Mait J N, Rhodes W T
Appl Opt. 1989 Apr 15;28(8):1474-88. doi: 10.1364/AO.28.001474.
A two-step iterative algorithm is described for designing pupil functions for synthesis of bipolar point spread functions in a two-channel incoherent spatial filtering system. The first step uses the method of projections onto convex sets to determine an appropriate bias for the system pupil functions. This bias determines the complementary point spread functions and therefore the magnitude of the corresponding coherent spread functions. The phase of the coherent spread functions is obtained in the second step via iterative application of a magnitude constraint and a finite support constraint. With both magnitude and phase of the coherent spread functions established, the corresponding pupil functions are realized via Fourier transformation. Results are presented for a 2-D bandpass filter and compared with analytically obtained results. The comparison suggests that the iterative algorithm yields bias functions having less energy than bias functions determined by previously proposed analytic methods.
描述了一种两步迭代算法，用于设计双信道非相干空间滤波系统中用于合成双极点扩散函数的光瞳函数。第一步使用凸集投影法来确定系统光瞳函数的适当偏差。该偏差决定互补点扩散函数，从而决定相应相干扩散函数的大小。第二步通过迭代应用幅度约束和有限支撑约束来获得相干扩散函数的相位。在确定了相干扩散函数的幅度和相位后，通过傅里叶变换实现相应的光瞳函数。给出了二维带通滤波器的结果，并与解析得到的结果进行了比较。比较表明，该迭代算法产生的偏差函数比先前提出的解析方法确定的偏差函数具有更少的能量。
