中文摘要：

为了在不精确知道图像的点扩展函数的情况下，去除图像的离焦模糊，提出了基于高维空间几何理论的图像复原算法。依据同源连续性原理，通过分析高维空间中向量的方向和点的位置关系来研究模糊图像与原图像的空间关系，并且结合最近邻算法，将该算法应用于去除最近邻算法所得图像的本层模糊。实验表明该算法能够有效地去除离焦模糊。本算法不同于传统的图像复原算法，可广泛应用于模糊图像的复原中。

英文摘要：

To recover blurred image but the PSF of the image was not known, using the method which according to homeomorphisms and the principle of homology continuity（ PHC ） in high-dimensional space geometry. The relation of vector and point in space could show the relation of the fuzzy image and the original image, contacting nearest-neighbor algorithm to use the method in computational optical section microscopy （COSM）. The results show that the method is efficient and correct. The method is not the same as traditional method, it can be used in the restoration of blurred image broadly.