DOST在声呐图像滤波中的应用  

The Application of the DOST in Sonar Image Filtering

作　　者：田原嫄[1] 王玉聪 郭海涛 

机构地区：[1]东北电力大学机械工程学院,吉林 吉林 [2]海南热带海洋学院海洋科学技术学院,海南 三亚

出　　处：《图像与信号处理》2024年第2期238-246,共9页Journal of Image and Signal Processing

摘　　要：声呐图像的滤波作为图像预处理的首要环节极其重要。但声呐图像具有存在对比度低、灰度值范围小且有用信息不明显的缺陷,传统滤波方法在声呐图像滤波中不能得到良好的效果。故本文将DOST应用于声呐图像的滤波中。通过小波阈值方法确定阈值,再用硬阈值函数去除高频DOST系数,最后进行逆变换,得到滤波后的图像。文中还用多种评价指标对比了多种滤波方法,并用真实声呐图像进行了分析,证明了其的实用性。The filtering of sonar images is extremely important as the primary step in image preprocessing. However, sonar images have the drawbacks of low contrast, small grayscale range, and unclear useful information. Traditional filtering methods cannot achieve good results in sonar image filtering. Therefore, this article applies the Discrete Orthogonal S-Transform (DOST) to the filtering of sonar images. Determine the threshold using wavelet thresholding method, remove high-frequency DOST coefficients using a hard thresholding function, and finally perform inverse transformation to obtain the filtered image. The article also compared multiple filtering methods using various evaluation indicators and analyzed them using real sonar images, proving their practicality.

关 键 词：声呐图像滤波 DOST 自适应阈值 图像评价指标 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]