The Design Of A Processing Platform For3D Point Clouds

With the development of3D digital geometry technology, the RE (reverse engineering) is widely used into the modern manufacturing. The processing on3D point clouds is the core content of RE, and it is significant to independently develop an advanced3D point cloud data processing platform based on the domestic and foreign advanced technology.The purpose of this paper is to develop a3D point cloud data processing platform with perfect functions, robust and reliable performances and use manner in accordance with China’s designers’ use habit, as well as providing integrated interfaces for the subsequent applications. Based on the3D point cloud processing functions and interfaces provided by open source software PCL(Point Cloud Library), this paper employs the object oriented programming and OpenGL graphics display technology and has completed a3D point cloud data processing software platform-systerm. The main functions of the3D point cloud data processing have been implemented.Firstly, this paper introduces the basic knowledge of PCL and the software architecture, mainly including the data structure of the point cloud, the parameter estimation algorithm, the description and extraction of the feature, point cloud segmentation and so on. Secondly, this paper discusses the overall design of the processing platform, the basic methods and data processing flow adopted in the platform design. Thirdly, this paper studies the algorithm process and implementation based on PCL modules, such as the filter module, registration module, and surface module in detail. Finally, this paper adopts several specific cases to verify those functions which are realized in the platform. The developed platform realizes the following functions:the import and render display based on OpenGL for3D point cloud data, removing noise points in point clouds based on the methods of Voxel grid and statistical analysis technology, combining the point cloud data from multiple-views into a same coordinate system in order to complete the registration of multiple-view point clouds, repairing the holes in the point cloud or simplifying the data model by using the resampling technology, triangular mesh processing and storage for, the completed point clouds, etc.The developed3D point cloud data processing platform is developed through programming and debugging, and tested by a number of examples, which demonstrates the developed platform has realized the above functions, and verifies the correctness and practicability of the proposed method...