基于含参数光学变换的超分辨成像技术研究

We are inspired to extensively explore super-resolution imaging technology of the reconstruction of higher resolution image with several lower resolution images, by analyzing two kinds of classical super resolution imaging methods (structure light technology and ptychographic iterative engine technology). The proposed application project applies fractional transforms (fractional Fourier transform, gyrator transform and Fresnel diffraction) to construct a single-channel multiple-parameter measurement system for implementing the super resolution imaging. By changing the fractional order of optical transforms (or distance), recording the output intensity images of the same object with different parameters is achieved. Subsequently the parallel iterative amplitude-phase retrieval algorithm is designed to invert the light field of object plane, by using some intensity images measured at the output of the system. Here the amplitude data is from single side of imaging system (output plane) and is used to enhance the convergence speed of retrieval algorithm. Finally, the measurement structure mentioned above is introduced into optical microscopy system in order to obtain high resolution image. The experimental demonstration with various resolution images will be performed to determine the resolution of the proposed system. This project avoids the use of grating in structure light technology. Moreover the parallel computational method replaces serial calculation structure in the ptychographic iterative engine technology in order to receive a more accurate convergence result from the retrieval algorithm.

通过分析两类典型的超分辨成像方法(结构光技术和重叠关联迭代引擎技术)，启发我们深入开展以多幅低分辨图像进行高分辨重构的超分辨成像技术研究。本申请项目利用分数阶光学变换(包括分数傅里叶变换、gyrator变换和菲涅尔衍射)构造单通道多参数测量系统以实现超分辨成像。通过改变光学变换的分数阶(或距离参数)实现对同一目标的不同参数下的强度输出图像测量。然后利用系统输出平面所测量的多幅强度图像，设计并行迭代振幅-相位恢复算法反演目标平面光场信息，其中振幅数据均来自成像系统输出面一侧且用于加快恢复算法的收敛速度。最后将上述测量结构融入到光学显微镜系统中以实现高分辨率图像的获取，针对不同分辨率标准图像进行实验验证以确定所研究系统的分辨率。本项目避免了结构光技术中光栅的使用，同时利用并行计算方法代替重叠关联迭代引擎技术中的串行计算结构使得恢复算法获得更为精确的收敛结果。

通过分析两类典型的超分辨成像方法(结构光技术和重叠关联迭代引擎技术)，启发我们深入开展以多幅低分辨图像进行高分辨重构的超分辨成像技术研究。本申请项目利用分数阶光学变换(包括分数傅里叶变换、gyrator变换和菲涅尔衍射)构造单通道多参数测量系统以实现超分辨成像。通过改变光学变换的分数阶(或距离参数)实现对同一目标的不同参数下的强度输出图像测量。然后利用系统输出平面所测量的多幅强度图像，设计并行迭代振幅-相位恢复算法反演目标平面光场信息，其中振幅数据均来自成像系统输出面一侧且用于加快恢复算法的收敛速度。最后将上述测量结构融入到光学显微镜系统中以实现高分辨率图像的获取，针对不同分辨率标准图像进行实验验证以确定所研究系统的分辨率。我们提出一种镜像结构光技术改善超分辨成像系统测量过程，同时利用并行计算方法代替重叠关联迭代引擎技术中的串行计算结构使得恢复算法获得更为精确的收敛结果。此我们还研究了含参数变换在光学信息安全技术、光束整形中的应用。