基于全卷积神经网络的SAR海面溢油图像分割方法
|
摘要
针对合成孔径雷达(SAR)图像中普遍存在斑点噪声和强度不均匀等问题,提出一种基于全卷积神经网络语义分割框架(FCN)的SAR图像溢油分割方法。首先该方法采用迁移学习来提高泛化能力,从而有效地抑制了溢油区域普遍存在的斑噪声和强度不均匀现象;然后采用跳跃式架构来提高溢油区域的分割精度;最后基于一个包含4 200个样本的溢油数据集,将该方法与一些传统机器学习算法(如支持向量机(SVM)、随机森林(RF)和分类回归树(CART)等)和BP神经网络进行对比实验。实验结果表明,该方法相对其他传统方法在像素精度方面提升了7%,针对SAR图像中存在的斑点噪声、强度不均匀及弱边界现象的暗斑分割效果具有显著的改善。
Aiming at speckle noise and intensity unevenness in synthetic aperture radar(SAR) images, an oil spill segmentation method based on Full Convolutional Neural network(FCN) was proposed. Firstly, the migration learning was used to improve the generalization ability, thereby effectively suppressing the universal phenomenon of speckle noise and intensity unevenness in SAR images of oil spill. Then a jump architecture was used to improve the segmentation accuracy. Finally based on an oil spill dataset containing 4 200 samples, the method was compared with some traditional machine learning methods(such as SVM(Support Vector Machine), RF(Random Forest), CART(Classification And Regression Tree)) and BP neural network. The experimental results show that the proposed method improves the pixel accuracy by 7% compared with other traditional methods, and it has significant improvement on dark spot segmentation effect with speckle noise, uneven intensity and weak boundary phenomenon of SAR images.
Aiming at speckle noise and intensity unevenness in synthetic aperture radar(SAR) images, an oil spill segmentation method based on Full Convolutional Neural network(FCN) was proposed. Firstly, the migration learning was used to improve the generalization ability, thereby effectively suppressing the universal phenomenon of speckle noise and intensity unevenness in SAR images of oil spill. Then a jump architecture was used to improve the segmentation accuracy. Finally based on an oil spill dataset containing 4 200 samples, the method was compared with some traditional machine learning methods(such as SVM(Support Vector Machine), RF(Random Forest), CART(Classification And Regression Tree)) and BP neural network. The experimental results show that the proposed method improves the pixel accuracy by 7% compared with other traditional methods, and it has significant improvement on dark spot segmentation effect with speckle noise, uneven intensity and weak boundary phenomenon of SAR images.
引文
[1] ALBAIGéS J.The basics of oil spill cleanup[J].International Journal of Environmental Analytical Chemistry,2014,94(14/15):1512-1514..
[2] GADE M,ALPERS W.Using ERS-2 SAR images for routine observation of marine pollution in European coastal waters[J].Science of the Total Environment,1999,237/238(4):441-448.
[3] FERRARO G,MERYER-ROUX S,MUELLENHOFF O,et al.Long term monitoring of oil spills in European seas[J].International Journal of Remote Sensing,2009,30(3):627-645.
[4] YIN J,YANG J,ZGOU Z S,et al.The extended Bragg scattering model-based method for ship and oil-spill observation using compact polarimetric SAR[J].IEEE Journal of Selected Topics in Applied Earth Observations & Remote Sensing,2015,8(8):3760-3772.
[5] BERN T I,WAHL T,ANDERSEN T,et al.Oil spill detection using satellite based SAR:experience from a field experiment[J].Photogrammetric Engineering & Remote Sensing,1993,59:3(3):423-428.
[6] SOLBERG A H S,STORVIK G,SOLBERG R,et al.Automatic detection of oil spills in ERS SAR images[J].IEEE Transactions on Geoscience & Remote Sensing,1999,37(4):1916-1924.
[7] KANAA T F N,TONYE E,MERCIER G,et al.Detection of oil slick signatures in SAR images by fusion of hysteresis thresholding responses[C]// Proceedings of the 2003 IEEE International Geoscience and Remote Sensing Symposium.Piscataway:IEEE,2003,4:2750-2752.
[8] LI C,KAO C-Y,GORE J C,et al.Minimization of region-scalable fitting energy for image segmentation[J].IEEE Transactions on Image Processing,2008,17(10):1940-1949.
[9] CHAN T F,VESE L A.Active contours without edges[J].IEEE Transactions on Image Processing,2001,10(2):266-277.
[10] TARAVAT A,OPPELT N.Adaptive Weibull multiplicative model and multilayer perceptron neural networks for dark-spot detection from SAR imagery[J].Sensors,2014,14(12):22798-22810.
[11] SINGHA S,BELLERBY T J,TRIESCHMANN O.Satellite oil spill detection using artificial neural networks[J].IEEE Journal of Selected Topics in Applied Earth Observations & Remote Sensing,2013,6(6):2355-2363.
[12] LONG J,SHELHAMER E,DARRELL T.Fully convolutional networks for semantic segmentation[C]// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition.Washington,DC:IEEE Computer Society,2015:3431-3440.
