Open lxw919 opened 1 year ago
The SLAM part in the paper is just a demo to show that our depth is sufficiently consistent. I do not plan to release the code, because it is not clean. You could save the depth prediction, and treat it as RGB-D data to run orblsam.
Hi! I want to use your system to test the global 6-dimensional camera trajectory and dense 3D map output by monocular RGB video. Could you please share the code of your system combined with orbslam2? thanks!