J4 ›› 2018, Vol. 35 ›› Issue (1): 7-12.

• 图像与信息处理 • 上一篇    下一篇

基于可控核双边滤波Retinex水下图像增强算法

石磊1,奚茂龙2,孙 俊2   

  1. 1. 无锡职业技术学院控制技术学院
    2. 江南大学
  • 收稿日期:2016-11-01 修回日期:2017-03-08 出版日期:2018-01-28 发布日期:2018-04-04
  • 通讯作者: 石磊

Underwater image enhancement algorithm based on controllable nuclear bilateral filtering Retinex

  • Received:2016-11-01 Revised:2017-03-08 Published:2018-01-28 Online:2018-04-04

摘要: 在水下环境中,水中各类干扰因素导致采集到的图像质量并不理想。为提高水下图像的成像质量,提出一种可控核函数双边滤波Retinex水下灰度图像增强方法。对于采集到的水下图像,首先引入改进的Sigmoid函数提升对数变换后的照度图像的亮度,同时提出一种可控核函数的双边滤波函数来估计图像的光照强度,克服传统高斯函数存在的边缘模糊和光晕伪影问题,之后,对照度估计图像采用增益函数进行校正处理。实验结果表明提出的算法有效提升了水下图像的对比度和清晰度,符合人眼视觉特性。

关键词: 图像处理 水下图像增强 Retinex Sigmoid函数 可控核函数

Abstract: The quality of underwater imaging is degraded due to many kinds of interference factors in the water in the underwater environment. For the collected underwater image, the underwater image enhancement method is presented in this paper,which base on the Retinex of ganglion receptive fields algorithm. To be the image under the enhanced water, first introduced after logarithmic transformation of improved Sigmoid function of illumination image brightness, and puts forward the bilateral filtering function of the controllable kernel function to estimate the image of the light light intensity to overcome the traditional Gaussian function has the edge blurring problem, and the illumination estimation function by different gain function correction processing is performed. The experimental results show that the proposed algorithm can effectively improve the contrast and definition of underwater image, which is consistent with the human visual characteristics.

Key words: image processing underwater image enhancement Retinex Sigmoid function Controlled kernel function