In an example embodiment, a process may select high density points from a point cloud. The process may create one or more clusters from the high density points and identify a circular cluster from the created clusters. The process may identify which points in the circular cluster are inner edge points and determine a center of an ellipse that fits the inner edge points. The process may define a search space utilizing the center of the ellipse. The process may determine the estimated x, y, and z coordinates for the position of the scanner in the search space utilizing a non-linear least square solver with different combinations of a relationship that is true for any pair of points of the cluster. An application may utilize the determined position with an object/file format (e.g., LSA format) to generate a high resolution 3D mesh of a scene.