基于局部模糊和多尺度Retinex的水下图像增强算法.doc基于局部模糊和多尺度Retinex的水下图像 增强算法 张学武,庄宏海,刘小敏,周卓赞,张卓,李敏,范新南 (河海大学物联网工程学院,常州213022) 摘要:针对水下图像模糊,对比度低下等特点,提出一种基于局部模糊和多尺度Retinex的 增强算法。算法中,根据局部区域的平均值和标准差,调整模糊增强的参数,并根据局部区 域所在的人类视觉区域,自适应的调整Retinex算法的参数取值。在此基础上,加权融合两 种算法所得的结果,实现水下图像的清晰化。本文的方法与传统方法相比,在细节增强和色 彩保持方面具有更好的效果。 关键词:局部模糊;多尺度;Retinex;增强算法 中图分类号:TP391 Based on local fuzzy and multiscale Retinex underwater image enhancement algorithm Zhang Xuewu, Zhuang Honghai, Liu Xiaomin, Zhou Zhuoyun, Zhang Zhuo, Li Min, Fan Xinnan (College of Internet of Things Engineering,Hohai University, Chang Zhou 213022) Abstract: According to the characteristics of the underwater images such as fuzzy and low contrast, this paper presents an image enhancement algorithm based on local fuzzy and multi-scale Retinex. In this algorithm, we are adjusting the parameter of the fuzzy enhancement base on the average and standard deviation of the local area. According to the local area where human visual area, adaptive adjusting the parameter of the Retinex algorithm. On the basis, we are weighting the results of the two algorithms to realize the underwater images clearly. Compared with the traditional method, this paper has a better effect in detail enhancement and color. Key words: local fuzzy; multi-scale;Retinex; enhancement algorithm 0引言 随着科技的进步对水下图像的要求越来越高,但由于水体本身的一些特性及光照不均匀 等摄影时存在的客观问题,使得水下图像对比度差等缺陷。这些缺陷影响着水下图像的分割、 特征提取以及目标识别。所以在图像处理前需要对图像进行增强处理,这