摘要
压缩感知是一种能够在某个特定域中压缩和恢复稀疏信号的技术。针对在使用传统观测矩阵进行数据压缩时,其数据恢复效果并不理想,且观测矩阵的随机性会导致数据传输量较大、硬件实现因难等问题,提出一种新的观测矩阵生成方法。将信道编码中的LDPC校验矩阵与对角块矩阵结合,生成一种尺度较小且易于硬件实现的观测矩阵,这种矩阵不仅高度稀疏,而且元素二值化。通过多组图像重构仿真实验对比发现,LDPC对角块矩阵重构结果优于其他传统观测矩阵的重构结果。
Compressive sensing is a technique that is suitable for compressing and recovering signals having sparse representations in certain bases.In view of two main problems in currently existing measurement matrices for compressive sensing of natural images,such as difficulty of hardware implementation and low sensing efficiency,this paper proposed a simple measurement matrix.By combining the diagonal block matrix with the LDPC check matrix in the channel coding,a new measurement matrix that facilitates the hardware implementation is generated.The diagonalizable LDPC measurement matrix is highly sparse and binary,and reduces the data storage space and computing time.Through the comparison of multiple sets of images,the reconstruction results of this method are much better than the others.
出处
《计算机科学》
CSCD
北大核心
2017年第7期279-282,共4页
Computer Science
基金
国家自然科学基金(61273251)
国家自然科学基金青年基金(61401220)
民用航天"十二五"技术预先研究项目
江苏省自然科学基金青年基金(BK20140884)资助
关键词
压缩感知
观测矩阵
对角化
LDPC
Compressive sensing
Measurement matrix
Diagonalization
LDPC
