基于小波阈值自适应修正的模糊图像修复算法  

Fuzzy Image Restoration Algorithm Based on Wavelet Threshold Adaptive Correction

作　　者：张婷曼[1] 

机构地区：[1]延安大学西安创新学院,西安710100

出　　处：《控制工程》2015年第6期1166-1170,共5页Control Engineering of China

基　　金：西安市产业技术创新计划(CX(1245(7))

摘　　要：通过对图像修复,使得图像丢失的信息点得到恢复和再现,起到图像保护的目的。传统的图像修复算法采用采用模板尺寸匹配和纹理特征提取的图像修复算法,由于块匹配过程中阈值不能自适应修正,对破损区域边缘上的像素点的修复效果不好。提出一种基于小波阈值自适应修正的模糊图像修复算法。进行了待修复图像的边缘检测和小波降噪处理,在对待修复图像的破损区域进行Morlet小波特征提取,分析图像边缘轮廓上破损点的向量量化信息,采用小波阈值自适应修正方法实现图像修复优化判决和最佳块匹配。仿真结果表明采用该方法进行图像修复,能较为有效准确地反应损坏的信息特征,避免了传统方法出现严重的结构断裂以及不连续现象,图像模糊性得到改善,信息恢复能力提高,性能定量评价表明,采用该方法峰值信噪比较高,均方误差较低,修复误差较小,性能指标改善突出。The image restoring can recover and reconstruct the lost message, by which the images are protected. The size of template matching and texture feature extraction algorithm for image restoration are adopted by the traditional image, but block matching threshold is not adaptive correction, so the broken repair effect ofpixels on the edge of the area is not good. A fuzzy image restoration algorithm based on wavelet threshold adaptive correction is proposed. The restored image edge detection and wavelet denoising in dealing with image damaged area of Morlet wavelet feature extraction and image edge breakage of vector quantization information are analyzed, using adaptive wavelet threshold correction method for image restoration optimization decision and the best block matching. Simulation results show that the method of image restoration can more accurately reflect the damage characteristic information, prevent serious structural fracture and discontinuity of traditional method, improve the fuzzy image, and enhance the ability of recovering information. And quantitative performance evaluation show that the method of peak signal has a relatively higher noise, lower square error, smaller repair error, and more outstanding performance indicators.

关 键 词：图像修复 小波 自适应 阈值 

分 类 号：TN919.8[电子电信—通信与信息系统]