Open GuillaumeHauss opened 6 years ago
Well, it seems that this function is handling the major magic step to get the ID of the marker
int FiducidalMarkers::hammDistMarker(Mat bits)
in the same arucodifmarker.cpp file. Now, there is some kind of mask set in the function, that has explicitly 5 columns, used in the computation of the Hammind distance. After some digging in the opencv library, I found this file which houses the function
int normHamming(const uchar* a, int n, int cellSize)
It also uses some predefined tables, but I cannot understand their origin. Do someone has a thought on that?
Hi all,
As mentioned in my previous issue, I'm looking to adapt the Aruco Detector to make it able to detect and identify markers with N-bits resolution, N in [4; 7]. So far, the bit resolution is hardcoded in the function detect in the file
arucofidmarkers.cpp
, as shown in the following snippet of codeMy initial guess was to change every
7
inN
, every6
inN-1
and every5
inN-2
, after addingN
as a parameter of the function. But this wasn't working. Does anyone has a better approach in mind?I'll keep digging and keep you posted if I find a workaround :)
Thank you very much for your help