【摘要】 为解决声纳图像易受噪声污染、对比度降低的问题,提出一种形态小波域图像去噪算法。首先构造了可以用于灰度图像处理的形态中值小波,并在实现完备重构的基础上对中值小波进行了提升;然后利用提升后的形态小波,以移位抽取的方式对声纳图像进行完整的形态多分辨率分析;经阈值处理后,最终通过各重构图像的加权平均得到理想的去噪效果。仿真实验表明,形态小波域去噪算法在去除噪声的同时能完好保留图像边缘等重要细节,与传统的小波阈值去噪方法比较,对比度有所提高,图像更加清晰。更多还原

【Abstract】 Sonar images are susceptible to noise pollution,thus resulting in lower contrast.To resolve this problem,this paper proposes an image denoising algorithm in morphological wavelet domain.Firstly,a morphological median wavelet is constructed for gray image processing.On the basis of achieving reconstruction without the redundance,the median wavelet is lifted.Then,the sonar image is analyzed in different multiresolutions by shifting decomposition and the lifted morphological wavelet.Finally, after threshold processing,a desired denoising effect is obtained by the weighted average of different reconstructed images.Computer simulations show that the denoising algorithm in the morphological domain can retain important details,such as image edge and so on.Compared with the traditional wavelet denoising method using the threshold,the contrast is enhanced and the image is clearer.更多还原