基于差分算子边缘约束先验的图像盲复原算法

Image restoration algorithm based on edge constraint prior of differential operator

为了有效克服图像在复原过程的边缘退化、振铃效应等影响,使空间域的边缘先验信息能够灵活的添加到图像复原算法中,依据差分算子描述检测边缘的特性,将数字摄影测量学中的拉普拉斯算子、Robert梯度算子、方向差分算子作为一种新的边缘约束先验引入到图像复原过程中,同时使用Toeplitz矩阵实现图像在空间域解卷积的过程,提出了一种以差分算子为边缘约束先验的空域图像复原算法。模拟数据的实验结果体现了更多的细节信息,相关评价指标表明了该算法的有效性。

In order to overcome the edge degradation of the image in the recovery processing and add the edge prior information of the space domain into the image restoration algorithms,Laplacian operator,Robert operator and direct differential operator of digital photogrammetry is introduced as an edge constraint prior into the process of image restoration,analysis and examine the process to realize the deconvolution of image in the space domain using Toeplitz matrix and promoted a space domain image restoration algorithm using differen-tial operator as edge constraint prior.The experimental results of the simulated data demonstrate more detailed information and proved the validity of the algorithm.