Study On The 3-D Scan Technology And 3-D Surface Reconstruction Algorithm

Three-dimensional (3-D) scan technology is a solid measurement. Compared with traditional measurement technology, it can finish the 3-D measurement of the points of complex object and is a high precision, quick or non-contact measurement. 3-D scan technology plays a very important role in many fields, such as national defence, industry, medicine, game & entertainment and so on. Meanwhile, the surface reconstruction is a significant section of reverse engineering. In this paper the following studies on 3-0 dimensional technology and surface reconstruction algorithm are carried through. (I) This paper presents the signification of 3-D scan technology, the main fields of application and developing status in the world. Several usual approaches of acquiring 3-0 information are reviewed. The principle of acquiring 3-D information methods based on computer vision (active and passive) is introduced. (2) The purpose of 3-0 scan is to reconstruct the model with the discrete points?coordinate of the object. Approximating the object surface by using many little triangular facets is one of the important reconstruction the algorithms. A new surface-triangulating algorithm based on the most similarity principle is given out. The speed of the new surface-triangulating algorithm is faster than the one of the traditional Delaunay triangulating algorithm. (3) Due to the virtues of the neural network, such as self-adapted learning, deducing and parallel computing, the multilayer perceptrons trained with the back-propagation algorithm and the radial-basis function neural network are employed to reconstruct 3-0 surface. In order to evaluate the two neural network surface reconstruction (NNSR) algorithms, many computer simulation experiments are carried out. In these experiments, we used a serial of points from a mathematically known surface in training these neural networks ~nd reconstructing the surface. These algorithms were compared with the local 2-0 lagrange surface reconstruction algorithm. The results show that the precision and speed of the two new algorithms are higher than the traditional algorithm. Specially, when the reconstructed surfaces are complex or incomplete or partially worn or damaged, the NNSR Algorithms are more effective.