Hi @RobotMesh. Do you have any example Python vision sensor program for IQ? I’ve got the basics working in Blockly but when I try to create something based on the example linked below, I get an error on the IQ screen that says:
EXCEPTION: 0xea
Name err
File 0x09 Line 1358
main:22 main
https://www.robotmesh.com/studio/5be40cf0c8f17a1f5799ca9b
I don’t believe the vision sensor is supported for IQ yet.
Then never mind I guess. Thought I’d read somewhere that the vision sensor wasn’t yet compatible for IQ.
@calvc01
Hmm, it definitely worked when I wrote it, and we don’t seem to have made any API changes… I’ll look into it on Monday.
@Xenon27
You might be remembering that it is not yet legal for the IQ competition this year.
I guess so. Sorry for any confusion
Cool, thanks. In the meantime, I’ll have a bit more of a play and see if I can narrow it down at all.
I tested a version without motors this morning and didn’t run into any problems. Any luck on your end?
Thanks for checking John. Will try again shortly.
Obviously something I’m doing! Will report back soon.
Deffo operator error! Not sure where I went wrong though but OK now.
Apologies for the false alarm and thanks for looking into it so quickly as well.
1 Like
So, all good now, lots of good stuff going on. Is the limit 4 objects per signature in the array? It doesn’t seem to see any more after 4 but the sensor itself seems to detect hundreds much like Pixy.
It’s an issue of communication bandwidth. It’s detecting more than four, but it only has time to report ~four. It will report the largest four of whichever signature you ask it for, though, so you could keep track of eight+ objects as long as there are no more than four objects per color signature. You would just have to cycle through asking it about different signatures.
2 Likes
Cool, thanks.
So is it the I2C bandwidth that is the limitation vs. using SPI on Arduino with a Pixy2?