基于Encoder-Decoder网络的遥感影像道路提取方法
|
摘要
针对道路目标特点,设计实现了用于遥感影像道路提取的Encoder-Decoder深度语义分割网络。首先,针对道路目标局部特征丰富、语义特征较为简单的特点,设计了较浅深度、分辨率较高的Encoder-Decoder网络结构,提高了分割网络的细节表示能力。其次,针对遥感影像中道路目标所占像素比例较小的特点,改进了二分类交叉熵损失函数,解决了网络训练中正负样本严重失衡的问题。在大型道路提取数据集上的试验表明,所提方法召回率、精度和F1-score指标分别达到了83.9%、82.5%及82.9%,能够完整准确地提取遥感影像中的道路目标。所设计的Encoder-Decoder网络性能优良,且不需人工设计提取特征,因而具有良好的应用前景。
According to the characteristics of the road features, an Encoder-Decoder deep semantic segmentation network is designed for road extraction of remote sensing images. Firstly, as the features of the road target are rich in local details and simple in semantic features, an Encoder-Decoder network with shallow layers and high resolution is designed to improve the ability of representing detail information. Secondly, as the road area is small proportion in remote sensing images, the cross-entropy loss function is improved, which solves the imbalance between positive and negative samples in training process. Experiments on large road extraction dataset show that, the proposed method gets the recall rate 83.9%, precision 82.5% and F1-score 82.9%, which can extract the road targets in remote sensing images completely and accurately. The Encoder-Decoder network designed in this paper performs well in road extraction task and needs less artificial participation, so it has a good application prospect.
According to the characteristics of the road features, an Encoder-Decoder deep semantic segmentation network is designed for road extraction of remote sensing images. Firstly, as the features of the road target are rich in local details and simple in semantic features, an Encoder-Decoder network with shallow layers and high resolution is designed to improve the ability of representing detail information. Secondly, as the road area is small proportion in remote sensing images, the cross-entropy loss function is improved, which solves the imbalance between positive and negative samples in training process. Experiments on large road extraction dataset show that, the proposed method gets the recall rate 83.9%, precision 82.5% and F1-score 82.9%, which can extract the road targets in remote sensing images completely and accurately. The Encoder-Decoder network designed in this paper performs well in road extraction task and needs less artificial participation, so it has a good application prospect.
引文
[1] 吴亮, 胡云安. 遥感图像自动道路提取方法综述[J]. 自动化学报, 2010, 36(7): 912-922.WU Liang, HU Yunan. A survey of automatic road extraction from remote sensing image[J]. Acta Automatica Sinica, 2010, 36(7): 912-922.
[2] ZHOU Y T, VENKATESWAR V, CHELLAPPA R. Edge detection and linear feature extraction using a 2-D random field model[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989, 11(1): 84-95.
[3] STEGER C, GLOCK C, ECKSTEIN W, et al. Model-based road extraction from images[C]//Proceedings of Automatic Extraction of Man-Made Objects from Aerial and Space Images. Basel, Schweiz: Springer, 1995: 275-284.
[4] STEGER C. An unbiased detector of curvilinear structures[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(2): 113-125.
[5] TUPIN F, MAITRE H, MANGIN J F, et al. Detection of linear features in SAR images: application to road network extraction[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998, 36(2): 434-453.
[6] SHI Wenzhong, ZHU Changqing. The line segment match method for extracting road network from high-resolution satellite images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2002, 40(2): 511-514.
[7] HU Jiuxiang, RAZDAN A, FEMIANI J C, et al. Road network extraction and intersection detection from aerial images by tracking road footprints[J]. IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(12): 4144-4157.
[8] CHRISTOPHE E, INGLADA J. Robust road extraction for high resolution satellite images[C]//Proceedings of 2007 IEEE International Conference on Image Processing. San Antonio, TX: IEEE, 2007, 5: 437-440.
[9] POULLIS C, YOU Suya. Delineation and geometric modeling of road networks[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2010, 65(2): 165-181.
[10] 付仲良, 杨元维, 高贤君, 等. 道路网多特征匹配优化算法[J]. 测绘学报, 2016, 45(5): 608-615. DOI: 10.11947/j.AGCS.2016.20150388.FU Zhongliang, YANG Yuanwei, GAO Xianjun, et al. An optimization algorithm for multi-characteristics road network matching[J]. Acta Geodaetica et Cartographica Sinica, 2016, 45(5): 608-615. DOI: 10.11947/j.AGCS.2016.20150388.
