Abstract
Extracting roads from multi-source data, such as aerial images and vehicle trajectories, is an important way to maintain road networks in the filed of urban computing. In this paper, we revisit the problem of road extraction and aim to boost its accuracy by solving three significant issues: the insufficient complementarity among multiple sources, rough edges of extracted roads, and many false positives caused by confusing pixels. In particular, we design an end-to-end neural network model to achieve this goal. At first, this model leverages two encoding networks to extract relative information from the inputs of two sources respectively, and then applies the attention mechanism to fuse them for sufficiently capturing the complementary correlation. Next, we introduce an auxiliary task, predicting road edges based on fused representations, to make the extracted roads smooth and continuous. At last, to reduce false positives relative to confusing pixels, we propose a pixel-aware contrastive-learning module to distinguish positive (roads) and negative (objects similar to roads) pixels. In addition, to improve the model's learning effectiveness, we propose a model-agnostic transfer learning method, which first builds auxiliary tasks to pre-train the whole model, and then fine-tunes the model's parameters for the main task. Extensive experiments on real datasets verify the superiority of our method as well as the importance of solving the three issues outlined above.
- 2023. BJRoad. https://github.com/suniique/Leveraging-Crowdsourced-GPS-Data-for-Road-Extraction-from-Aeria.Google Scholar
- 2023. Porto. https://github.com/algorithm-panda/.Google Scholar
- Jimmy Lei Ba, Jamie Ryan Kiros, and Geoffrey E Hinton. 2016. Layer normalization. arXiv preprint arXiv:1607.06450 (2016).Google Scholar
- Favyen Bastani, Songtao He, Sofiane Abbar, Mohammad Alizadeh, Hari Balakrishnan, Sanjay Chawla, Sam Madden, and David DeWitt. 2018. Roadtracer: Automatic extraction of road networks from aerial images. In CVPR. 4720--4728.Google Scholar
- Davide Belli and Thomas Kipf. 2019. Image-conditioned graph generation for road network extraction. arXiv preprint arXiv:1910.14388 (2019).Google Scholar
- James Biagioni and Jakob Eriksson. 2012. Map inference in the face of noise and disparity. In SIGSPATIAL GIS. 79--88.Google Scholar
- Zdravko I Botev, Joseph F Grotowski, and Dirk P Kroese. 2010. Kernel density estimation via diffusion. AoS 38, 5 (2010), 2916--2957.Google Scholar
- Alexander Buslaev, Selim Seferbekov, Vladimir Iglovikov, and Alexey Shvets. 2018. Fully convolutional network for automatic road extraction from satellite imagery. In CVPRW. 207--210.Google Scholar
- Abhishek Chaurasia and Eugenio Culurciello. 2017. Linknet: Exploiting encoder representations for efficient semantic segmentation. In VCIP. IEEE, 1--4.Google Scholar
- Chen Chen, Cewu Lu, Qixing Huang, Qiang Yang, Dimitrios Gunopulos, and Leonidas Guibas. 2016. City-scale map creation and updating using GPS collections. In SIGKDD. 1465--1474.Google Scholar
- Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. 2018. Encoder-decoder with atrous separable convolution for semantic image segmentation. In ECCV. 801--818.Google Scholar
- Guangliang Cheng, Ying Wang, Yongchao Gong, Feiyun Zhu, and Chunhong Pan. 2014. Urban road extraction via graph cuts based probability propagation. In ICIP. IEEE, 5072--5076.Google Scholar
- Jonathan J Davies, Alastair R Beresford, and Andy Hopper. 2006. Scalable, distributed, real-time map generation. IEEE PERVAS COMPUT 5, 4 (2006), 47--54.Google ScholarDigital Library
- Vincent Dumoulin and Francesco Visin. 2018. A guide to convolution arithmetic for deep learning. stat 1050 (2018), 11.Google Scholar
