s-trinh

Any chance to confirm or disprove my findings?

Thanks for your answer. I will try to test it on my data when I will found some free time.

Thanks, subtracting `0.5` did the trick. But what about my Blender test case? I am using Blender to generate the images. I am retrieving the ground truth camera pose and...

Hi, Unfortunately, due to some very limited human resources no further development in this repo are planned from my knowledge. You should go with the manual way: https://visp-doc.inria.fr/doxygen/visp-daily/tutorial-tracking-mb-generic.html#mb_generic_advanced_cao In any...

@yangqingnian Could be some linear algebra issue? You should record all the necessary data + the intrinsic matrix in order to be able to reproduce your issue.

Maybe you can try to record the data when you have the issue with: - https://github.com/lagadic/visp/blob/699e1e69a05baaff7689108240d24762611d3882/example/device/framegrabber/saveRealSenseData.cpp - to read the data: https://github.com/lagadic/visp/blob/699e1e69a05baaff7689108240d24762611d3882/example/device/framegrabber/readRealSenseData.cpp

Not a tutorial code but I have added `tutorial-compare-auto-gamma.cpp` sample code that can be used to easily compare the results between the different auto Gamma methods. Can be used with...

Ok, I did not read correctly the issue. --- off-topic Unless I have missed something, but `undistortPoints` should not be involved in the process when calling `drawFrameAxes`? Because you go...

> The second one seems to be a refinement over the first one. Maybe if the fitting error after the second call is (way?) larger than the first one, warn...

Related links: - https://github.com/opencv/opencv/pull/17417 - https://github.com/opencv/opencv/issues/10270 --- Most likely the input data fall into a "degenerate case" for this method. One possible option would be IMO to try the various...