基于压缩感知的水下图像去噪

Underwater image denoising based on compressed sensing

水下环境复杂,水下摄像得到的图像较为模糊。采集数据时会采集到大量不包含任何有用信息的数据,噪声影响更严重。压缩感知理论提出,能用较低采样率高概率重构信号。为研究压缩感知对水下图像噪声的抑制作用,采用OMP,SP,COSAMP不同贪婪重构算法对水下图像进行不同采样率重构分析。实验结果表明,选取合适采样率既可以以少量数据重构图像,又可以抑制水下噪声,且OMP算法效果最好。

The images got by underwater camera are fuzzy due to complex underwater environment.A number of data without useful information may be collected when data collection is performed,and the noise effect may be more serious.There fore,a theory of compressed sensing is proposed,in which the signal can be reconstructed in high probability by a low sampling rate.In order to research the effect of compressed sensing on underwater image denoising,OMP,SP,COSAMP greedy algorithms are used for the reconstruction analysis of underwater images in different sampling rates.The experimental results show that the image can be reconstructed by a small amount data and the underwater noise can be inhibited if an appropriate sampling rate is chosen.The best effect was got by OMP algorithm.