Laser scanning is one of the main methodologies that is being widely used by surveying engineers for geometric documentation of sights and objects. Laser scanners provide large point clouds that depict the surface of the area or object under investigation. Modern scanners are able to create point clouds of such a density that in many cases the production of the 3D model is not necessary. Basic operations, such as cross sections and distance measurements, can be made directly to the point cloud.

On the other hand, that big amount of datasets requires the appropriate tools of elaboration and extraction of the significant information. In this paper, we made an approach to distinguish, among each other, objects that exist in the same scanning scene. The detection of the points that are on objects' edges provide firstly the knowledge of the space that objects occupy and thus we are able to separate them from their surrounding space and secondly provide the user with the convenience of calculating their geometric parameters. The suggested method makes use of the k nearest neighbors of a point and the closeness of those neighbors to the point of interest