A New Method for Boundary Reconstruction and Center Line Extraction Case Study: Street's Center Line Extraction

Points sampled from an object that contains information about the object. These points don’t have topological information or special arrangement and are merely a set of points with coordinates. One of the most popular approaches for reconstructing boundary and extracting the central axis, using Delaunay triangulation and Vorony diagram for sampled points. The main problem with this approach is high sensitivity to noise and distortion. Therefore it is necessary to perform a filtration step aimed at removing excess branches of central axis. However, the filtration method as a preprocessing step or post processing cause to topological and geometrical changes in central axis. In this study, a new method for reconstructing boundary and extracting the central axis will be presented. As a practical example, the algorithm presented in this study will be used to extract the street center lines. Conceptual structure and the results of this study indicate that the stability, flexibility, high accuracy, even when there is noise and distortion border.