摘要
运动图像的盲复原,一直是图像处理领域的难点,本文根据运动模糊图像特点,提出了一种鲁棒、高效保留图像细节和纹理特征的图像盲复原算法.在图像的复原过程中运用了基于最大似然函数的Richardson-Lucy滤波方法只需知道模糊类型,不需要其它任何先验知识,就能做到有效复原,适用范围广,能对线性匀速运动类型的模糊、振动模糊、旋转模糊都有效,鲁棒性好,抗噪能力强.
Blind restoration of motion image always is a challenging problem in image processing, the paper, in view of the characteristic of motion image, puts forward a algorithm of blind restoration of motion image with good robustness and preserving image details and texture characteristic efficiently. The algorithm applys Richardson - Lucy filter based on the maximum - likelihood function during image restoration. It can restore efficiently a motion blurred image without any prior knowledge as long as the type of blur is known, its applicability for processing motion blurred images has a more wide range including the uniform linear motion images, vibration blurred images and rotation blurred images and it has a good robustness, strong anti -noise ability.
出处
《南华大学学报（自然科学版）》
2010年第2期61-67,共7页
Journal of University of South China：Science and Technology
关键词
图像盲复原
倒谱法
运动模糊
鲁棒性
blind restoration of image
cepstrum method
motion blur
robustness
