基于传统暗原色先验原理的图像去雾算法存在的“halo”效应,且图像中明亮区域存在颜色失真现象,针对此问题,本文提出了多尺度窗口的自适应透射率修复交通图像去雾方法。首先,利用新的8方向边缘检测算子求取图像中景深突变区域,根据暗通...基于传统暗原色先验原理的图像去雾算法存在的“halo”效应,且图像中明亮区域存在颜色失真现象,针对此问题,本文提出了多尺度窗口的自适应透射率修复交通图像去雾方法。首先,利用新的8方向边缘检测算子求取图像中景深突变区域,根据暗通道先验理论和前一步求得的景深突变区域,在景深变化较大区域使用5 X 5的窗口,景深变化较小区域则使用15 x 15的窗口得到暗原色估计图。同时,针对暗通道先验原理对近景部分存在白色区域时透射率估计不准确的问题,引人了自适应透射率修复方法,通过引导滤波器得到边缘增强后的暗原色图像,并利用其与原暗原色图像的纹理差对近景区域的透射率进行修正,完成图像去雾。实验结果表明:双边滤波和梯度双边滤波两种算法均存在halo现象,并且在包含白色物体的明亮区域色彩失真严重,客观评价指标失去意义;相比于引导滤波,本文去雾算法的各项指标均有所提高,其中平均梯度平均提高了8.305%,PSNR平均提高了12.455%,边缘强度因子平均提高了7.77%。本文算法有效解决了复原图像中“halo”效应现象和明亮区域颜色失真现象,去雾效果最优。展开更多
遥感图像在成像过程中,容易受到云层和雾霾天气的影响,形成带雾图像;同时在下传时,会受到多种因素影响(如发送接收误码、电离层和对流层的随机变化对信号形成扰动等),使图像信息丢失或掺杂噪声。本文针对信息丢失的带雾单色遥感图像,提...遥感图像在成像过程中,容易受到云层和雾霾天气的影响,形成带雾图像;同时在下传时,会受到多种因素影响(如发送接收误码、电离层和对流层的随机变化对信号形成扰动等),使图像信息丢失或掺杂噪声。本文针对信息丢失的带雾单色遥感图像,提出了基于矩阵复原和暗通道理论的单色遥感图像去雾算法,通过基于交替方向乘子法(Alternating Direction Method of Multipliers,ADMM)的矩阵复原算法与传统暗通道理论相结合,有效实现了信息丢失下的遥感雾图复原。通过主观评价和客观评价相结合的方式,将本文算法与经典算法对比。结果表明,本文算法得到的结果在直观视觉上效果更好,且相对于信息丢失30%的雾图,6个场景的平均信息熵提升1.6652,平均峰值信噪比提升11.7029,平均结构相似性提升0.8146,客观评价指标结果优异。进一步在不同比例信息丢失情况下进行实验,结果表明,即使在信息大量丢失的情况下,依然能够得到清晰的复原去雾图像。展开更多
To solve the problem of color distortion after dehazing in the sky region by using the classical dark channel prior method to process the hazy images with large regions of sky,an improved dark channel image dehazing m...To solve the problem of color distortion after dehazing in the sky region by using the classical dark channel prior method to process the hazy images with large regions of sky,an improved dark channel image dehazing method based on Gaussian mixture model is proposed.Firstly,we use the Gaussian mixture model to model the hazy image,and then use the expectation maximization(EM)algorithm to optimize the parameters,so that the hazy image can be divided into the sky region and the non-sky region.Secondly,the sky region is divided into a light haze region,a medium haze region and a heavy haze region according to the different dark channel values to estimate the transmission respectively.Thirdly,the restored image is obtained by combining the atmospheric scattering model.Finally,adaptive local tone mapping for high dynamic range images is used to adjust the brightness of the restored image.The experimental results show that the proposed method can effectively eliminate the color distortion in the sky region,and the restored image is clearer and has better visual effect.展开更多
