Page 1 of 1

FrameBuf access

Posted: Fri Oct 13, 2017 9:27 am
by javierf
Hi all,

I'm developing a computer vision sdk and I'd like to start by doing basic image/matrix manipulation operations (similar to what the cv::Mat class does in OpenCV). In order to do so, I think that the best way would be to work directly on top of the framebuf class. However, since OpenMV sill uses MicroPython 1.8.X and not 1.9.X (where the imagebuff class is very well defined https://docs.micropython.org/en/latest/ ... e-framebuf )I wonder if this is possible? Any ideas :roll: ?

Re: FrameBuf access

Posted: Fri Oct 13, 2017 9:58 am
by kwagyeman
The next version of the firmware is using MP v1.9.2.

That said, I think it would be better to use our image class and work in top of that... We have a lot of code written and will gladly accept PRs.

Re: FrameBuf access

Posted: Fri Oct 13, 2017 10:03 am
by javierf
I 100% agree. There's no point in re-inventing the wheel. Do you know when the next firmware (with 1.9.2) will be released? Is there a beta version I could start working on that gives access to Framebuf? Thank you

Re: FrameBuf access

Posted: Fri Oct 13, 2017 11:27 am
by kwagyeman
Pull the current github account and you can start working. Um, the MP framebuf class just exists for letting you write to a tiny LCD screen. So, if you're interested in making our image class work with that I'd just write a method to convert between the two formats.

Let me know if you actually want to commit a large number of methods or whatnot and I'll walk you through my ideas about things. There's a lot of work that needs to be done on more basic image manipulation stuff. But, I don't have the time and it doesn't demo well so I don't work on it since it won't sell cameras.

Re: FrameBuf access

Posted: Fri Oct 13, 2017 2:14 pm
by javierf
Thank you. Is there any documentation or steps I could follow to build from source the firmware?

Re: FrameBuf access

Posted: Sat Oct 14, 2017 10:50 am
by kwagyeman