Adding multiple point clouds from single Realsense
I have a series of point clouds by moving one realsense camera. when i display them together it seems the orientation changes. is there any proper way to combine mutiple point clouds using sing realsense camera?
-
Hi Dewan8511 If you are using a single camera and are able to compile a C++ application then the RealSense SDK's rs-kinfu (KinectFusion) may suit your needs. Instead of combining multiple point clouds, it builds up a point cloud as the camera is moved around the scene by fusing frames. When you are satisfied with the amount of detail on the cloud, it can then be exported to a .ply point cloud file.
https://github.com/IntelRealSense/librealsense/tree/master/wrappers/opencv/kinfu
If you would prefer to combine multiple point clouds then you could perform an operation called an affine transform to set all point clouds to the same position and rotation in 3D space so that they are aligned. The RealSense SDK has an instruction called rs2_transform_point_to_point that is an affine transform.
https://github.com/IntelRealSense/librealsense/issues/5583
Point cloud libraries such as Open3D and PCL, for which the RealSense SDK has compatibility 'wrappers', can also combine - or 'stitch' - multiple point clouds through mechanisms such as ICP registration. Here is an example of a tutorial for Open3D:
http://www.open3d.org/docs/release/tutorial/pipelines/icp_registration.html
Please sign in to leave a comment.
Comments
1 comment