Journal of Cadastre & Land InformatiX (지적과 국토정보)

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

DOI QR Code

Detecting high-resolution usage status of individual parcel of land using object detecting deep learning technique

객체 탐지 딥러닝 기법을 활용한 필지별 조사 방안 연구

  • Jeon, Jeong-Bae (Korea Land and Geospatial Infomatix Corporation )
  • 전정배 (한국국토정보공사 공간정보연구원)
  • Received : 2024.04.23
  • Accepted : 2024.06.24
  • Published : 2024.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

This study examined the feasibility of image-based surveys by detecting objects in facilities and agricultural land using the YOLO algorithm based on drone images and comparing them with the land category by law. As a result of detecting objects through the YOLO algorithm, buildings showed a performance of detecting objects corresponding to 96.3% of the buildings provided in the existing digital map. In addition, the YOLO algorithm developed in this study detected 136 additional buildings that were not located in the digital map. Plastic greenhouses detected a total of 297 objects, but the detection rate was low for some plastic greenhouses for fruit trees. Also, agricultural land had the lowest detection rate. This result is because agricultural land has a larger area and irregular shape than buildings, so the accuracy is lower than buildings due to the inconsistency of training data. Therefore, segmentation detection, rather than box-shaped detection, is likely to be more effective for agricultural fields. Comparing the detected objects with the land category by law, it was analyzed that some buildings exist in agricultural and forest areas where it is difficult to locate buildings. It seems that it is necessary to link with administrative information to understand that these buildings are used illegally. Therefore, at the current level, it is possible to objectively determine the existence of buildings in fields where it is difficult to locate buildings.

본 연구에서는 드론영상을 기반으로 YOLO 알고리즘을 통해 시설물과 농경지를 대상으로 객체탐지를 실시하고, 이를 법정지목과 비교를 수행하여 영상기반의 조사 가능성을 검토하였다. YOLO 알고리즘을 통해 객체를 탐지한 결과 건축물의 경우에는 기존 수치지형도에서 제공하고 있는 건축물 중 96.3%에 해당하는 객체를 탐지하는 것으로 분석되었다. 또한 수치지형도에서는 건축물이 위치하지 않지만, 영상에서 건축물이 존재하는 136개의 건축물을 추가로 탐지하는 것으로 나타나 정확도가 높은 것으로 나타났다. 비닐하우스의 경우에는 총 297개를 탐지했으나, 일부 과수형 비닐하우스의 경우에 탐지율이 낮은 것으로 분석되었다. 마지막으로 농경지는 가장 낮은 탐지율을 보였다. 농경지는 시설물 대비 넓은 면적과 불규칙한 형상으로 학습데이터의 일관성이 낮아 정확도가 시설물에 비해 작은 것으로 판단된다. 따라서 농경지의 경우에는 박스형태의 탐지가 아닌 Segmentation 탐지가 더욱 효과적으로 활용될 것으로 보인다. 마지막으로 탐지된 객체를 법정지목과 비교를 수행하였다. 그 결과 건축물이 입지가 어려운 농경지 및 임야에서 건축물이 존재하는 것으로 분석되었다. 그러나 이 건축물이 불법으로 활용됨을 파악하기 위해선 행정정보와 연계가 필요할 것으로 보여진다. 따라서 현재 수준에서는 건축물이 입지하기 어려운 필지에 건축물의 존재유무를 객관적으로 판단할 수 있는 수준까지 조사가 가능한 것으로 볼 수 있다.

