运动物体的压缩感知成像  

Compressed sensing imaging of moving objects

作　　者：乔蕊[1] 张锦华[1] 

机构地区：[1]周口师范学院计算机科学与技术学院,河南周口466001

出　　处：《量子电子学报》2015年第1期17-23,共7页Chinese Journal of Quantum Electronics

基　　金：河南省软科学研究计划项目(132400411253)

摘　　要：在使用压缩感知成像系统进行成像时,物体的运动会导致获取的物体图像分辨率下降.为了克服运动的干扰需要额外装置获取物体运动的参数,但在很多情况下无法获取物体的运动速度.针对这一问题,提出利用物体运动与散斑运动的相关性,采用散斑运动方式来抵消物体运动从而消除运动造成的分辨率下降.采用压缩感知算法利用多组不同散斑运动补偿情况下的参数获取不同的物体图像,最后利用物体运动和散斑运动的相关性判据参数,最终获得消除运动干扰的物体图像。算法求解中同时考虑了系统点扩散函数的影响,采用解卷积算法获取了消除运动和系统点扩散函数影响的物体图像。最后数字仿真结果证实了所提算法的有效性。When compressed sensing imaging system is used to imaging, the related motion between the object and imaging system will lead degradation of resolution of the object image. In order to overcome the interference of motion, additional device is employed to acquire the motion parameter of the object, but in many cases the parameter cannot be obtained. Aiming at this problem, it was proposed to use the correlation of object and speckle motion, and apply speckle motion method to offset the object motion interference. Compressed sensing algorithm parameters used multiple groups of different speckle motion parameters to recover the corresponding object image, and at last the correlation criterion parameters was applied to get the object image without motion interference. The influence of the system point spread function was considered in the algorithm, and the deconvolution method was utilized to obtain the object image without the motion and system point spread function influence. Finally, the digital simulation results confirmed the effectiveness of the proposed algorithm.

关 键 词：图像处理 运动模糊 压缩感知 点扩散函数 解卷积 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]