Presented in the paper is an algorithm for generating a section curve by slicing a point cloud which may include tens of thousands of points. Although there have been couple of previous results, they are very sensitive on the density variations and local noises. In the paper, three technological requirements are identified; 1) dominant point sampling, 2) local vibration avoidance and 3) robustness on the density variations. To satisfy these requirements, we propose a new slicing algorithm based on a Virtual Edge (VE) approach, which is known to have a weakness for density variations. To complement the weakness of VE approach, Density Gauge (DG) sphere and point supplementing techniques have been developed and applied. The DG sphere and point supplementing techniques help to identify accurate virtual edges, which should be sliced to generate contour points. To compute the section curves from the contour points, the proposed algorithm employs the 2D Delaunay triangulation technique. The algorithm has been implemented and tested with various examples.