Evaluation and Optimization of Laser Scan Data

The digitalization of objects from the real world is of great importance, e. g., in many multimedia applications, in 3D computer graphics, and in industrial measurement. In particular, methods of optical measurement are frequently used in this context. This type of 3D scanning yields a point cloud that may contain errors due to system specific characteristics. Considering the example of an optical laser scanner, the data acquisition process is analyzed in order to identify system parameters that can be used for a qualitative evaluation. Several methods for quantitative analysis of the data are discussed that are partially based on a B-spline approximation of points in the point cloud. Furthermore, methods are presented that are particularly well qualified to detect geometrical characteristics such as curvature and edges. These characteristics and the approximated quality of single points are used to reconstruct a revised and optimized point cloud. Finally, the effectiveness of the proposed methods is evaluated based on two exemplary point cloud data sets.