【摘要】 与普通光学图像相比,声纳图像受到噪声污染更加严重,为了更好的去除侧扫声纳图像噪声,提高图像质量,保持图像原始信息,该文通过将图像变换到多小波域,结合热传导方程的差分格式与图像的分形维数,提出了一种声纳图像软阈值去噪算法,并将该算法与单小波去噪算法做了比较.该算法只需要含噪图像本身,不需要任何其它先验知识,是一种自适应的去噪算法.仿真试验表明,与单小波去噪算法相比,该算法具有更好的去噪效果,同时较好的保持了声纳图像的原始信息.更多还原

【Abstract】 Compared to the optical image,the sonar image is polluted by the noise more seriously. To remove the noise of the sonar image well,improve the quality and keep the original information, the paper propose a sonar image denoising algorithm using multiwavelets threshold. And then,the algorithm is campared to the algorithm using single wavelet.With no need of empirical knowledge of the image,the algorithm can automatically determine the threshold, so it is an adptive algorithm.The result of the simulation experiment show that the effect of the algorithm is perfect than that using single wavelet,and the original information is keep after denoising using the algorithm.更多还原