基于深度学习的图像去雾算法
|
摘要
针对有雾天气会使图像质量降低、影响图像信息的提取、导致图像的应用价值减少的问题,基于深度学习提出了一种多尺度卷积神经网络,并基于此多尺度卷积神经网络给出了新的图像去雾算法:使用该多尺度卷积神经网络对原有雾图像进行特征提取、特征融合并最终实现图像细节的重建,得到粗透射率图;利用原有雾图像中像素点的位置和亮度值得到大气光值;利用导向滤波得到细透射率图,并结合大气光值反演出无雾图像;最后对无雾图像进行直方图颜色校正.实验结果表明:相比传统去雾算法,新的图像去雾算法对图像细节的处理更加自然且具有很好的视觉效果.
To solve the problem of foggy weather damaging the image quality,affecting the extraction of image information and reducing its applicability,an image defogging algorithm based on deep learning is proposed. Firstly,the original fog image is subjected to single-scale and multi-scale convolution for feature extraction. Then the multiscale convolution kernel is used to reconstruct the image detail to obtain a rough transmittance propagation map,the position and brightness values of the pixel in the original fog image are used to obtain the atmospheric light value,and the guided transmission is used to obtain the fine transmittance propagation map and to invert the fog-free image with the previously obtained atmospheric light value. Finally,the histogram color correction is performed on the fog-free image. The experimental results show that,compared with the traditional dehazing algorithm,the algorithm is more natural and has a good visual effect.
To solve the problem of foggy weather damaging the image quality,affecting the extraction of image information and reducing its applicability,an image defogging algorithm based on deep learning is proposed. Firstly,the original fog image is subjected to single-scale and multi-scale convolution for feature extraction. Then the multiscale convolution kernel is used to reconstruct the image detail to obtain a rough transmittance propagation map,the position and brightness values of the pixel in the original fog image are used to obtain the atmospheric light value,and the guided transmission is used to obtain the fine transmittance propagation map and to invert the fog-free image with the previously obtained atmospheric light value. Finally,the histogram color correction is performed on the fog-free image. The experimental results show that,compared with the traditional dehazing algorithm,the algorithm is more natural and has a good visual effect.
引文
[1]吴迪,朱青松.图像去雾的最新研究进展[J].自动化学报,2015,41(2):221-239.WU D,ZHU Q S. The latest research progress of image dehazing[J]. Acta Automatica Sinica,2015,41(2):221-239.
[2]禹晶,李大鹏,廖庆敏.基于物理模型的快速单幅图像去雾方法[J].自动化学报,2011,37(2):143-149.YU J,LI D P,LIAO Q M. Physics-based fast single image fog removal[J]. Acta Automatica Sinica,2011,37(2):143-149.
[3]王万国,王滨海,张晶晶,等.基于直方图规定化的图像去雾算法[J].计算机技术与发展,2014,24(9):241-244.WANG W G,WANG B H,ZHANG J J,et al. Image haze removal algorithm based on histogram specification[J].Computer Technology and Development,2014,24(9):241-244.
[4]张赛楠,吴亚东,张红英,等.改进的单尺度Retinex雾天图像增强算法[J].激光与红外,2013,43(6):698-702.ZHANG S N,WU Y D,ZHANG H Y,et al. Haze image enhancement method based on improved single-scale Retinex[J]. Laser&Infrared,2013,43(6):698-702.
[5]孙晓晓,杨峰.基于小波变换融合技术的去雾方法[J].山东师范大学学报(自然科学版),2016,31(2):40-43.SUN X X,YANG F. Defogging based on wavelet transform fusion technology[J]. Journal of Shandong Normal University(Natural Science),2016,31(2):40-43.
[6] TAN R T. Visibility in bad weather from a single from a single image[C]∥Proceedings of IEEE Computer Vision and Pattern Recognition. Anchorage,Alaska:IEEE,2008:1956-1963.
[7] FATTAL R. Single image dehazing[J]. ACM Transactions on Graphics,2008,27(3):721-728.
[8] HE K M,SUN J,TANG X O. Single image haze removal using dark channel prior[C]∥Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Miami,Florida:IEEE,2009:1956-1963.
[9] HE K M,SUN J,TANG X O. Guided image filtering[C]∥Proceeding of the 11th European Conference on Computer Vision:Part I. Berlin:Springer,2010:1-14.
[10] MENG G,WANG Y,DUAN J,et al. Efficient image dehazing with boundary constraint and contextual regularization[C]∥Proceeding of IEEE International Conference on Computer Vision. Washington:IEEE,2014:617-624.
[11] TANG K,YANG J,WANG J. Investigating haze-relevant features in a learning framework for image dehazing[C]∥Proceeding of IEEE Conference on Computer Vision and Pattern Recognition. Columbus:IEEE,2014:2995-3002.
