|
Hello!
Does anyone know a source where RGB-D SLAM, which is used in rtabmap, is explained, from the images to the trajectory and pointcloud, in an easy way? I read many papers, but my image processing skills are very bad, so I don't understand much. Or can someone explain it to me briefly? The explanation has not to be very scientific, just to get a overview if you want use it.
What I understand is that the first step is the feature detection on a new rgb image, than a comparison of the detected features between the new and the old image. Then, the detected keypoints depth data are taken from the 3D image and processed in the RANSAC step?
I would be very grateful.
best regards
Lucas
|