Hello there. Is there a way to calculate the hu moments from a lower level aspect? implementing it in micropython would be for our case too slow. Or do I have to make my own implementation for it in the firmware?
Thanks in advance.
Hi, this feature was requested long ago. But, I haven’t worked on it. Many of the blob computed results are the moments already. Additionally, you can get the partial x/y histograms of the blob. Which moments missing do you need?
Hello there again. I’ve been working on implementing Hu moments based on this paper. The paper details the calculation of “Moment invariants,” “normalized centered moments,” and finally the “Hu moments.” I also saw that you already calculate some of the moments and the centroids and I’ve been trying to implement the rest of them myself, but I feel like it’s going a bit over my head, haha.
If you want to add them all you’ll need to implement the calculations in the blob code: openmv/src/omv/imlib/blob.c at master · openmv/openmv (github.com)
Hi, excuse me, do you think implementing Hu moments to detect letters on an OpenMV H7 Plus would be efficient, or is a machine learning model better?