Processing an image on STM32 - C

Question

I am doing some blob analysis on an embedded STM32H7 platform using C.

What do you think would be best way to accomplish this? Would it be possible to port openCV to STM32 and use those functions maybe? Let say I just want to use adaptive thresholding, would it be better to just write my own code for that or use some library that is already there?

Just to point out that I am obligated to use C. I know about some Python libraries for STM32 but I must use C for this project.

Thank you!

Answer 1

Writing your own functions is trivial. I have done minor on the subject and we had about 8 weeks to do it.

Porting OpenCV is not a good idea. It uses dynamic memory allocation everywhere which makes conservative use impossible and is written in C++.

Adaptive tresholding would take max a week of effort.

Answer 2

You shouldn`t try to implement everything from zero

Try the opensource implementation first. If you can think about sth, that means some 1 else already did it. Someone else has done it.

See this link for details.

http://shervinemami.info/embeddedVision.html