一种有效保持边缘特征的散焦模糊图像复原方法pdf.PDF

37 7 Vol. 37 No. 7 2010 7 Computer Science July 2010 1 1, 2 1 ( 厦门大学信息科学与技术学院 厦门361005) 1 ( 厦门大学水声通信与海洋信息技术教育部重点实验室 厦门361005) 2 图像复原过程中图像的主观视觉质量与图像的局部细节信息之间密切相关针对散焦 糊图像, 提出 一种 新的图像复原方法所提方法在传统双边总变分正则化方法基础上, 通过引入 一种具有结构自适应的局部权值函数, 构造了一种新的图像复原目标函数该目标函数综合考虑了图像的全局与局部统计特性, 即在整体保真情况下还充 分考虑了图像的局部结构信息,使得所提复原方法能更有效地保持图像的边缘等细节信息与传统BTV 正则化方法 的比较实验表明, 所提方法在边缘保持方面更有效, 复原后的图像具有更好的主客观视觉质量 图像复原, 点扩散函数, 双边总变分,局部权值函数 TN911.73 A Novel Edgepreserving Algorithm for Defocus Blurred Image Restoration 1 1, 2 1 XIAO Quan DING Xinghao LIAO Yinghao ( School of Information Science and Technology, Xiamen University, Xiam en 361005, China) 1 ( Key Laboratory of U nder water Acoustic Com munication and Marine Information T echnology, Xiamen U niver sity, Xiamen 361005, China) 2 Abstract Relevant research on image restoration indicates that image s subjective visual qualit y is closely related to its local details. A novel restoration algorithm for defocus blurred image w as proposed.T he proposed algorithm based on BTV regularization framew ork by introducing a local adaptive w eight ed function constructs a new cost function for ima ge restoration.This cost funct ion which not only takes into account the global dat afidelity , but also considers the local st atistical properties of image, meaning to fully consider the local st ructural features of image under global datafidelity , hence behaves much bet ter in edge preservation. Experimental results confirm the effectiveness of the proposed method. The image is restored w ith bett er subjective and objective visual quality, com