APEX方法是一种非迭代式的点扩展函数(PSF)估计方法, 适用于实时性要求较高的场合, 但其估计出的PSF在有限支持域内不能保证退化过程能量守恒。在APEX方法的基础上, 针对光电图像的退化特征, 在估计PSF时加入非负限制和归一化处理, 并对APEX方法中的近似常数A和频域中心区域Ω的选取加以改进, 增强了其实用性。最后用慢演化连续边界(SECB)方法重构复原图像, 同时就SECB方法中的两个重构因子K和s的选取给出参考结论。通过仿真, 验证了改进算法不仅增强了稳定性, 而且复原效果也有较大改进。

APEX method is a kind of non-iterative restoration approach, can be applicable to some real-time situation. There is a problem that its result point spread function (PSF) can not be sure energy is conversation in finite support area. The PSF was restrained non-minus and linear unitary based on the APEX method, pointing to the degradation character of opto-electronic image, and the approximate constant A and frequency center domain Ω was improved. At last, reconstruct the restoration image utilizing slow evolution of continuation boundary (SECB) method, give the reference conclusion for how to select the reconstruction factor K and s. The simulation indicates the improved algorithm enhanced its stability, the effect of restoration also has been improved.