Abstract: Method that merges two or more point clouds captured from a scene, eliminates redundant points, and retains points that best represent the scene. The method may generally include a detection step, which locates points from different clouds that are close and thus potentially redundant, followed by a selection step that identifies preferred points. Clouds may be represented as range images, which may simplify both steps. Closeness testing may be optimized by dividing range images into tiles and testing tile bounding volumes for intersections between clouds. Selection of preferred points may incorporate user input, or it may be fully or partially automated. User selection may be performed using 2D drawing tools on range images to identify images with preferred views of a scene. Automated selection may assign a quality measure to points based for example on the surface resolution of each point cloud scan at overlapping points.

Abstract: Method that merges two or more point clouds captured from a scene, eliminates redundant points, and retains points that best represent the scene. The method may generally include a detection step, which locates points from different clouds that are close and thus potentially redundant, followed by a selection step that identifies preferred points. Clouds may be represented as range images, which may simplify both steps. Closeness testing may be optimized by dividing range images into tiles and testing tile bounding volumes for intersections between clouds. Selection of preferred points may incorporate user input, or it may be fully or partially automated. User selection may be performed using 2D drawing tools on range images to identify images with preferred views of a scene. Automated selection may assign a quality measure to points based for example on the surface resolution of each point cloud scan at overlapping points.

Abstract: Method that merges two or more point clouds captured from a scene, eliminates redundant points, and retains points that best represent the scene. The method may generally include a detection step, which locates points from different clouds that are close and thus potentially redundant, followed by a selection step that identifies preferred points. Clouds may be represented as range images, which may simplify both steps. Closeness testing may be optimized by dividing range images into tiles and testing tile bounding volumes for intersections between clouds. Selection of preferred points may incorporate user input, or it may be fully or partially automated. User selection may be performed using 2D drawing tools on range images to identify images with preferred views of a scene. Automated selection may assign a quality measure to points based for example on the surface resolution of each point cloud scan at overlapping points.