In general, terrestrial laser scanners (TLSs) acquire point clouds with sufficient overlaps among scans to improve point density and to produce conjugate features to be used for point cloud registration. In most cases, however, the overlapped data are maintained unfiltered even after the registration is finished and point density exceeded a sufficient level. In the present study, a method is introduced to manage multi-scan data, in which overlapped points are detected and removed. To detect the overlapped points, each point in a lately scanned point cloud is examined if it is too closely located with any one in former scanned point cloud, then overlapped points are removed. The remaining points are merged into the former point cloud, and the process is sequentially applied to all scans. For fast detection of overlapped points, octree is utilized as a 3-dimensional indexing structure.