Formulation of point cloud generation by open3d

Question

I am working on 3d photography, and in order to generate point cloud I am using (pcd = o3d.geometry.PointCloud.create_from_rgbd_image(rgbd, intrinsic_2, extrinsic)).

According to the explanation of this link ( http://www.open3d.org/docs/0.6.0/python_api/open3d.geometry.create_point_cloud_from_rgbd_image.html#open3d.geometry.create_point_cloud_from_rgbd_image ) first step is to convert 2d data to 3d data by using following formulation:

z = d / depth_scale

x = (u - cx) * z / fx

y = (v - cy) * z / fy

And then???! what is the role of Extrinsic matrix? I can not undrestand! can anyone help me?

Answer 1

The extrinsic matrix is to render the point cloud from the camera pose. If there is just one pointcloud it doesn't really matter, but if you want to stick multiple pointclouds together they must come with their extrinsic matrix.