Mahalati Reza Nasiri, Gu Ruo Yu, Kahn Joseph M
E. L. Ginzton Laboratory and Department of Electrical Engineering Stanford University, Stanford, CA 94305, USA.
Opt Express. 2013 Jan 28;21(2):1656-68. doi: 10.1364/OE.21.001656.
We experimentally demonstrate endoscopic imaging through a multi-mode fiber (MMF) in which the number of resolvable image features approaches four times the number of spatial modes per polarization propagating in the fiber. In our method, a sequence of random field patterns is input to the fiber, generating a sequence of random intensity patterns at the output, which are used to sample an object. Reflected power values are returned through the fiber and linear optimization is used to reconstruct an image. The factor-of-four resolution enhancement is due to mixing of modes by the squaring inherent in field-to-intensity conversion. The incoherent point-spread function (PSF) at the center of the fiber output plane is an Airy disk equivalent to the coherent PSF of a conventional diffraction-limited imaging system having a numerical aperture twice that of the fiber. All previous methods for imaging through MMF can only resolve a number of features equal to the number of modes. Most of these methods use localized intensity patterns for sampling the object and use local image reconstruction.
我们通过实验证明了利用多模光纤（MMF）进行内窥镜成像，其中可分辨图像特征的数量接近在光纤中每个偏振传播的空间模式数量的四倍。在我们的方法中，一系列随机场图案被输入到光纤中，在输出端产生一系列随机强度图案，这些图案用于对物体进行采样。反射功率值通过光纤返回，并使用线性优化来重建图像。分辨率提高四倍的原因是场强转换中固有的平方运算使模式混合。光纤输出平面中心的非相干点扩散函数（PSF）是一个艾里斑，相当于数值孔径是光纤两倍的传统衍射极限成像系统的相干PSF。以前所有通过MMF成像的方法只能分辨与模式数量相等的特征数量。这些方法大多使用局部强度图案对物体进行采样，并使用局部图像重建。
