Research On Point Cloud Data Processing And Reconstruction

With the rapid development of three-dimensional measurement technology, the scattered point cloud of entity models can be accessed easily. As an extremely important part of reverse engineering, the processing of mass point cloud and three-dimensional reconstruction has attracted a widespread attention. How to store, transmit and reconstruct the mass point cloud in the limited network bandwidth and computer resource has become one of the most active research points in the field of reverse engineering. After the intense study of the relative technology, two practical optimization algorithms are presented in this paper. The two algorithms focus on the lossless compression of point cloud data and three-dimensional reconstruction respectively. The main research content of this paper is as follows:1) A lossless compression algorithm for point cloud data is proposed. Firstly, the point cloud model is spilt into many equal sized surface patches. Over the points set of each patch, a minimal spanning tree is constructed and encoded in the depth-first order. Prediction processing is done along the tree structure after optimal linear predictor is constructed. Then both predicted and actual positions are broken into sign, exponent and mantissa. The mantissa is compressed by using arithmetic coding in the different contexts according to the relationship between sign and exponent. Experiments show that the lossless compression algorithm has a pretty good performance in time usage and compression ratio.2) A simple reconstruction algorithm is presented to accomplish point cloud data triangulation meshing. Firstly, the boundary of the point data is extracted. Then the points except boundary is simplified by the simplification algorithm based on fuzzy entropy iteration. Combinedwith the boundary of point cloud, the simplified 3D data points are constructed as triangular mesh using the edge extension criterion. And the criterion can make sure the triangular mesh is local optimum in theory. Texture mapping is realized after the triangulation and the 3D models are saved as standard format files. The proposed algorithm is proved to be highly efficient and exact.Besides, a complete point cloud data processing and reconstruction system has been accomplished based on the proposed lossless compression and reconstruction algorithms. The practical feasibility of the system is tested on both the well-known point cloud models and the point cloud models which are obtained by the grating projection measurement system in our laboratory. Experimental results show that the proposed algorithms are feasible and effective, the system is reliable and stable.