Hi! Thank you for giving this good tutorial!
Here I found a problem. When I tried the triangulation of the Python version and C++ version, I found that the output was not the same. The Python version will give a negative depth triangulation result.
After I add the camera intrinsic matrix to the function "cv.recoverPose", the triangulation works well.
The code for this line
https://github.com/mint-lab/3dv_tutorial/blob/f295338ae4a8c28bf7da0d3f2dea1506be652ef1/examples/triangulation.py#L14C43-L14C43
_, R, t, _ = cv.recoverPose(E, pts0, pts1)
should be:
_, R, t, _ = cv.recoverPose(E, pts0, pts1, K)
Hope this will be helpful.
Hi! Thank you for giving this good tutorial!
Here I found a problem. When I tried the triangulation of the Python version and C++ version, I found that the output was not the same. The Python version will give a negative depth triangulation result.
After I add the camera intrinsic matrix to the function "cv.recoverPose", the triangulation works well.
The code for this line
https://github.com/mint-lab/3dv_tutorial/blob/f295338ae4a8c28bf7da0d3f2dea1506be652ef1/examples/triangulation.py#L14C43-L14C43
should be:
Hope this will be helpful.