[11] MNIH V, HINTON G E. Learning to detect roads in high-resolution aerial images[C]//Proceedings of the 11th European Conference on Computer Vision. Berlin: Springer, 2010: 210-223.
[12] GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH: IEEE, 2014: 580-587.
[13] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[C]//Proceedings of the 3rd International Conference on Learning Representations. San Diego:[s.n.], 2015.
[14] SZEGEDY C, LIU Wei, JIA Yangqing, et al. Going deeper with convolutions[C]//Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Boston, MA: IEEE, 2015: 1-9.
[15] SHELHAMER E, LONG J, DARRELL T. Fully convolutional networks for semantic segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(4): 640-651.
[16] NOH H, HONG S, GAN B. Learning deconvolution network for semantic segmentation[C]//Proceedings of 2015 IEEE International Conference on Computer Vision (ICCV). Santiago, Chile: IEEE, 2015: 1520-1528.
[17] BADRINARAYANAN V, KENDALL A, CIPOLLA R. SegNet: a deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(12): 2481-2495.
[18] RONNEBERGER O, FISCHER P, BROX T. U-Net: convolutional networks for biomedical image segmentation[C]∥Proceedings of Medical Image Computing and Computer-Assisted Intervention. Cham: Springer, 2015, 234-241.
[19] MNIH V, HINTON G E. Learning to label aerial images from noisy data[C]//Proceedings of the 29th International Conference on Machine Learning. Edinburgh, Scotland: ACM, 2012: 567-574.
[20] MNIH V. Machine learning for aerial image labeling[D]. Toronto: University of Toronto, 2013.
[21] WANG Jun, SONG Jingwei, CHEN Mingquan, et al. Road network extraction: a neural-dynamic framework based on deep learning and a finite state machine[J]. International Journal of Remote Sensing, 2015, 36(12): 3144-3169.
[22] WEI Yanan, WANG Zulin, XU Mai. Road structure refined CNN for road extraction in aerial image[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(5): 709-713.
[23] PANBOONYUEN T, VATEEKUL P, JITKAJORNWANICH K, et al. An enhanced deep convolutional Encoder-Decoder network for road segmentation on aerial imagery[C]//Proceedings of Recent Advances in Information and Communication Technology 2017. Bangkok: Springer, 2017: 191-201.
[24] CLEVERT D A, UNTERTHINER T, HZOCHREITER S. Fast and accurate deep network learning by exponential linear units (ELUs)[C]//Proceedings of International Conference on Learning Representations. San Juan:[s.n.], 2016.
[25] CHENG Guangliang, WANG Ying, XU Shibiao, et al. Automatic road detection and centerline extraction via cascaded end-to-end convolutional neural network[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(6): 3322-3337.
[26] KINGMA D P, BA J L. ADAM: a method for stochastic optimization[C]//Proceedings of International Conference on Learning Representations. San Diego:[s.n.], 2015.
[27] WIEDEMANN C. External evaluation of road networks[J]. Proceedings of the ISPRS Workshop on Photogrammetric Image Analysis, Munich, Germany, 2003, 34(3): 93-98.
[28] WEGNER J D, MONTOYA-ZEGARRA J A, SCHINDLER K. A higher-order CRF model for road network extraction[C]//Proceedings of 2013 IEEE Conference on Computer Vision and Pattern Recognition. Portland, OR: IEEE, 2013: 1698-1705.
[29] ZHONG Zilong, LI J, CUI Weihong, et al. Fully convolutional networks for building and road extraction: preliminary results[C]//Proceedings of 2016 IEEE International Geoscience and Remote Sensing Symposium. Beijing, China: IEEE, 2016: 1591-1594.
[2] ZHOU Y T, VENKATESWAR V, CHELLAPPA R. Edge detection and linear feature extraction using a 2-D random field model[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989, 11(1): 84-95.
[3] STEGER C, GLOCK C, ECKSTEIN W, et al. Model-based road extraction from images[C]//Proceedings of Automatic Extraction of Man-Made Objects from Aerial and Space Images. Basel, Schweiz: Springer, 1995: 275-284.
[4] STEGER C. An unbiased detector of curvilinear structures[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(2): 113-125.
[5] TUPIN F, MAITRE H, MANGIN J F, et al. Detection of linear features in SAR images: application to road network extraction[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998, 36(2): 434-453.
[6] SHI Wenzhong, ZHU Changqing. The line segment match method for extracting road network from high-resolution satellite images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2002, 40(2): 511-514.
[7] HU Jiuxiang, RAZDAN A, FEMIANI J C, et al. Road network extraction and intersection detection from aerial images by tracking road footprints[J]. IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(12): 4144-4157.
