压缩感知新技术专题讲座(四) 第7讲 压缩感知在图像信号处理中的应用  

Application of Compressed Sensing in Field of Image Signal Processing

作　　者：李莉[1] 杨吉斌[1] 赵斐[1] 

机构地区：[1]解放军理工大学指挥自动化学院信息作战系,江苏南京210007

出　　处：《军事通信技术》2012年第3期93-99,共7页Journal of Military Communications Technology

摘　　要：绝大部分自然图像信号都在某个变换域具有稀疏性或近稀疏性。基于压缩感知理论,可以用远低于采样定理要求的采样频率采集信号,并可在一定条件下高概率恢复信号,这将极大降低图像信号的采样频率以及数据存储和传输的代价。文章首先简述了压缩感知理论,然后分析了图像的稀疏性对图像重建质量的影响,最后着重从图像压缩、图像融合、图像去噪、图像识别以及图像复原几个方面分析了压缩感知理论在图像处理领域中的应用以及目前所面临的问题。Most natural image signals are sparse or nearly sparse in a transform domain. Based on the theory of compressed sensing, the signal with the sampling frequency much lower than the sampling theorem can be sampled, and signal with high probability under certain condi- tions recovered. This can greatly reduce the sampling frequency and the cost of data storage and transmission of the image signal. This paper firstly outlined the compressed sensing theory, then analysed the effect of the sparsity of the image on the reconstruction quality and finally analysed the application of the compressed sensing theory in the field of image processing from image com- pression, image fusion, image denoising, image recognition and image restoration.

关 键 词：压缩感知 稀疏表示 图像压缩 图像融合 图像去噪 图像识别 图像复原 

分 类 号：TN911.73[电子电信—通信与信息系统]