迁移学习方法提取高分一号影像汶川地震震后滑坡
Transfer learning method for landslide extraction from GF-1 images after the Wenchuan earthquake
- 2023年27卷第8期 页码：1866-1875
DOI： 10.11834/jrs.20211020
引用
扫 描 看 全 文
摘要
2008年汶川8.0级地震触发了大量的崩塌滑坡地质灾害，导致强震区震后地质灾害频发，因其对生命和财产的巨大威胁而广泛关注。利用遥感等技术快速提取滑坡信息，对于减少灾害造成的损失具有重要的现实意义。本文提出一种迁移学习方法，从自然场景数据集中学习特征，迁移到滑坡提取中。该方法首先在ImageNet上预训练ResNet网络，然后输入滑坡区影像样本，将预训练网络及参数迁移至LinkNet上，最终实现滑坡提取。通过对2013年—2015年3景影像的汶川地震震后滑坡提取实验进行分析及验证，结果显示相较于传统支持向量机和其他深度学习方法，本文提出的迁移学习方法有较优的提取精度，有利于后续研判及决策。
Abstract
Landslides are natural disasters that are driven by various factors and often leave catastrophic damages and casualties. A huge earthquake can trigger plenty of landslides. Therefore, landslide extraction is critical to provide timely information for post-disaster decision making. Remote sensing is a convenient tool for landslide information acquisition. However, landslide features are so intricate that landslide extraction mainly relies on a visual interpretation of aerial photographs or high-resolution remote sensing images, which requires vast manpower. Several landslide extraction methods are available today, including pixel-based methods, which have relatively low accuracy, and object-oriented methods, whose parameters need to be decided subjectively. With the continuous development of deep learning in image semantic segmentation, a precise and automatic remote sensing image binary classification becomes possible. Many researchers have investigated the use of deep learning for landslide extraction in different areas. However, a relatively small amount of landslide data can easily lead to model overfitting. Transfer learning, where knowledge is transferred from the source domain to the target domain, can alleviate this problem by using knowledge in the source domain to improve performance in the target task. A transfer learning deep network is then designed to improve the accuracy of landslide extraction.,First, three GF-1 images taken from 2013 to 2015 in the research area were processed successively by geometric correction, registration, and image fusion to obtain 4 bands of images (red, green, blue, and near-infrared) with a resolution of 2 m. Second, a proper network was designed. The encoder of ResNet that was trained on ImageNet was chosen as the encoder, and the decoder of LinkNet, whose residual structure and bypass links can improve performance, was selected as the decoder. The bypass links in the decoder can also address the spatial information loss in max-pooling in the encoder, and the residual structure allows the network to learn complex features. After pre-training the ResNet network on ImageNet, we adjusted the number of input channels of the first convolution layer to 4, drop the last fully connected layer, and then form our network with the decoder. We eventually inputted remote sensing landslide images to finetune our model.,When testing different network depths, the network does not always perform better as the depth increases. We chose ResNet50 as our encoder given its peak performance. Afterward, we compared our method with SVM. Without a pre-training encoder, our network improves U-Net and a mainstream transfer learning method, AlbuNet, thereby suggesting that deep learning methods outperform SVM, whereas transfer learning methods outperform deep learning methods that are trained on landslide images. The proposed method outperforms SVM in terms of precision, recall, and F1 measure by 17.16%, 18.58%, and 17.4%, respectively, outperforms the improved U-Net by 2.98%, 6.35%, and 4.61%, and outperforms AlbuNet by 0.9%, 1.98%, and 1.48%.,The ResNet50 encoder combined with the LinkNet decoder should be selected to form a landslide extraction network with a higher accuracy compared with the transfer learning network AlbuNet and other ResNet encoders of different depths. Transferring knowledge learned from ImageNet can also improve the performance of the landslide extraction deep learning network. The proposed method can be used conveniently for follow-up landslide risk assessment, disaster investigation, disaster warning, and decision making.
