Journal of Cadastre & Land InformatiX (지적과 국토정보)

LX Spatial Information Research Institute (한국국토정보공사 공간정보연구원)
권호 표지
DOI QR코드

DOI QR Code

Object Classification and Change Detection in Point Clouds Using Deep Learning

포인트 클라우드에서 딥러닝을 이용한 객체 분류 및 변화 탐지

  • Seo, Hong-Deok (Department of Spatial Information Engineering, Namseoul University) ;
  • Kim, Eui-Myoung (Department of Spatial Information Engineering, Namseoul University)
  • 서홍덕 (남서울대학교 공간정보공학과) ;
  • 김의명 (남서울대학교 공간정보공학과)
  • Received : 2020.09.24
  • Accepted : 2020.11.25
  • Published : 2020.12.30
Download PDF
⟨ Previous Next ⟩

Abstract

With the development of machine learning and deep learning technologies, there has been increasing interest and attempt to apply these technologies to the detection of urban changes. However, the traditional methods of detecting changes and constructing spatial information are still often performed manually by humans, which is costly and time-consuming. Besides, a large number of people are needed to efficiently detect changes in buildings in urban areas. Therefore, in this study, a methodology that can detect changes by classifying road, building, and vegetation objects that are highly utilized in the geospatial information field was proposed by applying deep learning technology to point clouds. As a result of the experiment, roads, buildings, and vegetation were classified with an accuracy of 92% or more, and attributes information of the objects could be automatically constructed through this. In addition, if time-series data is constructed, it is thought that changes can be detected and attributes of existing digital maps can be inspected through the proposed methodology.

머신러닝과 딥러닝 기술의 발달로 인하여 도시의 변화탐지에 이러한 기술을 적용하려는 관심과 시도가 증가하고 있다. 그러나 기존의 변화탐지와 공간정보 구축방법은 여전히 사람에 의해 수작업으로 수행되는 경우가 많아 비용과 시간이 많이 소요되고 있다. 또한 도시지역에서 건축물의 변화탐지를 효율적으로 수행하기 위해서는 많은 인원이 필요한 실정이다. 따라서, 본 연구에서는 포인트 클라우드에서 딥러닝 기술을 적용하여 공간정보 분야에서 활용도가 높은 도로, 건물, 식생의 객체를 분류하고 변화탐지를 수행할 수 있는 방법을 제안하였다. 실험 결과 약 92% 이상의 정확도로 도로, 건물, 식생을 분류하였으며 이를 통해 객체의 속성정보를 자동으로 구축할 수 있었다. 또한, 시계열 데이터가 구축된다면 제안한 방법론을 통해서 변화를 탐지할 수 있고 기 구축된 수치지도의 속성을 검수할 수 있을 것으로 판단된다.

