Three-dimensional object reconstruction from range images

This research work focuses on reconstructing surface models from range images of three-dimensional (3D) objects. This problem of surface reconstruction from range images is very important in the design of any 3D computer vision system and normally consists of several phases of data processing. This work presents novel algorithms to provide efficient solutions used in the following four key phases: data acquisition, data registration from multiple views, data integration, and surface reconstruction.; A major critical issue in range data acquisition concerns with the data accuracy. Factors that affect accuracy are carefully considered for a specific range data acquisition system, the 4-Dimensional Imager system. The surface point distance errors, due to field curvature and depth of field of ranging system, have been studied.{09}Two model-based approaches, the area model (AM) and line model (LM), are proposed to model the systematical errors at different distances and orientation angles. Error lookup table is built with these models and used to reduce the systematical error of acquired data. Up to half of the systematical errors can be reduced using the AM method and almost all the systematical errors with the LM method.; Range data for large object that are acquired from different acquisition passes require data registration. Accurate data registration from multiple views for object reconstruction is still a difficult R&D problem. With the given system (4DI system), a registration method involving geometric transformation is presented for reconstructing large object from multiple views with high accuracy. The algorithm provides highly precise numerical values for the elements of the transformation matrices. These values are determined by the specific system parameters estimated through carefully designed tests. Experimental results show that the algorithm can reconstruct a large object within the required industrial accuracy level of approximately 3 mil (0.076 mm) tolerance.; Overlapping data need to be removed after data registration process as they destroy the surface singularity, reduce reconstruction accuracy, and introduce difficulties for further processing. To remove the overlapping data efficiently, a novel data integration approach which is based on predefined criteria and the nearest neighbor searching is developed. Our method manipulates surface points directly, thus, it provides a simple and fast way for overlap removal and has the lowest computational complexity, O(MN), when compared with the other methods. The approach is successfully applied to various range data sets of objects with different geometrical shapes. Experimental results show that the algorithm can significantly improve the accuracy of the reconstruction result.; To reliably and accurately reconstruct object surface from a set of 3D spatial points, an algorithm for meshing, a kind of incremental-based surface reconstruction method, is proposed. It considers the shape change at the boundary of mesh area and forces mesh area to propagate according to a priority driven based strategy. New criteria for triangulation are developed to construct triangle at each step of mesh growing. Unlike other incremental-based methods, the algorithm has no assumptions on the geometry or topology of input data and places no limitation on the input data acquired with non uniform sampling. The efficiency of the proposed algorithm is demonstrated by both small and large data sets.