Abstract:

The reconstruction of such features as protrusion, pocket, extrusion and revolution depends on the re-construction the section profiles describing the feature shapes. As being dense and out-of-order, the collected data points cannot be directly parameterized and fitted by sp ine curve after the profile sketch p lane is determined. This paper addresses the p rofile reconstruction based on dense data points. In addition, a novel app roach is proposed. By this app roach, the“thick boundary”of the section p rofile is firstly extracted in the p rofile sketch p lane, and the obtained data points are then thinned and parameterized, thus imp lementing the reconstruction of the section profilerep resented by the B-spline. Experimental results prove that the proposed method can effectively reconstruct the profileswith the above-mentioned features.

Key words: reverse engineering, profile reconstruction, boundary extraction, data point thinning