Surface Reconstruction of Defective Point Clouds Based on Dual Off-Set Gradient Functions

摘要: Surface reconstruction is an interesting and challenging task in extensively applied fields including rapid prototype manufacturing, computer vision, virtual reality aided design (CAD). A typical procedure begins with scanning, which the point data are sampled from physical objects by digitizing measurement systems (such as laser-range scanners hand-held digitizers). And then, generated a smooth, water-tight proper resulting surface suitable method. In industry most difficulty comes defective samples that subject to noise, holes overlapping regions. The often unavoidable due sampling inaccuracy, scan mis-registration accessibility constraints of scanning device. They make existing methods not practical for engineering application because oriented or neighbour information points, highly based on, hard evaluate. For instance, many rely on consistent normals, pose demand triangular meshes data. However, confuse point’s relationship, some jagged, self-intersect regions could exist corresponding mesh estimation normals becomes ill-posed problem. Only few need such specific information, but they have resort complex time-consuming steps, like re-sampling, distance-computing, mesh-smooth deformable models. Even if these can generate surface, reasonableness fitting guaranteed. fact, issues, especially “badscanning” data, lead long time, massive manual work poor model quality. Given challenges, this paper propose novel method takes input clouds without any output smooth surface. main idea (1) technique implicit function, convenient guarantee result; (2) approach indirect, two off-set surfaces best fit instead direct approximation. As shown Fig.1 (1D situation simple expression),