Keywords

References

  1. Kim SS, Chung YI, Chung YJ. 2023. Detecting Greenhouses from the Planetscope Satellite Imagery Using the YOLO Algorithm. Journal of the Korean Association of Geographic Information Studies. 26(4): 27-39 
  2. Kim JS, Hong IY. 2021. Analysis of Building Object Detection Based on the YOLO Neural Network Using UAV Images. Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography. 39(6): 391-392 
  3. Kim HY, Na R, Joo DH, Choi GH, Oh YG. 2022. A Study on the Detection of Solar Power Plant for High-Resolution Aerial Imagery Using YOLO v2. Journal of Korean Society of Rural Planning. 28(2): 87-96 
  4. Kim HO, Yeom JM. 2012. A Study on Object-Based Image Analysis Methods for Land Cover Classification in Agricultural Areas. Journal of the Korean Association of Geographic Information Studies. 15(4): 26-41 
  5. Park SC. Park YB, Jang SY, Kim TH. 2022. A Study on Evaluating the Possibility of Monitoring Ships of CAS500-1 Images Based on YOLO Algorithm: A Case Study of a Busan New Port and an Oakland Port in California. Korean Journal of Remote Sensing. 38(6): 1463-1478 
  6. Paek JG. 2006. Study on the Improvement of Land Use Control Functions by Urban Land Management System. Korea Research Institute for Human Settlements Report 
  7. Bae SY, Kim JH. 2015. A Study in Improvement and analysis of Current Situation between Land Category and Building Usage. Journal of the Korean Society of Cadastre. 31(2): 59-74 
  8. Oh CY, Park SY, Kim HS, Lee YW, Choi CU. 2010. Comparison of Landcover Map Accuracy Using High Resolution Satellite Imagery. Journal of the Korean Association of Geographic Information Studies. 13(1): 89-100 
  9. Yoo SJ, Kim J, Noh MJ, Do JE. 2015. A Report on Official Land Value Rising Effects of Land Use Category System. Journal of the Korean Society of Cadastre. 31(2): 87-104
  10. Yim JS, Han WS, Hwang JH, Chung SY, Cho HK, Shin MY. 2009. Estimation of Forest Biomass based upon Satellite Data and National Forest Inventory Data. Korean Journal of Remote Sensing. 25(4): 311-320
  11. Jeon JB, Yoo EJ, Lim GB, 2023. Research on the Application of Local Government Administrative Work using National Survey in Land Use. Spatial Ingormation Research Institute Report
  12. Aszkowski, Przemyslaw, et al. 2023. Deepness: Deep neural remote sensing plugin for QGIS. SoftwareX 23 (2023): 101495.
  13. Badgujar, Chetan M., Alwin Poulose, and Hao Gan. 2024. Agricultural Object Detection with You Look Only Once (YOLO) Algorithm: A Bibliometric and Systematic Literature Review. arXiv preprint arXiv:2401.10379.
  14. Chen, S., Zhan, R., & Zhang, J. 2018. Geospatial object detection in remote sensing imagery based on multiscale single-shot detector with activated semantics. Remote Sensing. 10(6), 820.
  15. Garcia-Ayllon Veintimilla, S., Miralles-Garcia, J. L., & Sowinska-Swierkosz, B. 2023. Challenges in sustainable urban planning and territorial management for the XXI century. Frontiers in Environmental Science. 11, 1252835.
  16. Mani, J. K., & Varghese, A. O. 2018. Remote sensing and GIS in agriculture and forest resource monitoring. Geospatial technologies in land resources mapping, monitoring and management, 377-400.
  17. Parra, L. 2022. Remote Sensing and GIS in Environmental Monitoring. Applied Sciences, 12(16), 8045.
  18. Petso, Tinao, and Rodrigo S. Jamisola. 2023. A Review on Deep Learning on UAV Monitoring Systems for Agricultural Applications. Artificial Intelligence for Robotics and Autonomous Systems Applications . 335-368.
  19. Reddy, G. O. 2018. Geospatial technologies in land resources mapping, monitoring, and management: An overview. Springer International Publishing. p. 1-18.
  20. Redmon, Joseph, et al. 2016. You only look once: Unified, real-time object detection. Proceedings of the IEEE conference on computer vision and pattern recognition.
  21. Skidmore, A. K., Bijker, W., Schmidt, K., & Kumar, L. 1997. Use of remote sensing and GIS for sustainable land management. ITC journal, 3(4), 302-315.
  22. Terven, Juan, Diana-Margarita Cordova-Esparza, and Julio-Alejandro Romero-Gonzalez. 2023. A comprehensive review of yolo architectures in computer vision: From yolov1 to yolov8 and yolo-nas. Machine Learning and Knowledge Extraction 5.4: 1680-1716.
  23. Van Maarseveen, Martin, Javier Martinez, and Johannes Flacke. 2019. GIS in sustainable urban planning and management: a global perspective. Taylor & Francis.
  24. Wu, T., & Dong, Y. 2023. YOLO-SE: Improved YOLOv8 for remote sensing object detection and recognition. Applied Sciences, 13(24), 12977.
  25. Zhang, X., Han, L., Han, L., & Zhu, L. 2020. How well do deep learning-based methods for land cover classification and object detection perform on high resolution remote sensing imagery?. Remote Sensing, 12(3), 417.
  26. Zheng, Z., Lei, L., Sun, H., & Kuang, G. 2020. A review of remote sensing image object detection algorithms based on deep learning. In 2020 IEEE 5th International Conference on Image, Vision and Computing (ICIVC). pp. 34-43. IEEE.