光场成像的轴向超分辨率方法研究

项目名称： 光场成像的轴向超分辨率方法研究

项目编号： No.61501370

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 无线电电子学、电信技术

项目作者： 肖照林

作者单位： 西安理工大学

项目金额： 21万元

中文摘要： 光场成像具有先拍摄后聚焦的可交互式重聚焦能力，轴向分辨率是衡量其数字重聚焦性能的重要指标。现有光场成像系统存在成像分辨率折衷的瓶颈，导致光场图像质量不高，且重聚焦能力不足，从而限制了光场应用的拓展。本项目拟研究轴向分辨率模型及数据处理阶段的轴向超分辨率方法。主要内容包括：1）针对双平面模型光场成像系统，研究光场采样特性、数据处理过程及成像方法对轴向分辨率的影响，构建非线性轴向分辨率模型，并探索轴向分辨率特性的量化评价方法，以确定影响轴向分辨率的关键因素；2）在不改变硬件系统设计的前提下，研究高精度光场数据解耦及轴向超分辨率图像生成方法，并进一步探寻光场图像质量的优化方法。本项目属于计算摄影学领域的前沿理论与方法研究，旨在解决双平面光场成像系统轴向分辨率不足的问题，并为突破光场成像分辨率折衷瓶颈提供新的思路。研究成果对丰富光场成像理论、提升光场图像质量、拓展光场成像应用具有重要意义。

中文关键词： 计算摄影；光场成像；数字重聚焦；超分辨率；轴向分辨率

英文摘要： Light field imaging is well known for its “focus-after-shoot ” interactive image generation ability. The axial resolution is an important measurement of the digital refocusing capability. At present, one bottle-neck problem is the tradeoff between its spatial and angular sampling. This problem significantly deteriorates the image resolution and axial resolution, which limit the applications of light field imaging. So, we attempt to solve the low axial resolution problem by researching on the resolution model and axial super-resolution algorithm of the plenoptic imaging system. The major research contents include: 1) For the two-paralleled-plane light field imaging system, we tend to research on its sampling characteristics, data processing and imaging algorithm. Then, we would construct a non-linear axial resolution model, and explore its quantization and verification method. 2) Without changing the system optical design, we tend to research on light field data decoupling and axial super-resolution algorithm. To enhance the image quality, we will further explore the light field imaging optimization methods. This project belongs to light field imaging theory and approach study in computational photography. The light field axial super resolution is one of key problems to break through the resolution tradeoff in the data processing stage. Thus, the research contents have theoretical significance to light field imaging, and many potential applications can be benefited from this work.

英文关键词： Computational photography;Light field imaging;Digital refocusing;Super resolution;Axial resolution