利用NSCT循环抽样进行水声图像去噪

The underwater acoustic image denoising based on NSCT cycle sampling

在海上军事及海底探测领域,基于声呐图像的识别应用越来越多。但是声呐图像混入了多种海上噪声,在图像对比时首先需要去除噪声。NSCT是一种基于拉普拉斯逆变换的滤波方法,对于海上声呐图像首先按照周期进行循环抽样,然后利用NSCT变换得到图像边缘系数,再通过滤波平移对噪声进行去除。本文在研究声呐图像及海上噪声特性的基础上,利用NSCT滤波算法对其进行去噪处理,得到空域平均后的去噪图像,并给出仿真结果。

In the field of maritime military and seabed exploration,the applications of sonar image recognition are more and more,but due to the variety of marine sonar image noise,it must to remove noise first based on applications of sonar image. NSCT is a filtering method based on Laplasse transform inverse filter,first using the cycle of circular sampling and then the edge image is obtained by using NSCT transform coefficients,the noise is removed by filtering translation. Based on the study of the sonar image and the noise characteristics of the sea,This paper use the NSCT filtering algorithm to deal with the noise,and the noise image is obtained after spatial averaging,at last the simulation results are given.