I noticed the OpenMV Cam M7 on Sparkfun earlier for $65. It looks like it’s more capable than the Pixy that a lot of teams have picked up, for essentially the same price. It has its own IDE, and has pins which could be used for a RTC or something. Sadly it doesn’t look like it runs OpenCV, but on the flipside it is all Python.
Do you guys think this could be a good alternative for teams looking to upgrade from a Pixy? Is it capable enough to discern multiple targets (like in Stronghold)? It might be useful for teams without students/mentors who are familiar with Linux/C++.
Our team will likely stick with the Jetson TX1, but I was wondering what CD thinks about it.
It’s nice that it comes with a variety of I/O. I noticed a number of teams using the Pixy used an Arduino as a sub processor to connect to the RIO as the Pixy didn’t have the IO they wanted.
Have a look over in this thread. The JeVois camera seems to be picking up some traction within the community. It runs openCV, is programmable in python and C++, is about the size of a pack of tick-tacks, and only runs $50. There is talk of some whitepapers coming soon (pre kickoff) on how to integrate it with GRIP and the roboRIO. If all of this really does come to fruition, I can see this becoming the new pixy.
First of all, recommend listening to a podcast discussing it here: http://embedded.fm/212
I have one and have been playing with it. The most obvious limitation is that it does not have an ethernet connection, so you’re really stuck to communicating with it via the USB port or one of the communication protocols it supports. One of those is CAN, which is pretty awesome, so (in theory), you shouldn’t have to deal with getting SPI or I2C going on the Roborio.
Having looked at the JeVois, the M7 that OP is asking about, and the Pixy… I’d pick the JeVois out of the three… it’s absurdly impressive.
We’re also currently beta testing a thing that might be out before the season and if it is then it might even best the JeVois depending on the price point - the software is definitely better and more FRC focused.
Also, OP - you’re like a mile from us - come visit our lab and we’ll be happy to talk about vision all day long. ([email protected] and set something up)
I’ve been experimenting with the OpenMV M7 for vision tracking this year but I’ve had some troubles with communication with the roborio. I was wondering if anyone had any suggestions, sample code, or the likes that might help. It would be much appreciated!
