在传统的基于波前探测的解卷积方法中, 由波前探测得到的点扩展函数被认为是精确的, 并用维纳滤波进行复原, 但是点扩展函数不可避免地存在误差, 所以最终的复原目标图像质量不佳。为了解决该难题, 提出了基于目标和点扩展函数联合估计的图像近视解卷积算法。它运用了点扩展函数和目标的先验信息, 对点扩展函数和目标进行了规整和进一步约束, 从而得到更优的恢复图像质量。对该方法的原理和实现过程进行了阐述, 并将其运用于室内点源目标数据中。实验结果证明, 与维纳滤波方法相比, 该方法使图像恢复的效果得到明显改善。

In the classical deconvolution based on wave-front sensing (DWFS), point spread function (PSF) is considered to be accurately achieved and therefore Wiener filter can be used in retrieval. However, for the inevitable noise with PSF, the object image cannot be restored perfectly. A myopic deconvolution method based on joint estimation of the object and PSF to overcome this limitation is proposed. It takes into account the available a priori information on the object to be restored as well as on PSF. The a priori information is used as constraint of the object and PSF, and therefore a better estimation of the restored object can be obtained. The principles and experimental procedure are presented. The experimental results for the indoor point source show that the myopic DWFS can give a better object restoration than Wiener filter method.