您的位置:首页  > 论文页面

基于鸡群优化的梯度域引导滤波Retinex图像自适应增强方法研究

发表时间:2018-11-30  浏览量:45  下载量:7
全部作者: 胡峰,黄勇,杨勇,代劲
作者单位: 重庆邮电大学计算机科学与技术学院计算智能重庆市重点实验室
摘 要: 针对现有图像增强方法在增强后出现过度增强和颜色失真,以及现有色调平衡方法对图像增强方法中的图像亮度敏感等问题,提出一种新的基于鸡群优化(chicken swarm optimization,CSO)算法的梯度域引导滤波(gradient domain guided image filter,GDGIF)图像自适应增强方法。以无参考图像质量评价模型(blind image quality measure of enhanced images,BIQME)为目标函数,将主观图像增强的问题转化为最优化的问题。首先对图像的灰阶亮度进行梯度域引导滤波Retinex图像增强和Gamma校正,其次对增强后的图像进行色调平衡,最后使用CSO算法对目标函数进行优化求解,获得图像增强的最优梯度域引导滤波参数和色调平衡参数,从而实现图像的自适应增强。在不同实验场景下的对比结果表明,本文方法不仅在主观感受上实现了较好的图像增强,同时也在客观评价指标信息熵、无参考图像对比度扭曲程度(no-reference image quality metric for contrast distortion,NIQMC)和BIQME方面有了较大的提高。
关 键 词: 计算机应用;图像自适应增强;梯度域引导滤波;Gamma校正;简单色调平衡;鸡群优化
Title: Research on gradient domain guided image filter Retinex image self-adaption enhancement method based on chicken swarm optimization
Author: HU Feng, HUANG Yong, YANG Yong, DAI Jin
Organization: Chongqing Key Laboratory of Computational Intelligence, College of Computer Science and Technology, Chongqing University of Posts and Telecommunications
Abstract: For the problems of existing image enhancement methods in enhanced, color distortion and the sensitivity of the existing color balance methods to the brightness of the image in the image enhancement methods, a novel gradient domain guided image filter (GDGIF) Retinex image self-adaption enhancement method based on chicken swarm optimization (CSO) algorithm is proposed. Blind image quality measure of enhanced images (BIQME) is adopted as objective function, in which the subjective image enhancement problem is transformed to an optimization problem. Firstly, the gray scale brightness of the image is graded by gradient domain guided image filter Retinex image enhancement and Gamma correction. Secondly, the enhanced image is subjected to hue balance. Finally, the CSO algorithm is used to optimize the objective function in order to obtain the optimal gradient domain guided filter parameters and the hue balance parameters of the image enhancement, so as to achieve the image self-adaption enhancement. The compared results in different experimental scenarios indicate that this method not only achieves better image enhancement on subjective feelings, but also improves the information entropy, no-reference image quality metric for contrast distortion (NIQMC) and BIQME of objective evaluation index.
Key words: computer applications; image self-adaption enhancement; gradient domain guided image filter; Gamma correction; simplest color balance; chicken swarm optimization
发表期数: 2018年11月第22期
引用格式: 胡峰,黄勇,杨勇,等. 基于鸡群优化的梯度域引导滤波Retinex图像自适应增强方法研究[J]. 中国科技论文在线精品论文,2018,11(22):2254-2269.
 
2 评论数 0
暂无评论
友情链接