摘要
介绍了具有非负和有限支撑约束的递归逆滤波器盲图像复原算法。在此基础上 ,利用分块奇异值分解和压缩技术 ,提出一种去噪声方法 ,使得可以复原被噪声污染的具有全黑、全白或全灰的背景和有限支撑的目标图像。计算机仿真结果表明 ,新的方法具有更好的图像复原性能。
A restoration algorithm for recursive inverse filtering blind image is introduced for non-negative and finite support images. Based on the algorithm, a de-noise method for using blocking singular value decomposition and compress techniques is proposed. In this method, the blur image is divided into blocks and SVD is performed in every block. Singular values which are less than some given value are set to zeroes. New image is reconstructed using the new singular matrix. The filtered blur image is obtained. Computer simulations show that the proposed approach can restore blur and noise images with whole black, whole white or whole gray background.
出处
《数据采集与处理》
CSCD
2002年第1期1-4,共4页
Journal of Data Acquisition and Processing
基金
国家自然科学基金 (编号 :30 0 0 0 0 4 1 )
山东省自然科学基金 (编号 :Y2 0 0 0 G1 2
1 3)
山东大学青年科学基金资助项目
