基于独立分量分析的数字图像盲复原方法研究

Independent Component Analysis is a novel method for blind source separation based on high statistics，in general methods of ICA, its application scope is limited, because it doesn’t have a constraint of non-negativity of the source random victors and only if the source vectors are independent of each other, ICA can achieve good results. In this paper, the method of ICA is application to blind image restoration with the constraint of non-negativity. Besides this, by a matrix of decor relation, it eliminates the bad effect of the source signal’s correlation to the algorithm. The experiment results demonstrate that the algorithm can realize blind image restoration better.