Keywords

References

  1. Ministry of Land, Infrastructure and Transport. 2017. Strategies for the development of spatial information-based convergence industries and measures to improve the legal system. Spatial Information Industry Promotion Institute. 1-179.
  2. Kim HJ, Lee JM, Bae KH, Eo YD. 2018. Application Research on Obstruction Area Detection of Building Wall using R-CNN Technique. Journal of Cadastre & Land InformatiX. 48(2): 213-225. https://doi.org/10.22640/LXSIRI.2018.48.2.213
  3. Ryu SH, Yoon JB. 2017. The Effect of regularization and identity mapping on the performance of activation functions. Journal of the Korea Academia-Industrial cooperation Society. 18(10):75-80. https://doi.org/10.5762/KAIS.2017.18.10.75
  4. Park SH, Choi MH, Oh BW. 2016. A Case Study of the Transformation Process of National Spatial Information Policy. Journal of the Korean Cadastre Information Association. 18(1):99-112.
  5. Seo KH. 2018. GeoAI-based Land Use Change Monitoring Innovation and Utilization Plan. Korea Research Institute For Human Settlements. KRIHS POLICY BRIEF. 694:1-6.
  6. Seo HD, Kim EM. 2019. Object Classification Using Point Cloud and True Ortho-image by Applying Random Forest and Support Vector Machine Techniques. Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography. 37(6):405-416. https://doi.org/10.7848/KSGPC.2019.37.6.405
  7. Seo HD, Kim EM. 2020. 3D Coordinate Transformation Based on Point-to-Line Using Quaternion and Procrustes Algorithms. Journal of Institute of Control, Robotics and Systems. 26(10): 863-870. https://doi.org/10.5302/J.ICROS.2020.20.0051
  8. Yang SC. 2019. A Study on Improving Availability of Open Data by Location Intelligence. Journal of Cadastre & Land InformatiX. 49(1):93-107.
  9. Oh YK. 2018. Spatio-temporal Analysis of Forest Change using Spatial Information : A case study of Heongseong and Wonju. Journal of Cadastre & Land InformatiX. 48(2):21-30. https://doi.org/10.22640/LXSIRI.2018.48.2.21
  10. Lee DK, Cho EJ, Lee DC. 2018. Evaluation of Building Detection from Aerial Images Using Region-based Convolutional Neural Network for Deep Learning. Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography. 36(6):469-481. https://doi.org/10.7848/KSGPC.2018.36.6.469
  11. Jo WH, Lim YH, Park KH. 2019. Deep learning based Land Cover Classification Using Convolutional Neural Network - a case study of Korea -. Journal of the Korean Geographical Society. 54(1):1-16.
  12. Chai BS, Mun JK, Lee MH. 2015. Illegal Act and Management System within Greenbelt and Improvement of Policy. The Korea Spatial Planning Review. 84:145-163. https://doi.org/10.15793/kspr.2015.84..009
  13. Cheonan-Si. 2020. Cheonan-Si statistics [Internet]. [https://www.cheonan.go.kr/stat.do]. Last accessed 4 September 2020.
  14. Han SH. 2016. Introduction to Photogrammetry and Remote Sensing. Goomibook. p. 348-350.
  15. Hong SP, Kim EM. 2018. Classification of 3D Road Objects Using Machine Learning. Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography. 36(6):535-544. https://doi.org/10.7848/KSGPC.2018.36.6.535
  16. Agarap AF. 2018. Deep learning using rectified linear units (relu). arXiv preprint. arXiv:1803.08375.
  17. Ertugrul OF. 2018. A novel type of activation function in artificial neural networks: Trained activation function. Neural Networks. 99:148-157. https://doi.org/10.1016/j.neunet.2018.01.007
  18. Jaderberg M, Simonyan K, Zisserman A. 2015. Spatial transformer networks. In Advances in Neural Information Processing Systems, 7-12 December 2015, Montreal, Canada. 2017-2025.
  19. Liu Z, Tang H, Lin Y, Han S. 2019. Point-voxel CNN for efficient 3D deep learning. In Advances in Neural Information Processing Systems, 8-14 December 2019, Vancouber, Canada. 965-975.
  20. Qi CR, Su H, Mo K, Guibas, LJ. 2017. Pointnet: Deep learning on point sets for 3d classification and segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, 21-26 July 2017, Honolulu, USA. 652-660.
  21. Choy CB, Xu D, Gwak J, Chen K, Savarese S. 2016. 3d-r2n2: A unified approach for single and multi-view 3d object reconstruction. In European conference on computer vision, Springer, Cham, 8-16 October 2016, Amsterdam, Netherlands. 628-644.
  22. Riegler G, Osman Ulusoy A, Geiger A. 2017. Octnet: Learning deep 3d representations at high resolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, CVPR, 21-26 July 2017, Honolulu, USA, 3577-3586.
  23. Fusiello A, Crosilla F, Malapelle F. 2015. Procrustean point-line registration and the NPnP problem, In 2015 International Conference on 3D Vision, 19-22 October 2015, Lyon, France, 250-255.
  24. Habib A, Mazaheri M. 2015. Quaternion-based solutions for the single photo resection problem, Photogrammetric Engineering & Remote Sensing. 81(3):209-217. https://doi.org/10.14358/PERS.81.3.209-217
  25. Vignesh T, Thyagharajan, KK, Ramya K. 2019. Change detection using deep learning and machine learning techniques for multispectral satellite images, International Journal of Innovative Technology and Exploring Engineering. 9(1S):90-93. https://doi.org/10.35940/ijitee.A1021.1191S19
  26. Wikipedia. 2004. Support vector machine; [accessed: 2020 September 3]. https://en.wikipedia.org/wiki/Support_vector_machine.
  27. Zhang W, Qi J, Wan P, Wang H, Xie D, Wang X, Yan G. 2016. An easy-to-use airborne LiDAR data filtering method based on cloth simulation, ISPRS Annals of the Photogrammetry, Remote Sensing, 8(6): 501.

Cited by

  1. 기구축 공간데이터를 활용한 3차원 건물모델의 위치정합 기법 연구 - ICP 알고리즘 구현 중심으로 - vol.51, pp.1, 2021, https://doi.org/10.22640/lxsiri.2021.51.1.67
  2. 항공영상을 이용한 딥러닝 기반 건물객체 추출 기법들의 비교평가 vol.39, pp.3, 2020, https://doi.org/10.7848/ksgpc.2021.39.3.157