结合压缩感知和曲波的天文图像去噪

在天文图像去噪中,为了提高迭代曲波阈值算法的去噪重建性能,提出了基于循环平移和曲波维纳滤波的压缩感知迭代重构算法。首先,使用基于曲波阈值的循环平移方法对重构图像进行调整以抑制重构图像中的伪吉布斯效应;接着,用提出的曲波维纳滤波算子替代小波阈值在迭代过程中对图像曲波系数进行筛选以进一步提高重构图像的质量。通过对添加高斯白噪声的Lena图像和月球图像进行重构实验,分析本文算法和当前主流算法的性能。实验结果表明,与传统的压缩感知迭代曲波阈值算法相比,本文算法能够获得较优的去噪性能,有效地保护天文图像的细节信息,峰值信噪比大约提高了2.6~3.2dB。

Astronomical image denoising with compressed sensing and curvelet

In astronomical image denoising, to improve denoising construction performance of iterative curvelet threshold (ICT)algorithm, a compressed sensing iterative reconstruction algorithm by combining cycle spinning and curvelet wiener filtering was proposed.Firstly, cycle spinning method based on curvelet threshold was used to adjust reconstructed images for inhibiting Pseudo-gibbs effect of reconstructed images;then, proposed curvelet wiener filtering operators were used to replace wavelet threshold for sieving image curvelet coefficient to further improve the quality of reconstructed image.The reconstruction experiment on Lena image and moon image with Gaussian white noise was conducted, and the result shows that compared with traditional compressed sensing ICT algorithm, the peak signal noise ratio of proposed algorithm increases by 2.6~3.2 dB approximately.So the proposed method can acquire better denoising performance, and can protect detail information of astronomical images effectively.