关键词
遥感滑坡提取迁移学习ImageNet高分一号
Keywords
remote sensinglandslide extractiontransfer learningImageNetGF-1
references
An L Q, Zhang J F and Zhao F J. 2011. Extracting secondary disaster of Wenchuan earthquake: application of object-oriented image-classifying technology. Journal of Natural Disasters, 20(2): 160-168
安立强, 张景发, 赵福军. 2011. 汶川地震次生灾害提取——面向对象影像分类技术的应用. 自然灾害学报, 20(2): 160-168 [DOI: 10.13577/j.jnd.2011.0225http://dx.doi.org/10.13577/j.jnd.2011.0225]
Dai W Y, Xue G R, Yang Q and Yu Y. 2007. Co-clustering based classification for out-of-domain documents//Proceedings of the 13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Jose: ACM: 210-219 [DOI: 10.1145/1281192.1281218http://dx.doi.org/10.1145/1281192.1281218]
Fan J R, Zhang X Y, Su F H, Ge Y G, Tarolli P, Yang Z Y, Zeng C and Zeng Z. 2017. Geometrical feature analysis and disaster assessment of the Xinmo landslide based on remote sensing data. Journal of Mountain Science, 14(9): 1677-1688 [DOI: 10.1007/s11629-017-4633-3http://dx.doi.org/10.1007/s11629-017-4633-3]
He K M, Zhang X Y, Ren S Q and Sun J. 2016. Deep residual learning for image recognition//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE: 770-778 [DOI: 10.1109/CVPR.2016.90http://dx.doi.org/10.1109/CVPR.2016.90]
Huang T, Bai X F, Zhuang Q F and Xu J H. 2018. Research on landslides extraction based on the Wenchuan earthquake in GF-1 remote sensing image. Bulletin of Surveying and Mapping, 2: 67-71, 82
黄汀, 白仙富, 庄齐枫, 徐敬海. 2018. 高分一号汶川极震区滑坡提取研究. 测绘通报, (2): 67-71, 82 [DOI: 10.13474/j.cnki.11-2246.2018.0046http://dx.doi.org/10.13474/j.cnki.11-2246.2018.0046]
LeCun Y, Bengio Y and Hinton G. 2015. Deep learning. Nature, 521(7553): 436-444 [DOI: 10.1038/nature14539http://dx.doi.org/10.1038/nature14539]
Li Q, Zhang J F, Luo Y and Jiao Q S. 2019. Recognition of earthquake-induced landslide and spatial distribution patterns triggered by the Jiuzhaigou earthquake in August 8, 2017. Journal of Remote Sensing, 23(4): 785-795
李强, 张景发, 罗毅, 焦其松. 2019. 2017年“8. 8”九寨沟地震滑坡自动识别与空间分布特征. 遥感学报, 23(4): 785-795 [DOI: 10.11834/jrs.20197345http://dx.doi.org/10.11834/jrs.20197345]
Li S, Deng B K, Xu H Q and Wang Z F. 2015. Fast interpretation methods of landslides triggered by earthquake using remote sensing imagery. Remote Sensing Information, 30(4): 25-28
李松, 邓宝昆, 徐红勤, 王治福. 2015. 地震型滑坡灾害遥感快速识别方法研究. 遥感信息, 30(4): 25-28 [DOI: 10.3969/j.issn.1000-3177.2015.04.005http://dx.doi.org/10.3969/j.issn.1000-3177.2015.04.005]
Li W L, Wu J and Lü B X. 2011. Research on landslide triggered by earthquake: review and prospect. Journal of Catastrophology, 26(3): 103-108
李为乐, 伍霁, 吕宝雄. 2011. 地震滑坡研究回顾与展望. 灾害学, 26(3): 103-108 [DOI: 10.3969/j.issn.1000-811X.2011.03.021http://dx.doi.org/10.3969/j.issn.1000-811X.2011.03.021]
Liu P, Wei Y M, Wang Q J, Chen Y and Xie J J. 2020. Research on post-earthquake landslide extraction algorithm based on improved U-net model. Remote sensing, 12(5): 894 [DOI: 10.3390/rs12050894http://dx.doi.org/10.3390/rs12050894]
Liu Y and Wu L Z. 2016. Geological disaster recognition on optical remote sensing images using deep learning. Procedia Computer Science, 91: 566-575 [DOI: 10.1016/j.procs.2016.07.144http://dx.doi.org/10.1016/j.procs.2016.07.144]
Mondini A C, Guzzetti F, Reichenbach P, Rossi M, Cardinali M and Ardizzone F. 2011. Semi-automatic recognition and mapping of rainfall induced shallow landslides using optical satellite images. Remote Sensing of Environment, 115(7): 1743-1757 [DOI: 10.1016/j.rse.2011.03.006http://dx.doi.org/10.1016/j.rse.2011.03.006]
Pan S J and Yang Q. 2010. A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 22(10): 1345-1359 [DOI: 10.1109/tkde.2009.191http://dx.doi.org/10.1109/tkde.2009.191]
