压缩感知理论在光场成像中的应用的研究.pdfVIP

压缩感知理论在光场成像中的应用研究 学生姓名：余临倩 专业班级：电信1002 导师姓名：覃亚丽 摘要：新兴的压缩感知理论，以远低于奈奎斯特采样定律的采样速率对信号进行 采样，且可根据少量的采样数据完美重构出原始图像。是信号处理领域的一大开 拓性成果。而光场成像作为一种计算成像，可收集光场中的四维信号，降低传感 器要求并获得任意景深的计算图像，是摄影领域的一大开拓性研究。但是光场成 像却会产生大量采样数据。因此，本文研究了如何利用压缩感知理论的特性，对 光场成像的前端数据采集进行非自适应性线性投影的方式以伪随机性质进行采 样，再根据重构算法对原始图像进行恢复。本文的主要工作是对上述思想的理论 研究，以及构造了基于微透镜阵列的光场成像系统图、用Matlab 软件对图像的 仿真验证算法。 关键词：压缩感知、光场成像、景深、非自适应、线性投影、伪随机、重构算法 Research on Compressive Sensing Applying for Optical Field Imaging Abstract: The burgeoning theory of Compressive Sensing can reduce the sampled data which is far less than the way sampled by the theory of Nyquist, also, it can perfectly reconstruct the original image through the few sampled-data. Therefore, it’s a great challenge and the inspiring news during the signal processing field. On the other hand, the optical imaging, as the computational imaging, can assemble the 4-dimensional signal in the light field, which will reduce the demand for the quality of the sensors. Moreover, it render obtain the computational images whatever the depth of the field they have. Thus, it’s a pioneering research in the photography. Because the large number of sampled-data will be acquired by the light field imaging, we research a method using the idea of compressive sensing into the optical field. When we collect the data during the optical field, we will exploit the non-adaptive linear projection which has the property of pseudo-random to obtain the less data. Then we can reconstruct the original images. So in the dissertation we will study the theoretical research upon the above-mentioned notion, and construct the system diagram of optical field imaging based on the