基于压缩感知的声纳成像信息重建技术研究

Research for Sonar Image Information Reconstruction Based on Compressed Sensing

人类对海洋的探索发现中,由于水下环境复杂使得水下声数据面临着高速采集、压缩和图像重构的问题。新兴的压缩感知理论提出,能用低采样率高概率地重构原始信号。论文研究了DCT和DWT两种变换基对声纳图像进行稀疏表示,并改进了一种在低采样率下重建稀疏图像的测量矩阵,通过仿真实验论证了该矩阵相对其他矩阵在低采样率下有效提升了重构图像质量,为利用少量数据重构水下图像提供了便利。

Humans are faced the problems with high-speed sampling,compressing and image reconstruction due to complex underwater environment in the process of the discovery for oceans. Recently the new theory of compressed sensing is proposed,in which the signal can be reconstructed in high-probability by a low sampling rate. This paper has researched the sparse representation for the sonar image by the conversion called DCT and DWT. It improves a measurement matrix when the sparse image can be reconstructed. Simulation results demonstrate that the matrix can obviously enhance the sparse image reconstruction quality against others,which provides convenience for the reconstruction by using less data.