- Chelsea Finn, Pieter Abbeel, and Sergey Levine. 2017. Model-agnostic meta-learning for fast adaptation of deep networks. In ICML. PMLR, 1126--1135.Google Scholar
- Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In CVPR. 770--778.Google Scholar
- Songtao He, Favyen Bastani, Sofiane Abbar, Mohammad Alizadeh, Hari Balakrishnan, Sanjay Chawla, and Sam Madden. 2018. RoadRunner: improving the precision of road network inference from GPS trajectories. In SIGSPATIAL GIS. 3--12.Google Scholar
- Songtao He, Favyen Bastani, Satvat Jagwani, Mohammad Alizadeh, Hari Balakrishnan, Sanjay Chawla, Mohamed M Elshrif, Samuel Madden, and Mohammad Amin Sadeghi. 2020. Sat2graph: Road graph extraction through graph-tensor encoding. In ECCV. Springer, 51--67.Google Scholar
- Abdeltawab M Hendawi, Jie Bao, and Mohamed F Mokbel. 2013. iRoad: a framework for scalable predictive query processing on road networks. PVLDB 6, 12 (2013), 1262--1265.Google ScholarDigital Library
- Stefan Hinz and Albert Baumgartner. 2003. Automatic extraction of urban road networks from multi-view aerial imagery. ISPRS 58, 1--2 (2003), 83--98.Google Scholar
- Jilin Hu, Chenjuan Guo, Bin Yang, and Christian S Jensen. 2019. Stochastic weight completion for road networks using graph convolutional networks. In ICDE. IEEE, 1274--1285.Google Scholar
- Hoyoung Jeung, Man Lung Yiu, Xiaofang Zhou, and Christian S Jensen. 2010. Path prediction and predictive range querying in road network databases. The VLDB Journal 19 (2010), 585--602.Google ScholarDigital Library
- Meng Lan, Yipeng Zhang, Lefei Zhang, and Bo Du. 2020. Global context based automatic road segmentation via dilated convolutional neural network. INFORM SCIENCES 535 (2020), 156--171.Google ScholarCross Ref
- Peike Li, Yunqiu Xu, Yunchao Wei, and Yi Yang. 2020. Self-correction for human parsing. IEEE Transactions on Pattern Analysis and Machine Intelligence 44, 6 (2020), 3260--3271.Google ScholarCross Ref
- Justin Liang, Namdar Homayounfar, Wei-Chiu Ma, Shenlong Wang, and Raquel Urtasun. 2019. Convolutional recurrent network for road boundary extraction. In CVPR. 9512--9521.Google Scholar
- Yinyi Lin, Luoma Wan, Hongsheng Zhang, Shan Wei, Peifeng Ma, Yu Li, and Zhuoyi Zhao. 2021. Leveraging optical and SAR data with a UU-Net for large-scale road extraction. INT J APPL EARTH OBS 103 (2021), 102498.Google Scholar
- Hao Liu, Jindong Han, Yanjie Fu, Jingbo Zhou, Xinjiang Lu, and Hui Xiong. 2020. Multi-modal transportation recommendation with unified route representation learning. PVLDB 14, 3 (2020), 342--350.Google Scholar
- Lingbo Liu, Zewei Yang, Guanbin Li, Kuo Wang, Tianshui Chen, and Liang Lin. 2022. Aerial images meet crowdsourced trajectories: a new approach to robust road extraction. TNNLS (2022).Google Scholar
- Ilya Loshchilov and Frank Hutter. 2017. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 (2017).Google Scholar
- Ali Akbar Matkan, Mohammad Hajeb, and Saeed Sadeghian. 2014. Road extraction from lidar data using support vector machine classification. PHOTOGRAMM ENG REM S 80, 5 (2014), 409--422.Google Scholar
- Juan B Mena. 2003. State of the art on automatic road extraction for GIS update: a novel classification. PATTERN RECOGN LETT 24, 16 (2003), 3037--3058.Google ScholarDigital Library
- Zelang Miao, Wenzhong Shi, Alim Samat, Gianni Lisini, and Paolo Gamba. 2015. Information fusion for urban road extraction from VHR optical satellite images. J-STARS 9, 5 (2015), 1817--1829.Google Scholar
- Fausto Milletari, Nassir Navab, and Seyed-Ahmad Ahmadi. 2016. V-net: Fully convolutional neural networks for volumetric medical image segmentation. In 3DV. IEEE, 565--571.Google Scholar
