Kumar Abhishek, Georgiev Stefan, Salas Matthias, Leitgeb Rainer A
Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Austria.
Wavesense Engineering GmbH, Vienna, Austria.
Biomed Opt Express. 2021 Feb 23;12(3):1577-1592. doi: 10.1364/BOE.416569. eCollection 2021 Mar 1.
A novel non-iterative digital adaptive optics technique is presented in which the wavefront error is calculated using the phase difference between the pupil field and its digital copies translated by a pixel along the horizontal and vertical direction in the pupil plane. This method provides slope data per pixel, thus can generate > 50k local slope data samples for a circular pupil of diameter 256 pixels with high accuracy and dynamic range. It offers more than 12x faster computational speed in comparison to the sub-aperture based digital adaptive optics method. Furthermore, it is independent of any system parameters, the light distribution in the pupil plane, or the intensity of the image. The technique is useful in applications such as interferometric or digital holography based microscopy, metrology, and as digital wavefront sensor in adaptive optics, where focusing of light in the sample is involved that creates a guide star or where the sample itself exhibits guide star-like structures. This technique is implemented in a point scanning swept-source OCT at 1060 nm, and a large wavefront error with a peak to valley of 20 radians and root mean square error of 0.71 waves is detected and corrected in case of a micro-beads phantom sample. Also, human photoreceptor images are recovered from aberrated retinal OCT volumes acquired at eccentricities of 2 and 2.5 degrees from the fovea .
本文提出了一种新型非迭代数字自适应光学技术,该技术利用瞳孔场与其在瞳孔平面中沿水平和垂直方向平移一个像素后的数字副本之间的相位差来计算波前误差。该方法提供每个像素的斜率数据,因此对于直径为256像素的圆形瞳孔,能够高精度且高动态范围地生成超过50k个局部斜率数据样本。与基于子孔径的数字自适应光学方法相比,其计算速度快12倍以上。此外,它与任何系统参数、瞳孔平面中的光分布或图像强度无关。该技术在诸如基于干涉或数字全息的显微镜、计量学等应用中很有用,并且在自适应光学中作为数字波前传感器也很有用,在这些应用中涉及到样品中光的聚焦从而产生导星,或者样品本身呈现出类似导星的结构。该技术在1060nm的点扫描扫频源光学相干断层扫描(OCT)中得以实现,在使用微珠模型样品的情况下,检测并校正了峰谷值为20弧度、均方根误差为0.71个波长的大波前误差。此外,还从距中央凹2度和2.5度偏心位置采集的畸变视网膜OCT体积中恢复了人类光感受器图像。