作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程

• 先进计算 • 上一篇    下一篇

三维可视化的水声数据降噪算法

周天琪,叶学义,高真,宋倩倩   

  1. (杭州电子科技大学通信工程学院模式识别与信息安全实验室,杭州 310018)
  • 收稿日期:2015-01-08 出版日期:2016-01-15 发布日期:2016-01-15
  • 作者简介:周天琪(1988-),男,硕士研究生,主研方向为水声数据可视化;叶学义,副教授、博士;高真、宋倩倩,硕士研究生。
  • 基金资助:
    国家自然科学青年基金资助项目(60802047)。

3D Visualized De-noising Algorithm for Underwater Acoustic Data

ZHOU Tianqi,YE Xueyi,GAO Zhen,SONG Qianqian   

  1. (Laboratory of Pattern Recognition & Information Security,School of Communication Engineering, Hangzhou Dianzi University,Hangzhou 310018,China)
  • Received:2015-01-08 Online:2016-01-15 Published:2016-01-15

摘要: 针对现有二维水声成像去噪算法不能同时有效抑制水声数据中加性噪声和乘性噪声的现状,提出结合模糊中值滤波和小波软阈值滤波的三维可视化水声数据降噪算法。根据复合噪声模型的特点,将三维水声数据分为多个相邻的二维切片数据,通过基于三维滤波窗口的模糊中 值滤波、考虑相邻切片系数相关性与小波系数尺度间相关性的小波软阈值滤波,分层去除水声数据中的加性噪声和乘性噪声。实验结果表明,该算法能降低原始水声数据中的噪声,提高数据连续性和对比度,为观察与探测水下目标提供良好的视觉条件。

关键词: 水声数据, 噪声模型, 模糊中值滤波, 小波变换, 降噪, 可视化

Abstract: Aiming at the problem that traditional underwater acoustic imaging de-noising algorithms can not effectively suppress the additive noise and multiplicative noise at the same time,a 3D visualized underwater acoustic data de-nosing algorithm based on fuzzy median filtering and wavelet soft-threshold filtering is proposed.This algorithm analyzes the characteristics of composite noise model,and divides the 3D data into several 2D data in the concrete operating process.In order to hierarchically remove the additive and multiplicative noise in the underwater acoustic data,it adopts a 3D filter window in the implementation of fuzzy median filtering,and not only considers the inter-scale correlation of wavelet coefficients of 2D plane data,but also considers the correlation of wavelet coefficients of adjacent plane data in the process of soft threshold filtering.Experimental results show that the proposed algorithm reduces the noise in the original data,improves the data continuity and contrast,and it provides a good visual conditions for the observation and detection of underwater acoustic target.

Key words: underwater acoustic data, noise model, fuzzy median filtering, wavelet transform, de-noising, visualization

中图分类号: