摘要
遗传算法在模糊图像恢复的应用上,如果算法设计的不合理,需要更多的迭代次数,影响算法本身的运行效率,也有可能会陷入局部收敛,影响图像恢复的效果。针对现有的遗传算法,结合图像本身的特点,提出了一种新的图像模糊恢复的遗传算法结构。该算法以二维的染色体编码方式,通过样本分布模板和多重随机参数,以提高迭代收敛的速度,同时避免局部收敛。实验结果表明,该算法在运动模糊图像的恢复中,要优于传统的逆滤波法,算法的抗噪声能力较强,对于运动参数估计的依赖性也较弱。
When applying genetic algorithms in fuzzy image restoration,if the algorithm is designed to be unreasonable,need more iterations,there are likely to fall into the local convergence affect the image restoration results.Against the existing genetic algorithms,combined with the characteristics of the image itself, presents a genetic algorithm to restore images blurred structure.The algorithm is 2-D chromosome encoding,through the distribution of the sample templates and multiple random parameters in order to improve the speed of iterative convergence,while avoiding local convergence.The experimental results show that the algorithm in the motion-blurred image restoration,superior to the traditional inverse filtering method.
出处
《计算机技术与发展》
2010年第6期5-8,共4页
Computer Technology and Development
关键词
遗传算法
运动模糊
随机种群
样本分布模板
genetic algorithm
motion blur
random population
sample distribution template