Search code examples
3dcomputer-visionpoint-cloudsopen3d

Can someone explain the meaning of x, y and z coordinates in open3d pointcloud?

I am working with a depth camera. The camera outputs a depth map which is 500x500 array. Each of the elements in this array represent the depth from the camera.

Now I convert this depth map to a pointcloud using open3d library. I know that each of these point have an x,y and z coordinate. I want to know out of these three, which one represent the depth from the camera? Also, are the x and y cooridnates here are in camera frame?

Solution

  • The depth map would provide the depth as the value indexed at [u,v] in the map, based on what is rendered in the depth image.

    A point in the point cloud would be:

    (u,v,depth[u+v*depth.w])

    According to Open3D documentation the Z value represents the depth as a value of the equation (d/depth_scale)