[8] CHRISTOPHE E, INGLADA J. Robust road extraction for high resolution satellite images[C]//Proceedings of 2007 IEEE International Conference on Image Processing. San Antonio, TX: IEEE, 2007, 5: 437-440.
[9] POULLIS C, YOU Suya. Delineation and geometric modeling of road networks[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2010, 65(2): 165-181.
[10] 付仲良, 杨元维, 高贤君, 等. 道路网多特征匹配优化算法[J]. 测绘学报, 2016, 45(5): 608-615. DOI: 10.11947/j.AGCS.2016.20150388.FU Zhongliang, YANG Yuanwei, GAO Xianjun, et al. An optimization algorithm for multi-characteristics road network matching[J]. Acta Geodaetica et Cartographica Sinica, 2016, 45(5): 608-615. DOI: 10.11947/j.AGCS.2016.20150388.
[11] MNIH V, HINTON G E. Learning to detect roads in high-resolution aerial images[C]//Proceedings of the 11th European Conference on Computer Vision. Berlin: Springer, 2010: 210-223.
[12] GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH: IEEE, 2014: 580-587.
[13] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[C]//Proceedings of the 3rd International Conference on Learning Representations. San Diego:[s.n.], 2015.
[14] SZEGEDY C, LIU Wei, JIA Yangqing, et al. Going deeper with convolutions[C]//Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Boston, MA: IEEE, 2015: 1-9.
[15] SHELHAMER E, LONG J, DARRELL T. Fully convolutional networks for semantic segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(4): 640-651.
[16] NOH H, HONG S, GAN B. Learning deconvolution network for semantic segmentation[C]//Proceedings of 2015 IEEE International Conference on Computer Vision (ICCV). Santiago, Chile: IEEE, 2015: 1520-1528.
[17] BADRINARAYANAN V, KENDALL A, CIPOLLA R. SegNet: a deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(12): 2481-2495.
[18] RONNEBERGER O, FISCHER P, BROX T. U-Net: convolutional networks for biomedical image segmentation[C]∥Proceedings of Medical Image Computing and Computer-Assisted Intervention. Cham: Springer, 2015, 234-241.
[19] MNIH V, HINTON G E. Learning to label aerial images from noisy data[C]//Proceedings of the 29th International Conference on Machine Learning. Edinburgh, Scotland: ACM, 2012: 567-574.
[20] MNIH V. Machine learning for aerial image labeling[D]. Toronto: University of Toronto, 2013.
[21] WANG Jun, SONG Jingwei, CHEN Mingquan, et al. Road network extraction: a neural-dynamic framework based on deep learning and a finite state machine[J]. International Journal of Remote Sensing, 2015, 36(12): 3144-3169.
[22] WEI Yanan, WANG Zulin, XU Mai. Road structure refined CNN for road extraction in aerial image[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(5): 709-713.
[23] PANBOONYUEN T, VATEEKUL P, JITKAJORNWANICH K, et al. An enhanced deep convolutional Encoder-Decoder network for road segmentation on aerial imagery[C]//Proceedings of Recent Advances in Information and Communication Technology 2017. Bangkok: Springer, 2017: 191-201.
[24] CLEVERT D A, UNTERTHINER T, HZOCHREITER S. Fast and accurate deep network learning by exponential linear units (ELUs)[C]//Proceedings of International Conference on Learning Representations. San Juan:[s.n.], 2016.
[25] CHENG Guangliang, WANG Ying, XU Shibiao, et al. Automatic road detection and centerline extraction via cascaded end-to-end convolutional neural network[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(6): 3322-3337.
[26] KINGMA D P, BA J L. ADAM: a method for stochastic optimization[C]//Proceedings of International Conference on Learning Representations. San Diego:[s.n.], 2015.
[27] WIEDEMANN C. External evaluation of road networks[J]. Proceedings of the ISPRS Workshop on Photogrammetric Image Analysis, Munich, Germany, 2003, 34(3): 93-98.
[28] WEGNER J D, MONTOYA-ZEGARRA J A, SCHINDLER K. A higher-order CRF model for road network extraction[C]//Proceedings of 2013 IEEE Conference on Computer Vision and Pattern Recognition. Portland, OR: IEEE, 2013: 1698-1705.
[29] ZHONG Zilong, LI J, CUI Weihong, et al. Fully convolutional networks for building and road extraction: preliminary results[C]//Proceedings of 2016 IEEE International Geoscience and Remote Sensing Symposium. Beijing, China: IEEE, 2016: 1591-1594.