So my team is relatively new (2018 rookies) and didn’t use vision tracking last year. This year, we’re considering using a sensor to assist in placing the game pieces. The best sounding opportunity I’ve found is the Limelight, and I’m trying to figure out how viable it is of an option. If any teams have used/are using the Limelight for this year or any prior, is it worth the price? And how much easier is to use than other alternatives such as a sensor to follow the white line or one to pick up the reflection from the vision targets?
I’d say that the LL is worth the price if you’re a lower resource/skill team. However, if you have some programming ability, the JeVois has a lot of FRC precedent, and the FRCVision Raspberry Pi image (Kickoff Release of WPILib FRCVision Raspberry Pi image) is pretty easy to use.
1 Like
My team uses the LimeLight. It’s an excellent tool, really easy to use, and makes life a lot easier. However, the biggest drawback is the price. If your team has someone willing to work on vision, and write vision code, there are definitely cheaper options, but you will have to put more time and effort into in.
1 Like
We bought two Limelights and we have been using and loving them. For us, they are definitely worth the price. It helps to get rookie members interested and able to do vision without having to jump right into OpenCV. In my experience, the Limelight is one of the easiest tools to use. Within 2 minutes of our hatch mockup being built we had filtered out everything but the reflective strips. Within 2 more minutes we were getting the values in our code, and about 5 minutes later we were successfully tracking the mockup. Aside from the product itself, we’ve been having some issues regarding using a certain camera alongside the Limelight. As soon as I posted on the forum, @Hjelstrom responded and helped me get in contact with them. I’m honestly not sure which is best, the product or the service. I can praise the Limelight all day, so I’m gonna stop here before I make this any longer 
2 Likes
I agree. You are deciding between time or money. The Limelight will save a lot of time but will cost much more. Otherwise, I would suggest using the Raspberry Pi or Jetson to do your vision code.
Also, don’t try to chew more than you can swallow. Making sure you have time for driver practice and finished product testing is the key to having successful season. There are teams who have spent their entire season trying to get vision code to work and ended up failing.
If you are set on using vision code, I would suggest buying the limelight since it makes vision code so much easier. This would give you time to make sure that you have a solid robot that will be test ready well before build season ends. However, making sure you have a perfected robot should take precedence.
I like to think of the Limelight as a multi-year investment.
2 Likes
How experienced is your team with vision tracking? Your team is a few years older than mine so you probably would have leaned towards vision tracking more across those years, but did you go into coding the Limelight with previous tracking experience? And if so, how difficult would you think it would be for a team relatively new to this to code?
Considering the difference in price between this and other options, would you say that it’s worth the price? If not, what alternative would you recommend we go for to get the perfect balance of price and helpfulness?
We never got vision working until we got the Limelight (for the sake of figuring out what we were doing wrong, I eventually got it working, but that’s besides the point). We made a really simple subsystem for getting the Limelight values and then we made a simple PID command to track a target. The about 10 minute setup was with some experience with PID commands but not much vision experience.
1 Like
Ok. Thank you for helping out. I’ll bring these points to my team and hopefully it’ll help us make a decision.
1 Like