[12] CAI B,XU X,JIA K,et al. Dehaze Net:an end-to-end system for single image haze removal[J]. IEEE Transactions on Image Processing,2016,25(11):5187-5198.
[13] MCCARTNEY E J. Optics of the atmosphere:scattering by molecules and particles[M]. New Jersey:Wiley,1976.
[14] NARASIMHAN S G,NAYAR S K. Contrast restoration of weather degraded images[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2003,25(6):713-724.
[15] NARASIMHAN S G,NAYAR S K. Vision and the atmosphere[J]. International Journal of Computer Vision,2002,48(3):233-254.
[16]周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(6):1229-1251.ZHOU F Y,JIN L P,DONG J. Review of convolutional neural network[J]. Chinese Journal of Computers,2017,40(6):1229-1251.
[17]郭璠,蔡自兴.图像去雾算法清晰化效果客观评价方法[J].自动化学报,2012,38(9):1410-1419.GUO F,CAI Z X. Objective assessment method for the clearness effect of image defogging algorithm[J]. Acta Automatica Sinica,2012,38(9):1410-1419.
[2]禹晶,李大鹏,廖庆敏.基于物理模型的快速单幅图像去雾方法[J].自动化学报,2011,37(2):143-149.YU J,LI D P,LIAO Q M. Physics-based fast single image fog removal[J]. Acta Automatica Sinica,2011,37(2):143-149.
[3]王万国,王滨海,张晶晶,等.基于直方图规定化的图像去雾算法[J].计算机技术与发展,2014,24(9):241-244.WANG W G,WANG B H,ZHANG J J,et al. Image haze removal algorithm based on histogram specification[J].Computer Technology and Development,2014,24(9):241-244.
[4]张赛楠,吴亚东,张红英,等.改进的单尺度Retinex雾天图像增强算法[J].激光与红外,2013,43(6):698-702.ZHANG S N,WU Y D,ZHANG H Y,et al. Haze image enhancement method based on improved single-scale Retinex[J]. Laser&Infrared,2013,43(6):698-702.
[5]孙晓晓,杨峰.基于小波变换融合技术的去雾方法[J].山东师范大学学报(自然科学版),2016,31(2):40-43.SUN X X,YANG F. Defogging based on wavelet transform fusion technology[J]. Journal of Shandong Normal University(Natural Science),2016,31(2):40-43.
[6] TAN R T. Visibility in bad weather from a single from a single image[C]∥Proceedings of IEEE Computer Vision and Pattern Recognition. Anchorage,Alaska:IEEE,2008:1956-1963.
[7] FATTAL R. Single image dehazing[J]. ACM Transactions on Graphics,2008,27(3):721-728.
[8] HE K M,SUN J,TANG X O. Single image haze removal using dark channel prior[C]∥Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Miami,Florida:IEEE,2009:1956-1963.
[9] HE K M,SUN J,TANG X O. Guided image filtering[C]∥Proceeding of the 11th European Conference on Computer Vision:Part I. Berlin:Springer,2010:1-14.
[10] MENG G,WANG Y,DUAN J,et al. Efficient image dehazing with boundary constraint and contextual regularization[C]∥Proceeding of IEEE International Conference on Computer Vision. Washington:IEEE,2014:617-624.
[11] TANG K,YANG J,WANG J. Investigating haze-relevant features in a learning framework for image dehazing[C]∥Proceeding of IEEE Conference on Computer Vision and Pattern Recognition. Columbus:IEEE,2014:2995-3002.
[12] CAI B,XU X,JIA K,et al. Dehaze Net:an end-to-end system for single image haze removal[J]. IEEE Transactions on Image Processing,2016,25(11):5187-5198.
[13] MCCARTNEY E J. Optics of the atmosphere:scattering by molecules and particles[M]. New Jersey:Wiley,1976.
[14] NARASIMHAN S G,NAYAR S K. Contrast restoration of weather degraded images[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2003,25(6):713-724.
[15] NARASIMHAN S G,NAYAR S K. Vision and the atmosphere[J]. International Journal of Computer Vision,2002,48(3):233-254.
[16]周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(6):1229-1251.ZHOU F Y,JIN L P,DONG J. Review of convolutional neural network[J]. Chinese Journal of Computers,2017,40(6):1229-1251.
[17]郭璠,蔡自兴.图像去雾算法清晰化效果客观评价方法[J].自动化学报,2012,38(9):1410-1419.GUO F,CAI Z X. Objective assessment method for the clearness effect of image defogging algorithm[J]. Acta Automatica Sinica,2012,38(9):1410-1419.