基于混合对比度增强的户外图像去雾方法  

Defogging Method of Outdoor Scene Images Based on Hybrid Contrast Enhancement

作　　者：翟艺书[1] 梁媛[2] 

机构地区：[1]天津工程师范学院理学院,天津300222 [2]浙江万里学院基础学院,浙江宁波315101

出　　处：《计算机仿真》2010年第5期227-230,共4页Computer Simulation

基　　金：天津市自然科学基金项目(08JCYBJC12100);天津市高校科技发展基金项目(20081003)

摘　　要：雾天拍摄的户外场景图像对比度和色彩降质严重,影响了视觉监控系统的可靠性和鲁棒性。针对雾天图像灰度动态范围较窄、对比度差的问题,为提高视觉精确性,提出一种基于混合对比度增强的户外降质图像去雾方法。在分析雾天图像特征的基础上,给出了一种全局去雾模型,然后基于遗传算法的全局搜索优势,对图像进行全局优化去雾处理。根据雾天图像降质与深度密切相关的先验知识,研究了一种模糊邻域对比度增强方法,进一步实现局部细节的增强处理,并进行仿真。结果表明,提出的混合去雾方法实现简单,能够有效增强雾天降质图像的全局对比度和细节结构,较好地改善视觉效果。Images captured in fog weather suffer from poor contrast and color, which drastically degrade the reliability and stability of outdoor surveillance system. According to the problem of smaller dynamic gray range and poorer contrast for fog - degraded images, this paper proposed a novel defogging method, which was based on the idea of Hybrid contrast enhancement. Firstly, a global model for defogging was proposed by analyzing the character of fog images, and then, the global contrast enhancement was obtained by optimizing the parameters in model; finally, according to the prior knowledge of atmosperical scattering model and the contrast exponential degrade law, a fuzzy contrast enhancement method was proposed for further local details. Experimental results demonstrate that the proposed hybrid method can enhance both the global contrast of the image and the details of the objects in scene images and performs better visal quality.

关 键 词：雾天降质图像 全局去雾模型 遗传算法 模糊对比度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]