[13] PERSELLO C,STEIN A.Deep fully convolutional networks for the detection of informal settlements in VHR images[J].IEEE Geoscience & Remote Sensing Letters,2017,14(12):2325-2329.
[14] HUANG H,DENG J,LAN Y,et al.A fully convolutional network for weed mapping of unmanned aerial vehicle (UAV) imagery[J].PLoS One,2018,13(4):e0196302.
[15] GUO H,WU D,AN J.Discrimination of oil slicks and lookalikes in polarimetric SAR images using CNN[J].Sensors,2017,17(8):E1837.
[16] RUSSELL B C,TORRALBA A,MURPHY K P,et al.LabelMe:a database and Web-based tool for image annotation[J].International Journal of Computer Vision,2008,77(1/2/3):157-173.
[17] DOMAHUE J,JIA Y,VINYALS O,et al.DeCAF:a deep convolutional activation feature for generic visual recognition[J].arXiv Preprint,2013,2013:arXiv:1310.1531.
[18] SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J].arXiv Preprint,2014,2014:arXiv:1409.1556.
[19] DENG J,DONG W,SOCHER R,et al.ImageNet:A large-scale hierarchical image database[C]// Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition.Washington,DC:IEEE Computer Society,2009:248-255.
[20] 陈杰,张渊智,李煜.基于特征概率函数的双阈值分割海面溢油检测[J].地球信息科学学报,2012,14(4):531-539.
[2] GADE M,ALPERS W.Using ERS-2 SAR images for routine observation of marine pollution in European coastal waters[J].Science of the Total Environment,1999,237/238(4):441-448.
[3] FERRARO G,MERYER-ROUX S,MUELLENHOFF O,et al.Long term monitoring of oil spills in European seas[J].International Journal of Remote Sensing,2009,30(3):627-645.
[4] YIN J,YANG J,ZGOU Z S,et al.The extended Bragg scattering model-based method for ship and oil-spill observation using compact polarimetric SAR[J].IEEE Journal of Selected Topics in Applied Earth Observations & Remote Sensing,2015,8(8):3760-3772.
[5] BERN T I,WAHL T,ANDERSEN T,et al.Oil spill detection using satellite based SAR:experience from a field experiment[J].Photogrammetric Engineering & Remote Sensing,1993,59:3(3):423-428.
[6] SOLBERG A H S,STORVIK G,SOLBERG R,et al.Automatic detection of oil spills in ERS SAR images[J].IEEE Transactions on Geoscience & Remote Sensing,1999,37(4):1916-1924.
[7] KANAA T F N,TONYE E,MERCIER G,et al.Detection of oil slick signatures in SAR images by fusion of hysteresis thresholding responses[C]// Proceedings of the 2003 IEEE International Geoscience and Remote Sensing Symposium.Piscataway:IEEE,2003,4:2750-2752.
[8] LI C,KAO C-Y,GORE J C,et al.Minimization of region-scalable fitting energy for image segmentation[J].IEEE Transactions on Image Processing,2008,17(10):1940-1949.
[9] CHAN T F,VESE L A.Active contours without edges[J].IEEE Transactions on Image Processing,2001,10(2):266-277.
[10] TARAVAT A,OPPELT N.Adaptive Weibull multiplicative model and multilayer perceptron neural networks for dark-spot detection from SAR imagery[J].Sensors,2014,14(12):22798-22810.
[11] SINGHA S,BELLERBY T J,TRIESCHMANN O.Satellite oil spill detection using artificial neural networks[J].IEEE Journal of Selected Topics in Applied Earth Observations & Remote Sensing,2013,6(6):2355-2363.
[12] LONG J,SHELHAMER E,DARRELL T.Fully convolutional networks for semantic segmentation[C]// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition.Washington,DC:IEEE Computer Society,2015:3431-3440.
[13] PERSELLO C,STEIN A.Deep fully convolutional networks for the detection of informal settlements in VHR images[J].IEEE Geoscience & Remote Sensing Letters,2017,14(12):2325-2329.
[14] HUANG H,DENG J,LAN Y,et al.A fully convolutional network for weed mapping of unmanned aerial vehicle (UAV) imagery[J].PLoS One,2018,13(4):e0196302.
[15] GUO H,WU D,AN J.Discrimination of oil slicks and lookalikes in polarimetric SAR images using CNN[J].Sensors,2017,17(8):E1837.
[16] RUSSELL B C,TORRALBA A,MURPHY K P,et al.LabelMe:a database and Web-based tool for image annotation[J].International Journal of Computer Vision,2008,77(1/2/3):157-173.
[17] DOMAHUE J,JIA Y,VINYALS O,et al.DeCAF:a deep convolutional activation feature for generic visual recognition[J].arXiv Preprint,2013,2013:arXiv:1310.1531.
[18] SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J].arXiv Preprint,2014,2014:arXiv:1409.1556.
[19] DENG J,DONG W,SOCHER R,et al.ImageNet:A large-scale hierarchical image database[C]// Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition.Washington,DC:IEEE Computer Society,2009:248-255.
[20] 陈杰,张渊智,李煜.基于特征概率函数的双阈值分割海面溢油检测[J].地球信息科学学报,2012,14(4):531-539.