Abstract
Satellite images find various applications today, of which segmenting the objects in a satellite image for map generation is widely used in disaster mitigation planning and recovery. Existing works mostly segment one class of object from satellite images, whereas actual images contain multiple classes of objects. To overcome this shortcoming, this paper describes a system which generates the base map from a satellite image for four classes of objects - Buildings, Roads, Greenery and Water Bodies. Binary and multi-class U-Net architectures are designed and trained with suitable datasets to segment the four classes of objects individually. The trained models can take an input satellite image (RGB, JPG format) of pixel resolution 30–50 cm, individually segment the four classes and generate the base map by combining segmented regions in different colours. The performance is analysed by comparing the Intersection Over Union (IOU) score and cross entropy loss with validation set and existing models. A real time image acquired from Google Earth Pro is tested and the results are subjectively evaluated to infer that 90% of the regions are segmented correctly.
Source Code - https://github.com/sudhamsugurijala/Sat_Image_Seg.
Supported by Sri Sivasubramaniya Nadar College of Engineering.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Dahiya, S., Garg, P.K., Jat, M.K.: Building extraction from high resolution satellite images using Matlab software. In: 14th International Multidisciplinary Scientific GeoConference SGEM, Albena, Bulgaria, pp. 71–78 (2014)
Xu, Y., Wu, L., **e, Z., Chen, Z.: Building extraction in very high resolution remote sensing imagery using deep learning and guided filters. Remote Sens. 10(1), 144 (2018)
Yüksel, B.: Automated building detection from satellite images by using shadow information as an object invariant. Master’s Thesis, The Graduate School of Natural and Applied Sciences of METU, Turkey (2012)
Aamir, M., Pu, Y.F., Rahman, Z., Tahir, M., Naeem, H., Dai, Q.: A framework for automatic building detection from low-contrast satellite images. Symmetry 11(1), 3 (2019)
Yadav, P. and Agrawal, S.: Road network identification and extraction in satellite imagery using Otsu’s method and connected component analysis. In: International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. XLII-5, pp. 91–98 (2018)
Wang, S., Yang, H., Wu, Q., Zheng, Z., Wu, Y., Li, J.: An improved method for road extraction from high-resolution remote-sensing images that enhances boundary information. Sensors 20(7), 2064 (2020)
Kalyan, M., Rajib, D., Subhasish, D., Anasua, S.: Land use land cover map segmentation using remote sensing: a case study of Ajoy river watershed, India. J. Intell. Syst. 30(1), 273–286 (2021)
Gupta, A., Watson, S., **, H.: Deep learning-based aerial image segmentation with open data for disaster impact assessment. Neurocomputing 439(7), 22–33 (2021)
Abdollahi, A., Pradhan, B., Shukla, N., Chakraborty, S., Alamri, A.: Multi-object segmentation in complex urban scenes from high-resolution remote sensing data. Remote Sens. 13(18), 3710, 1–22 (2021). https://doi.org/10.3390/rs13183710
Wang, F., **e, J.: A context and semantic enhanced UNet for semantic segmentation of high-resolution aerial imagery. J. Phys. Conf. Ser. 1607, 182475–182489 (2020)
Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015, Part III. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-24574-4_28
Rakhlin, A., Davydow, A., Nikolenko, S.: Land cover classification from satellite imagery with U-Net and lovász-softmax loss. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, Salt Lake City, UT, USA, pp. 262–266 (2018)
Pan, X., et al.: Building extraction from high-resolution aerial imagery using a generative adversarial network with spatial and channel attention mechanisms. Remote Sens. 11(8), 917 (2019)
Iglovikov, V., Shvets, A.: Ternausnet: U-Net with VGG11 encoder pre-trained on ImageNet for image segmentation, pp. 1–5. ar**v preprint ar**v:1801.05746 (2018)
Rusiecki, A.: Trimmed categorical cross-entropy for deep learning with label noise. Electron. Lett. 55(6), 319–320 (2019)
Ulmas, P., Liiv, I.: Segmentation of satellite imagery using U-Net models for land cover classification, pp. 1–11. ar**-challenge" data-track="click" data-track-action="external reference" data-track-label="https://www.aicrowd.com/challenges/map**-challenge">https://www.aicrowd.com/challenges/map**-challenge. Accessed November 2020
Roads Dataset Link. https://competitions.codalab.org/competitions/18467. Accessed November 2020
Greenery and Water Bodies Dataset Link. http://landcover.ai. Accessed November 2020
Author information
Authors and Affiliations
Corresponding authors
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 Springer Nature Switzerland AG
About this paper
Cite this paper
Srinivasan, K., Gurijala, S., Sai Chitti Subrahmanyam, V., Swetha, B. (2022). Generating the Base Map of Regions Using an Efficient Object Segmentation Technique in Satellite Images. In: Patel, K.K., Doctor, G., Patel, A., Lingras, P. (eds) Soft Computing and its Engineering Applications. icSoftComp 2021. Communications in Computer and Information Science, vol 1572. Springer, Cham. https://doi.org/10.1007/978-3-031-05767-0_27
Download citation
DOI: https://doi.org/10.1007/978-3-031-05767-0_27
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-05766-3
Online ISBN: 978-3-031-05767-0
eBook Packages: Computer ScienceComputer Science (R0)