基于方向信息测度的运动模糊图像恢复  

Motion Blurred Image Restoration Based on Orientation Information Measure

作　　者：郝毫刚[1] 李艳玲[1] 黄春艳[1] 

机构地区：[1]华北水利水电学院数学系,郑州450011

出　　处：《计算机工程》2010年第2期201-202,共2页Computer Engineering

基　　金：河南省教育厅自然科学研究基金资助项目(2008B120004)

摘　　要：针对水平直线匀速运动引起的模糊图像产生过程,在图像的空间域内,利用含约束条件的最小二乘法、边缘信息逼近的能量差和Lagrange乘子法构成代价函数,以求解复原图像,并逐步推演出其求解公式。该算法实现简单,不需要进行多次迭代运算和频域变换。实验结果证明,处理后的图像质量良好,能满足人们的视觉要求,避免了频域恢复中出现的高频振铃现象。Aiming at the generation process of blurred image aroused by the straight and level constant velocity motion,this paper constructs cost function by using constrained least square method,energy differences of edge information approximation and the Lagrange multipliers in image spacial domain to solve the restored image and deduce its solution formula gradually. The algorithm is simple,does not need too many times iterative computations or frequency transformation. Experimental results indicate that the processed image has good quality,can satisfy peoples visual demands and avoid the high frequency deranging phenomenon in the frequency domain restoration.

关 键 词：运动模糊 LAGRANGE乘子 方向信息测度 图像恢复 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]