- Kyriakos Mouratidis, Man Lung Yiu, Dimitris Papadias, and Nikos Mamoulis. 2006. Continuous nearest neighbor monitoring in road networks. PVLDB (2006).Google Scholar
- Mashaal Musleh, Sofiane Abbar, Rade Stanojevic, and Mohamed Mokbel. 2021. QARTA: an ML-based system for accurate map services. PVLDB 14, 11 (2021), 2273--2282.Google ScholarDigital Library
- Aaron van den Oord, Yazhe Li, and Oriol Vinyals. 2018. Representation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748 (2018).Google Scholar
- Dian Ouyang, Long Yuan, Lu Qin, Lijun Chang, Ying Zhang, and Xuemin Lin. 2020. Efficient shortest path index maintenance on dynamic road networks with theoretical guarantees. PVLDB 13, 5 (2020), 602--615.Google ScholarDigital Library
- Aditya Prakash, Kashyap Chitta, and Andreas Geiger. 2021. Multi-modal fusion transformer for end-to-end autonomous driving. In CVPR. 7077--7087.Google Scholar
- Olaf Ronneberger, Philipp Fischer, and Thomas Brox. 2015. U-net: Convolutional networks for biomedical image segmentation. In MICCAI. Springer, 234--241.Google Scholar
- Sijie Ruan, Cheng Long, Jie Bao, Chunyang Li, Zisheng Yu, Ruiyuan Li, Yuxuan Liang, Tianfu He, and Yu Zheng. 2020. Learning to generate maps from trajectories. In Proceedings of the AAAI conference on artificial intelligence, Vol. 34. 890--897.Google ScholarCross Ref
- Tao Sun, Zonglin Di, Pengyu Che, Chun Liu, and Yin Wang. 2019. Leveraging crowdsourced GPS data for road extraction from aerial imagery. In CVPR. 7509--7518.Google Scholar
- Yong-Qiang Tan, Shang-Hua Gao, Xuan-Yi Li, Ming-Ming Cheng, and Bo Ren. 2020. Vecroad: Point-based iterative graph exploration for road graphs extraction. In CVPR. 8910--8918.Google Scholar
- Laurens Van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. JMLR 9, 11 (2008).Google Scholar
- Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. NIPS 30 (2017).Google Scholar
- Ting Wang and Ling Liu. 2009. Privacy-aware mobile services over road networks. PVLDB 2, 1 (2009), 1042--1053.Google ScholarDigital Library
- Weixing Wang, Nan Yang, Yi Zhang, Fengping Wang, Ting Cao, and Patrik Eklund. 2016. A review of road extraction from remote sensing images. JTTE 3, 3 (2016), 271--282.Google Scholar
- Wenguan Wang, Tianfei Zhou, Fisher Yu, Jifeng Dai, Ender Konukoglu, and Luc Van Gool. 2021. Exploring cross-image pixel contrast for semantic segmentation. In ICCV. 7303--7313.Google Scholar
- Yong Wang, Guoliang Li, Kaiyu Li, and Haitao Yuan. 2022. A Deep Generative Model for Trajectory Modeling and Utilization. PVLDB 16, 4 (2022), 973--985.Google ScholarDigital Library
- Jan D Wegner, Javier A Montoya-Zegarra, and Konrad Schindler. 2013. A higher-order CRF model for road network extraction. In CVPR. 1698--1705.Google Scholar
- Yanan Wei, Zulin Wang, and Mai Xu. 2017. Road structure refined CNN for road extraction in aerial image. GRSL 14, 5 (2017), 709--713.Google Scholar
- Guojun Wu, Yichen Ding, Yanhua Li, Jie Bao, Yu Zheng, and Jun Luo. 2017. Mining spatio-temporal reachable regions over massive trajectory data. In ICDE. IEEE, 1283--1294.Google Scholar
- Hao Wu, Hanyuan Zhang, Xinyu Zhang, Weiwei Sun, Baihua Zheng, and Yuning Jiang. 2020. DeepDualMapper: A gated fusion network for automatic map extraction using aerial images and trajectories. In AAAI, Vol. 34. 1037--1045.Google ScholarCross Ref
- Wei Yang, Tinghua Ai, and Wei Lu. 2018. A method for extracting road boundary information from crowdsourcing vehicle GPS trajectories. Sensors 18, 4 (2018), 1261.Google ScholarCross Ref
