Co-registration of Surfaces by 3D Least Squares Matching

A method for the automatic co-registration of 3D surfaces is presented. The method utilizes the mathematical model of Least Squares 2D image matching and extends it for solving the 3D surface matching problem. The transformation parameters of the search surfaces are estimated with respect to a template surface. The solution is achieved when the sum of the squares of the 3D spatial (Euclidean) distances between the surfaces are minimized. The parameter estimation is achieved using the Generalized Gauss-Markov model. Execution level implementation details are given. Apart from the co-registration of the point clouds generated from spaceborne, airborne and terrestrial sensors and techniques, the proposed method is also useful for change detection, 3D comparison, and quality assessment tasks. Experiments using terrain data examples show the capabilities of the method. Introduction With the availability of the various sensors and automated methods, the production of large numbers of point clouds is no longer particularly notable. In many cases, the object of interest is covered by a number of point clouds, which are referenced in different spatial or temporal datums. Therefore, the issue of co-registration of point clouds (or surfaces) is an essential topic in 3D modeling. In terrestrial laser scanning practice, special targets provided by the vendors, e.g., Zoller Fröhlich, Leica, and Riegl, are mostly used for the co-registration of point clouds. However, such a strategy has several deficiencies with respect to fieldwork time, personnel and equipment costs, and accuracy. In a recent study, Sternberg et al. (2004) reported that registration and geodetic measurements comprise 10 to 20 percent of the total project time. In another study, a collapsed 1,000-car parking garage was documented in order to assess the damage and structural soundness of the building. The laser scanning took three days, while the conventional survey of the control points required two days (Greaves, 2005). In a project conducted by our research group at Pinchango Alto (Lambers et al., 2007), two persons set the targets in the field and measured them using the real-time kinematic GPS technique in one and one-half days. As well as fieldwork time, accuracy is another important concern. The target-based registration methods may not exploit the full accuracy potential of the data. The geodetic measurements naturally introduce some error, which might exceed the internal error of the scanner instrument. In PHOTOGRAMMETRIC ENGINEER ING & REMOTE SENS ING March 2010 307 Department of Information Technologies, Isik University, 34980 Sile, Istanbul, Turkey, and formerly with the Institute of Geodesy and Photogrammetry, ETH Zurich, Switzerland ([email protected]). Photogrammetric Engineering & Remote Sensing Vol. 76, No. 3, March 2010, pp. 307–318. 0099-1112/10/7603–0307/$3.00/0 © 2010 American Society for Photogrammetry and Remote Sensing Co-registration of Surfaces by 3D Least Squares Matching