I would like to use my Zed 2 camera in a fixed position, i.e., stationary mode, to extract some geometric features of the passing objects in front of the camera. I have an instance segmentation model that detects the target objects and provides a mask which is going to be applied to the 3D point clouds coming from each frame. Is it possible to extract the point cloud associated to one single frame in a none incremental way? Do you have any suggestion for this objective?
When you retrieve the point cloud with
retrieveImage it gets the pointcloud of the current frame. Wh do you mean “on a none incremental way”?
If you use our spatial mapping module however, you’ll be able to retrieve a fused point cloud from several frames. But not with the usual