I know that opensfm constructs a sparse pointcloud and we can use opensfm library, reconstruction.json, sparse pointcloud to project the points onto undistorted(when given 360 images as input) images. But for ODM, we get output of dense pointcloud and geo referenced pointcloud. I have couple of questions in this case
How to get convert the geo referenced pointcloud to local referenced pointcloud? Is there a transformation matrix with which we can do this sort of conversion?
Is there any procedure to project the 3d point to project onto the undistorted or the direct 360 images?