基于小波变换和改进SVD的红外图像去噪

Infrared Image Denoising Based on Wavelet Transform and Improved SVD

针对小波变换红外图像去噪需要已知噪声先验知识的缺点,提出了一种基于分块奇异值分解的正交小波变换红外图像去噪新算法。首先对红外图像进行离散正交小波变换,并对高频图像采用改进的分块奇异值分解估计小波系数,其中对奇异向量采用傅里叶变换进行了修正;最后将低频图像与估计的高频图像通过小波反变换得到去噪图像。仿真结果表明,该图像去噪算法能在无噪声先验知识条件下有效去除图像噪声,信噪比有了明显提高,并获得了良好的主观视觉效果。

Due to the problem of the knowledge of noise using wavelet transform denoising, a new method based on orthogonal wavelet transform using block-based singular value decomposition for infrared image denoising is proposed. Firstly, infrared image is decomposited using orthogonal wavelet transform. For the high frequency components of image decomposition ,the wavelet coefficients are estimated using improved block-based singular value decomposition. And the singular vectors are modified using fourier transform. Then the high frequency coefficients of signal was obtained. The denoised image is obtained through inverse wavalet transform by the low frequency image and the high frequency images. The experimental results show that the infrared image can be denoised effectively in this means without the knowledge of noise. The SNRs are improved substantially, and the visual quality is achieved well.