基于压缩感知理论的图像融合方法.pdfVIP

194 2012 ,48(12) Computer Engineering and Applications 计算机工程与应用 基于压缩感知理论的图像融合方法 张伟I 气曾凡仔1 ，曾庆光1 1 1 1 ZHANGWei ZENG Fanzi , ZENG Qingguang 气 1.湖南大学计算机与通信学院，长沙 410082 2.华北计算技术研究所后勤信息化事业部，北京 100083 I.School of Computer and Communicatíon, Hunan Uníversíty, Changsha 410082 , Chína 2.Department Of Logístics InfOrmation Business, North China Institute of COmputing Technology,Beijing 1 ∞083， Chína ZHANG Wei, ZENG Fanzi, ZENG Qingguang. Method of image fusion based on compressed sensing. Com- puìer Engineering and Applications ,2012 ,48( 12) : 194-197. Abstract: Based On COmpressed Sensíng ( CS) , a new image fusíon method is presented. The new method can take reduced samples, and has the advantages of símple structure and e础y implementatiOn. The method decOmpOses twO or more original ímages using wavelet transfOrm, and gets the sparse matrix by the wavelet cOefficient sparse repre- sentatíon, and fuses the sparse matrices wíth the cOefficients absOlute value maximum scheme. RandOmly Observed, it can receive the cOmpressed sample. At the fusing end, the fusíOn ímage ís recovered 企om the reduced samples by sOlvíng the OptímízatiOn. The prOpOsed methOd can cons往uct the fusiOn image with less measurements because the wavelet ∞efficients are sparse presentation. SimulatiOn results shOw the propO sed method exhibits its superiority overthe 往aditiOnal method of the maximum absOlute values fusíon wíth the same sampling rates , and under the lOw- er sampling, it can alsO achieve be阳r 缸síon perfOrmance. Key words: cOmpressed sensing; randomly observed; image fusiOn; absOlute value maximum scheme 摘 要:基于压缩感知(COmpressed Sensing , CS)理论，提出了一种采样点少且结构简单易实现的图像融合方 法.对需要处理的两幅或多幅困像进行小波变换，分别对得到的小波系数进行稀疏处理得到稀疏矩阵，通过 系数绝对值较大法进行融合，对融合后的系数矩阵通过随机观测获取压缩采样，而图像恢复则是对得到的压 缩采样通过求解最优化的问题得到.由于对iJ 、波系数进行了稀疏处理，故该方法可以用少量的采样点来恢复 图像.实验结果表明，在相同采样点下，该方法得到的图像质量(PSNR) 明显优于传统的系教绝对值较大法融 合;在少量采样点下，采用该方法也可以使融合的图