基于天空区域分割与边缘优化的图像去雾算法  

Image dehazing algorithm based on sky region segmentation and edge optimization

作　　者：赵小强[1,2,3] 周康毅 ZHAO Xiao-qiang;ZHOU Kang-yi(College of Electrical Engineering and Information Engineering,Lanzhou University of Technology,Lanzhou 730050,China;Key Laboratory of Gansu Advanced Control for Industrial Processes,Lanzhou University of Technology,Lanzhou 730050,China;National Experimental Teaching Center of Electrical and Control Engineering,Lanzhou University of Technology,Lanzhou 730050,China)

机构地区：[1]兰州理工大学电气工程与信息工程学院,甘肃兰州730050 [2]兰州理工大学甘肃省工业过程先进控制重点实验室,甘肃兰州730050 [3]兰州理工大学国家级电气与控制工程实验室教学中心,甘肃兰州730050

出　　处：《兰州理工大学学报》2025年第1期92-99,共8页Journal of Lanzhou University of Technology

基　　金：国家自然科学基金(62263021);甘肃省高校产业支撑计划项目(2023CYZC-24);甘肃省科技计划资助项目(21JR7RA206)。

摘　　要：针对暗通道先验算法对天空区域和景深突变区域处理能力不足,复原后的图像存在光晕以及颜色失真等降质问题,提出一种将区域分割与边缘优化相结合的去雾算法.首先,提出一种有效的自适应阈值分割算法,将有雾图像中的天空区域和非天空区域分割开来,在天空区域内对全局大气光值进行求解.其次,对有雾图像天空区域的暗通道和透射率进行分析,求出天空区域透射率权重图,进而求出该区域透射率;对非天空区域提出一种边缘优化算法来改进暗通道,抑制光晕现象的产生.然后,通过引导滤波细化透射率,利用大气散射模型复原出图像.最后,对复原后的图像进行色调映射,使其更加符合人眼视觉特性.实验结果表明,本文算法去雾效果明显,天空区域恢复自然,边缘光晕现象得到有效抑制,复原后的图像在主观视觉和客观指标方面均表现出色.Aiming at the insufficient processing ability of the traditional dark channel prior algorithm for the sky area and the sudden depth of field area,and the degraded problems such as halo and color distortion in the restored image,a dehazing algorithm that combines area segmentation and edge optimization algorithm is proposed in this paper.Firstly,an efficient adaptive threshold segmentation algorithm is proposed to separate the sky and non-sky regions in foggy images,and the global atmospheric light value is solved in the sky area.Secondly,the dark channel and transmittance of the sky area of the foggy image are analyzed to derive the transmittance weight map,and then the transmittance of the sky area is obtained.For the non-sky area.an edge optimization algorithm is proposed to improve the dark channel and suppress the appearance of the halo phenomenon.Then the transmittance is refined by guided filtering,and the image is restored by using the atmospheric scattering model.Finally,tone mapping is performed on the restored image to make it more consistent with the visual characteristics of human eyes.The experimental results show that the proposed algorithm in this paper significantly improves dehazing effectiveness,restores the natural appearance of the sky region,effectively suppresses edge halos,and delivers superior performance in both subjective vision and objective indicators.

关 键 词：暗通道先验 天空区域分割 透射率权重 边缘优化 色调映射 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]