| Volumetric scanning display technology can render a 3D image with real physical depth, and allow many viewers watch in all directions with eye naked. It is an important research direction of 3D display and has been increasingly used, thus to study and design an efficient data sources generating scheme is of great significance.Based on "3D volumetric scanning display system" subject, this thesis realizes a program of 3D sources acquisition and processing for different objects. To improve the data sources storage, transmission efficiency and display effect, the paper proposes a point cloud compression method which has broad applicability.After experimenting on many 3D sources acquisition ways, the thesis decides to choose 3ds Max modeling and Kinect scanning to obtain virtual objects data and real objects data. Read and process these files to get the point cloud data, then do cylindrical coordinate conversion and data mapping to create system data sources. Meanwhile, the dissertation presents a voxel homogenization algorithm based on density screening, to fix display distortion caused by uneven density.Based on analysis of many compression algorithms, this paper puts forward areal density based compression algorithm and voxel center based compression algorithm, and compare them with random compression algorithm and original octree algorithm. The article then expatiates on the evaluation criteria of data sources for volumetric scanning applications, including sampling ratio, coding ratio and homogeneous degree, voxel center based algorithm is selected by evaluation for the ultimate program.The latest open-source library PCL is introduced to achieve the compression algorithm, and the calculation method of compression accuracy is illustrated. At last, by simulation and experiment, the efficiency and accuracy of scheme is verified. |
PDF Full Text Request