摘要
介绍了一种基于压缩传感理论的THz成像方式,该理论突破了传统奈奎斯特采样定律的限制,可以以远小于图像总像素数的少量测量值来恢复出原始图像,从而很大程度上缩短了成像时间、提高了内存资源的利用率。通过对该理论的分析可知,变换基、测量矩阵和重构算法是该理论的三个重要因素,通过一系列仿真试验,分别对不同的变换基、测量矩阵和重构算法作了对比,根据对比结果,得到了最适合于文中被成像物体的变换基、测量矩阵以及重构算法。最后,利用离轴抛面镜代替聚乙烯透镜搭建了成像系统,得到了初步的试验结果。
A THz imaging system based on the theory of compressed sensing was described, the theory break through the limit of Nyquist sampling law, it can reconstruct the original image by only a small number of measurements, so that it can shorten the imaging time and improve the utilization of memory resources to a large extend. Through the analysis of the theory, the three important factor of the theory are the transform matrix, the measurement matrix, and the reconstruction algorithm. A series of simulation was done to contrast different transform matrixes, different measurement matrixes and different reconstruction algorithms respectively, and according to contrast results, the most suitable transform matrix, measurement matrix and reconstruction algorithm for the image object were decided. Finally, the imaging system was set up by off-axis throw lens instead of polyethylene lens, and preliminary experimental results were made by this experiment system.
出处
《红外与激光工程》
EI
CSCD
北大核心
2013年第6期1523-1527,共5页
Infrared and Laser Engineering
基金
国家973计划(2007CB310408)
国家自然科学基金(11004140、60977009)
北京市教育委员会科技面上项目(11224010011)
关键词
压缩传感
THZ
变换基
测量矩阵
重构算法
compressed sensing
THz
transform matrix
measurement matrix
reconstruction algorithm
