Thanks Al. The drive team never got that information to our programming team, and we needed that calibration time. Something to talk to the drive team about.