摘要
现有迭代盲复原算法存在计算复杂、参数选择困难、收敛性难以保证的缺点。基于带噪声图像退化模型在欧几里得向量范数约束下的最优化问题,提出快速混合迭代图像盲复原算法。讨论了迭代公式中的参数选取问题,给出了退化函数迭代初始的标准设计方法。这种算法具有通用性强、计算量小、收敛性好、设计简便的特点,复原质量主要受到点扩散函数初始估计尺寸的影响,一般不需要交互式地调整滤波器参数。仿真结果表明,复原出的图像细节清楚,视觉效果显著提高。
The existing iterative blind restoration algorithms have the disadvantages of complicated calculation,difficulty in parameter selection and poor convergence.Based on the optimization problem of the image degradation model with noise under the Euclidean vector norm constraint,a fast hybrid iterative image blind restoration algorithm is proposed.The problem of parameter selection in the iterative formula is discussed,and the standard design method of iterative initial value of degenerate function is given.This algorithm has the characteristics of strong universality,small computation,good convergence and simple design.The restoration quality is mainly affected by the initial estimated size of the point spread function,and it is generally not necessary to adjust the filter parameters interactively.The simulation results show that the details of the restored image are clear and the visual effect is significantly improved.
出处
《智能城市》
2017年第8期20-22,共3页
Intelligent City
基金
广西高校中青年教师基础能力提升项目(2017KY0351)
广西科技大学科学基金项目(20161312)
关键词
盲复原
快速
迭代
退化模型
欧几里得向量范数
blindrestoration
fast
iterative
the degradation model
the Euclidean vector norm
