Multi-degradation extremely-blurred image restoration method

The invention discloses a multi-degradation extremely-blurred image restoration method so that a problem that the existing criminal investigation software can not restore an original blurred image effectively can be solved. According to the method disclosed by the invention, with consideration of multi-fuzzy factors caused by optical diffraction, quantization, defocusing and relative motion, all sub degradation functions are initialized according to practically measured physical parameters; on the basis of a variational Bayesian algorithm, optimal distribution of all sub degradation functions is solved by means of alternating iteration; according to a relationship between sub degradation functions and system degenerate functions, all sub degradation functions are synthesized into a point spread function of the system; and then image restoration is carried out based on an L-R algorithm. 本发明公开了种多重退化的极模糊图像的复原方法，用于解决现有的刑侦软件不能有效复原模糊图像的问题。该方法主要将光学衍射、量化、散焦以及相对运动引起的多重模糊因素考虑在内，根据实测的物理参数初始化各子退化函数，利用变分贝叶斯算法采用交替迭代的方式求解出各