针对现有的深度学习去雾算法参数多,训练时间长,无法应用到实时计算机视觉系统等问题,本文提出了一种基于明暗通道的循环GAN网络(bright and dark channel CycleGAN network,BDCCN).BDCCN以CycleGAN为基础,采用固定参数和训练参数相结...针对现有的深度学习去雾算法参数多,训练时间长,无法应用到实时计算机视觉系统等问题,本文提出了一种基于明暗通道的循环GAN网络(bright and dark channel CycleGAN network,BDCCN).BDCCN以CycleGAN为基础,采用固定参数和训练参数相结合方式,基于明暗通道先验理论,改进循环感知损失,实现图像去雾.实验结果表明,本文算法计算量小,收敛快,在合成数据集和真实数据集上均表现优异.展开更多
文摘基于传统暗原色先验原理的图像去雾算法存在的“halo”效应,且图像中明亮区域存在颜色失真现象,针对此问题,本文提出了多尺度窗口的自适应透射率修复交通图像去雾方法。首先,利用新的8方向边缘检测算子求取图像中景深突变区域,根据暗通道先验理论和前一步求得的景深突变区域,在景深变化较大区域使用5 X 5的窗口,景深变化较小区域则使用15 x 15的窗口得到暗原色估计图。同时,针对暗通道先验原理对近景部分存在白色区域时透射率估计不准确的问题,引人了自适应透射率修复方法,通过引导滤波器得到边缘增强后的暗原色图像,并利用其与原暗原色图像的纹理差对近景区域的透射率进行修正,完成图像去雾。实验结果表明:双边滤波和梯度双边滤波两种算法均存在halo现象,并且在包含白色物体的明亮区域色彩失真严重,客观评价指标失去意义;相比于引导滤波,本文去雾算法的各项指标均有所提高,其中平均梯度平均提高了8.305%,PSNR平均提高了12.455%,边缘强度因子平均提高了7.77%。本文算法有效解决了复原图像中“halo”效应现象和明亮区域颜色失真现象,去雾效果最优。
文摘遥感图像在成像过程中,容易受到云层和雾霾天气的影响,形成带雾图像;同时在下传时,会受到多种因素影响(如发送接收误码、电离层和对流层的随机变化对信号形成扰动等),使图像信息丢失或掺杂噪声。本文针对信息丢失的带雾单色遥感图像,提出了基于矩阵复原和暗通道理论的单色遥感图像去雾算法,通过基于交替方向乘子法(Alternating Direction Method of Multipliers,ADMM)的矩阵复原算法与传统暗通道理论相结合,有效实现了信息丢失下的遥感雾图复原。通过主观评价和客观评价相结合的方式,将本文算法与经典算法对比。结果表明,本文算法得到的结果在直观视觉上效果更好,且相对于信息丢失30%的雾图,6个场景的平均信息熵提升1.6652,平均峰值信噪比提升11.7029,平均结构相似性提升0.8146,客观评价指标结果优异。进一步在不同比例信息丢失情况下进行实验,结果表明,即使在信息大量丢失的情况下,依然能够得到清晰的复原去雾图像。
基金National Natural Science Foundation of China(Nos.61841303,61963023)Project of Humanities and Social Sciences of Ministry of Education in China(No.19YJC760012)。
文摘To solve the problem of color distortion after dehazing in the sky region by using the classical dark channel prior method to process the hazy images with large regions of sky,an improved dark channel image dehazing method based on Gaussian mixture model is proposed.Firstly,we use the Gaussian mixture model to model the hazy image,and then use the expectation maximization(EM)algorithm to optimize the parameters,so that the hazy image can be divided into the sky region and the non-sky region.Secondly,the sky region is divided into a light haze region,a medium haze region and a heavy haze region according to the different dark channel values to estimate the transmission respectively.Thirdly,the restored image is obtained by combining the atmospheric scattering model.Finally,adaptive local tone mapping for high dynamic range images is used to adjust the brightness of the restored image.The experimental results show that the proposed method can effectively eliminate the color distortion in the sky region,and the restored image is clearer and has better visual effect.
文摘针对现有的深度学习去雾算法参数多,训练时间长,无法应用到实时计算机视觉系统等问题,本文提出了一种基于明暗通道的循环GAN网络(bright and dark channel CycleGAN network,BDCCN).BDCCN以CycleGAN为基础,采用固定参数和训练参数相结合方式,基于明暗通道先验理论,改进循环感知损失,实现图像去雾.实验结果表明,本文算法计算量小,收敛快,在合成数据集和真实数据集上均表现优异.
