基于边缘提取与图像融合的压缩感知补偿算法  

Compressive Sensing Compensation Algorithm Based on Image Edge Detection and Fusion Technology

作　　者：赵明昕[1] 金志刚[1] 卫津津[1] 

机构地区：[1]天津大学电子信息工程学院,天津300072

出　　处：《南开大学学报（自然科学版）》2016年第2期15-21,共7页Acta Scientiarum Naturalium Universitatis Nankaiensis

基　　金：自然科学基金项目(61201179)

摘　　要：针对测量数据极稀疏情况下,某些压缩感知方法在多种类噪声干扰环境下进行图像传输存在重构效果模糊、图像细节欠佳的问题,提出1种基于图像边缘提取与融合技术的压缩感知补偿算法.首先,在发送端预提取图像边缘特征作为压缩感知测量值的重要补充,随后,在接收端进行边缘解码与压缩感知重构,最后,使用空间域方法实现2者融合.实验结果表明,该方法可以以较小的传输代价获得更加清晰、全面的图像细节特征,可增强原算法抵抗噪声、尤其是稀疏性噪声的能力,具有一定的应用价值.In view of some compressive sensing （CS） method with extremely sparse measurements existing fuzzy reconstruction effect and poor image details in image transmission under various kinds of noise interference, a compressed sensing compensation algorithm based on image edge detection and fusion was proposed. First, the characteristic of image edge was pre-extracted at the sending end as the important compensation of CS measurements; Second, edge decoding and CS reconstruction were implemented at the receiver; Finally, those two results were fused by space-domain method. Experimental results demonstrate that the CS compensation algorithm is able to achieve images with more clear and comprehensive details and acceptable cost, and enhance the ability of original algorithm to resist noise, especially the sparse noise. This algorithm is easy to implement.

关 键 词：图像传输 压缩感知 边缘提取 图像融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]