Open AhmedSamara opened 8 years ago
@adwiii @tvbarnette999 @dmpage318 Can you guys try to work on a solution for this problem since lidar is currently stopped?
Using surf points to identify them as I was planning on doing in #442 because the edges of the blocks are so rough that the surf points are non-unique every time, we'd have to be able to look at every single block.
Since I'm re-working the vision anyway, I'm going to see if I can just identify the outline of blue objects and work from there.
https://gist.github.com/AhmedSamara/0ce6a836dd45e16b68ba
Can fuilford-bros upload the code they had too? Both of these look promising for arbitrary object tracking.
https://gist.github.com/adwiii/a12c523b0d675d25c45d
This is the code that @tvbarnette999 and I were working on. It takes whatever blocks it sees and attempts to coerce them into a rectangle and whether or not it can do that it displays the angle that it finds. The coercion is just to make the angle better.
The generic blocks do not have QR codes on them.
This is the method that we've been using for calculating the vector from the end-effector to the block, without that QR code we don't have a way to put the end effector in the correct place.
So we really don't have a way to handle this. The rest of the blocks should be possible to handle with Sean and I's senior design project, but that's a much higher risk area.
I'd hate to miss out on all of the points from what really should be a low-risk area.