表面不规则场景的光场非均匀采集方法研究

项目名称： 表面不规则场景的光场非均匀采集方法研究

项目编号： No.61471150

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 无线电电子学、电信技术

项目作者： 张桦

作者单位： 杭州电子科技大学

项目金额： 83万元

中文摘要： 现有的光场采集系统采用稀疏均匀摄像机阵列，可完成结构性场景的三维采集和重建。但对于表面不规则场景，均匀的采集密度使平坦区域过采样，存储处理数据冗余，而复杂区域欠采样，细节无法精确重建。本项目针对全景稀疏动态光场采集系统中表面不规则场景的获取问题，研究光场的非均匀采集方法，通过分析场景表示维度间的稀疏冗余特性，构建场景表面属性、摄像机部署、重建质量之间的关系模型，提出重建场景高保真复原的评价机制，在保证重建质量最优的前提下，最大化降低采集数据冗余度。本项目的研究为采集系统中摄像机数量、空间位置以及方向角的部署提供有效的指导意见，为非结构化等新型动态光场系统的建立奠定了前期理论研究基础。

中文关键词： 光场采集系统；稀疏采集；非均匀采集；三维重建质量；客观质量评价

英文摘要： The plenoptic sampling system acquires and reconstuct the structral scene by using the sparse uniform camera array. However, because of the the irregular surface of the scene, the uniform sampling makes the flat region oversample and data redundancy, which also makes the complex one under sample and miss reconstruction. The project aims to investigate the non-uniform sampling in the panoramic sparse dynamic plenoptic sampling system, and analyze the information redundancy of the 3D scene. A relationship model between the surface properties, the camera deployment and the reconstruction quality is defined. An objective quality assessment of the 3D reconstruction is proposed which makes the best quality and the lowest data redundantcy. The research work of this project will provide the effective guidance for the number, position and the direction of the cameras in the sampling system which will be the foundation of the future unstructured dynamic plenoptic sampling system.

英文关键词： Plenoptic Sampling;Sparse Sampling;Non-uniform Sampling;3D Reconstruction Quality;Objective Quality Assessment