Penatti O A B, Nogueira K and Santos J A D. 2015. Do deep features generalize from everyday objects to remote sensing and aerial scenes domains?//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Boston: IEEE: 44-51 [DOI: 10.1109/CVPRW.2015.7301382http://dx.doi.org/10.1109/CVPRW.2015.7301382]
Peng L, Xu S N, Mei J J and Su F H. 2017. Earthquake-induced landslide recognition using high-resolution remote sensing images. Journal of Remote Sensing, 21(4): 509-518
彭令, 徐素宁, 梅军军, 苏凤环. 2017. 地震滑坡高分辨率遥感影像识别. 遥感学报, 21(4): 509-518 [DOI: 10.11834/jrs.20176176http://dx.doi.org/10.11834/jrs.20176176]
Raina R, Ng A Y and Koller D. 2006. Constructing informative priors using transfer learning//Proceedings of the 23rd International Conference on Machine Learning. Pittsburgh: ACM: 713-720 [DOI: 10.1145/1143844.1143934http://dx.doi.org/10.1145/1143844.1143934]
Shvets A A, Rakhlin A, Kalinin A A and Iglovikov V I. 2018. Automatic instrument segmentation in robot-assisted surgery using deep learning//2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA). Orlando: IEEE: 624-628 [DOI: 10.1109/ICMLA.2018.00100http://dx.doi.org/10.1109/ICMLA.2018.00100]
Su F H, Liu H J and Han Y S. 2008. The extraction of mountain hazard induced by Wenchuan earthquake and analysis of its distributing characteristic. Journal of Remote Sensing, 12(6): 956-963
苏凤环, 刘洪江, 韩用顺. 2008. 汶川地震山地灾害遥感快速提取及其分布特点分析. 遥感学报, 12(6): 956-963 [DOI: 10.11834/jrs.200806128http://dx.doi.org/10.11834/jrs.200806128]
Wang Y, Wang X F and Jian J F. 2019. Remote sensing landslide recognition based on convolutional neural network. Mathematical Problems in Engineering, 2019: 8389368 [DOI: 10.1155/2019/8389368http://dx.doi.org/10.1155/2019/8389368]
Wang Y, Yang Y, Wang B S, Wang T, Bu X H and Wang C Y. 2019. Building segmentation in high-resolution remote sensing image through deep neural network and conditional random fields. Journal of Remote Sensing, 23(6): 1194-1208
王宇, 杨艺, 王宝山, 王田, 卜旭辉, 王传云. 2019. 深度神经网络条件随机场高分辨率遥感图像建筑物分割. 遥感学报, 23(6): 1194-1208 [DOI: 10.11834/jrs.20198141http://dx.doi.org/10.11834/jrs.20198141]
Wang Z H. 1999. Reviewing and prospecting for applying remote sensing to landslide and debris flow investigation. Remote Sensing for Land and Resources, 41(3): 10-15, 39
王治华. 1999. 滑坡、泥石流遥感回顾与新技术展望. 国土资源遥感, 41(3): 10-15, 39 [DOI: 10.3969/j.issn.1001-070X.1999.03.004http://dx.doi.org/10.3969/j.issn.1001-070X.1999.03.004]
Ye R Q, Deng Q L and Wang H Q. 2007. Landslides identification based on image classification: a case study on Guizhoulaocheng landslide in the three gorges reservoir area. Chinese Journal of Engineering Geophysics, 4(6): 574-577
叶润青, 邓清禄, 王海庆. 2007. 基于图像分类方法滑坡识别与特征提取——以归州老城滑坡为例. 工程地球物理学报, 4(6): 574-577 [DOI: 10.3969/j.issn.1672-7940.2007.06.012http://dx.doi.org/10.3969/j.issn.1672-7940.2007.06.012]
Yin J, Yang Q and Ni L. 2005. Adaptive temporal radio maps for indoor location estimation//Third IEEE International Conference on Pervasive Computing and Communications. Kauai: IEEE: 85-94 [DOI: 10.1109/PERCOM.2005.7http://dx.doi.org/10.1109/PERCOM.2005.7]
Yu H, Ma Y, Wang L F, Zhai Y S and Wang X Q. 2017. A landslide intelligent detection method based on CNN and RSG_R//2017 IEEE International Conference on Mechatronics and Automation (ICMA). Takamatsu: IEEE: 40-44 [DOI: 10. 1109/ICMA. 2017. 8015785http://dx.doi.org/10.1109/ICMA.2017.8015785]
文章被引用时，请邮件提醒。
提交
相关作者
暂无数据
相关机构
中国科学院空天信息创新研究院 遥感科学国家重点实验室
中国科学院空天信息创新研究院 遥感科学国家重点实验室
中国科学院大学 资源与环境学院
华中农业大学 资源与环境学院 宏观农业研究院
中国科学院空天信息创新研究院 数字地球重点实验室
- 地址：北京市海淀区北四环西路19号 《遥感学报》编辑部 邮编：100190
- 联系电话：010-58887052 Email：nrsb@aircas.ac.cn
- 技术支持由北京北大方正电子有限公司提供 京ICP备20021838号-9
京公网安备11010802024621 - 本系统建议在Chrome、 IE9+ 以上版本浏览器阅读本站内容，360浏览器请切换至极速模式
- Cookies帮助我们提供服务并提供个性化体验。使用本网站，即表示您同意我们使用Cookies
0
回到顶部