结合浓度划分与图像融合的多分支非均质图像去雾  

Multi-branch non-homogeneous image dehazing based on concentration partitioning and image fusion

作　　者：金鑫乐 刘春晓[1] 叶爽爽 王成骅 周子翔 Jin Xinle;Liu Chunxiao;Ye Shuangshuang;Wang Chenghua;Zhou Zixiang(School of Computer Science and Technology,Zhejiang Gongshang University,Hangzhou 310018,China)

机构地区：[1]浙江工商大学计算机科学与技术学院,杭州310018

出　　处：《中国图象图形学报》2025年第3期798-810,共13页Journal of Image and Graphics

基　　金：国家自然科学基金项目(61976188);浙江省自然科学基金项目(LY24F020004,LZ23F020004);浙江省重点研发计划资助(2023C01039);国家级大学生创新创业训练计划项目(GJ202313014);浙江工商大学“数字+”学科建设项目(SZJ2022B016);浙江省大学生科技创新活动计划暨新苗人才计划项目(2023R408035,2023R408072)。

摘　　要：目的目前的去雾算法已能够较好地处理均质的薄雾图像,但针对雾霾浓度不同的非均质雾霾图像往往具有较低的去雾性能。为此,提出了结合浓度划分与图像融合的多分支非均质图像去雾算法。方法本文将单幅非均质雾霾图像视为由多个具有均质薄雾或者均质浓雾的局部区域组成,通过分别解决单幅非均质雾图中的不同均质雾霾区域来进行整幅非均质图像去雾。首先在不同均质雾霾浓度的去雾数据集上训练了多个图像增强网络,以得到针对不同均质雾霾浓度的图像增强模型,它们对于相应雾霾浓度的图像区域具有较好的增强性能。由于单个图像增强模型只能较好地增强一幅非均质雾霾图像中具有对应雾霾浓度的图像区域,但对其他不同雾霾浓度的图像区域可能存在去雾力度不足或者过度增强的现象,本文又设计了一个图像融合网络,将多个初始图像增强结果中的优势区域进行融合,得到最终的图像去雾结果。结果大量的实验结果显示,在合成雾霾数据集FiveK-Haze上,本文算法与排名第2的SCAN(self-paced semi-curricular attention network)方法相比在峰值信噪比(peak signal-tonoise ratio,PSNR)和结构相似性(structural similarity index,SSIM)有参考指标上分别提高了5.2866 dB和0.1138。在真实雾霾数据集Real-World上,本文算法与排名第2的DEAN(detail-enhanced convolution and content-guided attention network)方法相比,在FADE(fog aware density evaluator)和HazDes无参考指标上分别降低了0.0793和0.0512。在室内合成测试数据集SOTS-indoor(synthetic objective testing set)上,本文算法的PSNR和SSIM指标比排名第2的DeFormer方法分别提高了2.5182 dB和0.0123。在室外合成测试数据集SOTS-outdoor上,本文算法在PSNR指标上比排名第2的SGID-PFF(self-guided image dehazing using progressive feature fusion)方法提高了2.832 dB,在SSIM指标上比排名第2的DeFormer方法提高了0.0238Objective When capturing images using a camera,atmospheric floating particles,such as smoke,dust,and fog,can affect image quality,leading to decreased clarity.These compromised images not only increase the likelihood of human visual misjudgment but also hinder the development of visual tasks such as remote sensing monitoring and autonomous driving.Current dehazing methods are effective for homogeneous thin hazy images but often perform poorly on the nonhomogeneous hazy images.Therefore,a multibranch non-homogeneous image dehazing method combined with concentration partitioning and image fusion is proposed to address these challenges.A single non-homogeneous hazy image is regarded as a combination of multiple local regions with homogeneous thin or dense haze.The entire non-homogeneous image is dehazed by separately addressing different homogeneous hazy regions in a single nonhomogeneous hazy image.Method Concentration partitioning and image fusion based multi-branch image dehazing neural network(CPIFNet),a twostage network framework for image enhancement and image fusion,is then designed.Experiment results revealed that training models based on homogeneous haze image datasets with different haze concentrations can lead to enhancement in image models with varying enhancement intensities.Homogeneous hazed image datasets with different haze concentrations are necessary to obtain varying enhancement models.FiveK-Haze is a synthesized dehazing dataset based on the atmospheric physical scattering model,encompassing nine types of different homogeneous hazed images with varying haze concentrations.The hazy images in the FiveK-Haze dataset are re-partitioned based on haze concentration,dividing the dataset into 1 to 5 different haze concentration levels to exclude the hazy image samples with excessive haze concentrations.Then,the image enhancement network is trained on those new homogeneous dehazing datasets to obtain image enhancement models for different haze concentrations.In the image enhancement networks,the deep imag

关 键 词：图像去雾 非均质雾霾图像 雾霾浓度划分 图像融合 多分支神经网络 

分 类 号：T391.41[一般工业技术]