- Xiaojing Yu, Xiang-Yang Li, Jing Zhao, Guobin Shen, Nikolaos M Freris, and Lan Zhang. 2022. ANTIGONE: Accurate Navigation Path Caching in Dynamic Road Networks leveraging Route APIs. In INFOCOM. IEEE, 1599--1608.Google Scholar
- Haitao Yuan and Guoliang Li. 2019. Distributed in-memory trajectory similarity search and join on road network. In ICDE. IEEE, 1262--1273.Google Scholar
- Haitao Yuan and Guoliang Li. 2021. A Survey of Traffic Prediction: from Spatio-Temporal Data to Intelligent Transportation. Data Sci. Eng. 6, 1 (2021), 63--85.Google Scholar
- Haitao Yuan, Guoliang Li, and Zhifeng Bao. 2022. Route Travel Time Estimation on A Road Network Revisited: Heterogeneity, Proximity, Periodicity and Dynamicity. PVLDB 16, 3 (2022), 393--405.Google ScholarDigital Library
- Haitao Yuan, Guoliang Li, Zhifeng Bao, and Ling Feng. 2020. Effective Travel Time Estimation: When Historical Trajectories over Road Networks Matter. In SIGMOD, David Maier, Rachel Pottinger, AnHai Doan, Wang-Chiew Tan, Abdussalam Alawini, and Hung Q. Ngo (Eds.). 2135--2149.Google Scholar
- Haitao Yuan, Guoliang Li, Zhifeng Bao, and Ling Feng. 2021. An Effective Joint Prediction Model for Travel Demands and Traffic Flows. In ICDE. 348--359.Google Scholar
- Haitao Yuan, Guoliang Li, Ling Feng, Ji Sun, and Yue Han. 2020. Automatic View Generation with Deep Learning and Reinforcement Learning. In ICDE. 1501--1512.Google Scholar
- Jiangye Yuan, DeLiang Wang, Bo Wu, Lin Yan, and Rongxing Li. 2011. LEGION-based automatic road extraction from satellite imagery. TGRS 49, 11 (2011), 4528--4538.Google ScholarCross Ref
- Qianqian Zhang, Qingling Kong, Chao Zhang, Shucheng You, Hai Wei, Ruizhi Sun, and Li Li. 2019. A new road extraction method using Sentinel-1 SAR images based on the deep fully convolutional neural network. EuJRS 52, 1 (2019), 572--582.Google ScholarCross Ref
- Zhengxin Zhang, Qingjie Liu, and Yunhong Wang. 2018. Road extraction by deep residual u-net. GRSL 15, 5 (2018), 749--753.Google Scholar
- Lisheng Zhao, Jiali Mao, Min Pu, Guoping Liu, Cheqing Jin, Weining Qian, Aoying Zhou, Xiang Wen, Runbo Hu, and Hua Chai. 2020. Automatic calibration of road intersection topology using trajectories. In ICDE. IEEE, 1633--1644.Google Scholar
- Lichen Zhou, Chuang Zhang, and Ming Wu. 2018. D-LinkNet: LinkNet with pretrained encoder and dilated convolution for high resolution satellite imagery road extraction. In CVPRW. 182--186.Google Scholar
- Zongwei Zhou, Md Mahfuzur Rahman Siddiquee, Nima Tajbakhsh, and Jianming Liang. 2019. Unet++: Redesigning skip connections to exploit multiscale features in image segmentation. IEEE T MED IMAGING 39, 6 (2019), 1856--1867.Google ScholarCross Ref
Recommendations
Automatic and Accurate Extraction of Road Intersections from Raster Maps
Since maps are widely available for many areas around the globe, they provide a valuable resource to help understand other geospatial sources such as to identify roads or to annotate buildings in imagery. To utilize the maps for understanding other ...
Road Extraction from Multi-source Remote Sensing Images Based on 2nd Generation Curvelet Fusion
ICNC '09: Proceedings of the 2009 Fifth International Conference on Natural Computation - Volume 06In view of the problem of non-target interference and misjudgment in road extraction from remote sensing images, a new method of road extraction based on curvelet fusion is presented. Firstly spectral similarity matrix is calculated in multispectral ...
Automatic Road Environment Classification
The ongoing development autonomous vehicles and adaptive vehicle dynamics present in many modern vehicles has generated a need for road environment classification - i.e., the ability to determine the nature of the current road or terrain environment ...
Comments