한국지리정보학회지 (Journal of the Korean Association of Geographic Information Studies)

  • 제26권4호
  • /
  • Pages.27-39
  • /
  • 2023
  • /
  • 1226-9719(pISSN)
  • /
  • 2287-6952(eISSN)
한국지리정보학회 (The Korean Association of Geographic Information Studies)
권호 표지
DOI QR코드

DOI QR Code

YOLO 알고리즘을 활용한 Planetscope 위성영상 기반 비닐하우스 탐지

Detecting Greenhouses from the Planetscope Satellite Imagery Using the YOLO Algorithm

  • 김성수 ((주)지오씨엔아이 공간정보기술연구소) ;
  • 정연인 (계명대학교 토목공학과) ;
  • 정윤재 ((주)지오씨엔아이 공간정보기술연구소)
  • Seongsu KIM (Geospatial Research Center, GEO C&I Co., Ltd.) ;
  • Youn-In CHUNG (Dept. of Civil Engineering, Keimyung University) ;
  • Yun-Jae CHOUNG (Geospatial Research Center, Geo C&I Co., Ltd.)
  • 투고 : 2023.09.12
  • 심사 : 2023.10.05
  • 발행 : 2023.12.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

원격탐사 자료 기반 비닐하우스 탐지 기술 개발은 불법 농경 시설물의 현황 파악과 비닐하우스에서 재배되는 농작물 수량 예측을 위해 중요하다. 본 연구에서는 딥러닝 알고리즘을 활용하여 김제시 지역을 촬영한 Planetscope 위성영상들로부터 비닐하우스를 탐지하기 위한 방법을 제안하였다. 우선, 5장의 Planetscope 위성영상을 기반으로 비닐하우스 객체를 포함한 훈련 영상들을 제작하였다. 그리고, 훈련 영상들을 이용하여 YOLO(You Only Look Once) 모델을 학습시킨다. 학습시킨 YOLO 모델을 테스트 Planetscope 위성영상에 적용하여 비닐하우스 객체들을 탐지한다. 본 연구에서 제안한 방법을 적용한 결과, 주어진 Planetscope 위성영상으로부터 총 76.4%의 비닐하우스가 탐지되었다. 추후 연구에서는 공간해상도 1m 이하의 고해상도 위성영상에서 더 많은 비닐하우스 객체를 탐지하기 위한 기술을 개발할 계획이다.

Detecting greenhouses from the remote sensing datasets is useful in identifying the illegal agricultural facilities and predicting the agricultural output of the greenhouses. This research proposed a methodology for automatically detecting greenhouses from a given Planetscope satellite imagery acquired in the areas of Gimje City using the deep learning technique through a series of steps. First, multiple training images with a fixed size that contain the greenhouse features were generated from the five training Planetscope satellite imagery. Next, the YOLO(You Only Look Once) model was trained using the generated training images. Finally, the greenhouse features were detected from the input Planetscope satellite image. Statistical results showed that the 76.4% of the greenhouse features were detected from the input Planetscope satellite imagery by using the trained YOLO model. In future research, the high-resolution satellite imagery with a spatial resolution less than 1m should be used to detect more greenhouse features.

키워드

과제정보

This study was carried out with the support of "Research Program for Agricultural Science & Technology Development (Project No. PJ0162342023)", National Academy of Agricultural Science, Rural Development Administration, Republic of Korea.

참고문헌

  1. Agroberichten Buitenland. Greenhouse in South Korea. 2023. https://www.agroberichtenbuitenland.nl/actueel/nieuws/2021/08/03/greenhouse-in-south-korea(Accessed August 9,
  2. Britannica. greenhouse. 2023. https://www.britannica.com/topic/greenhouse(Accessed August 9,
  3. Choung, Y.J. 2015. A Study on Extraction of Croplands Located nearby Coastal Areas Using High-Resolution Satellite Imagery and LiDAR Data. Journal of the Korean Association of Geographic Information Studies, 18(1):170-181. https://doi.org/10.11108/kagis.2015.18.1.170
  4. Datahunt. YOLO Object Detection, Object Recognition - Concept, Principle, Reasons to Pay Attention, Use Case. https://www.thedatahunt.com/trend-insight/guide-for-yolo-object-detection#yolo-wonri---yolo-algorijeum-jakdong-bangsik-sayong-bangbeop(Accessed August 9,
  5. Ham, G.W., Lee, J.M.., Bae, K.H. and Park, H.G. 2019. A Study on Agricultural Drought Monitoring using Drone Thermal and Hyperspectral Sensor. Journal of the Korean Association of Geographic Information Studies, 22(3):107-119.
  6. Lee, G.S. and Choi, Y.W. 2019. Analysis of Cropland Spectral Properties and Vegetation Index Using UAV. Journal of the Korean Association of Geographic Information Studies, 22(4):86-101.
  7. NIP Group. Why Location and Orientation of Your Greenhouse Matter. https://nipgroup.com/greenhouse-location-orientation/ (Accessed August 9, 2023).
  8. Prospiant. Common Greenhouse Structure. https://www.prospiant.com/complete-guide-to-greenhouse-structures/ (Accessed August 9, 2023).
  9. Redmon, J., Divvala, S., Girshick, R. and Farhadi, A. 2016. You Only Look Once: Unified, Real-Time Object Detection. Proceedings in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 27-30 June 2016, 779-788.
  10. ScienceDirect. Greenhouse Cultivation. 2023. https://www.sciencedirect.com/topics/earth-and-planetary-sciences/greenhouse-cultivation (Accessed August 9, 2023).
  11. SentinelHub. Planetscope. https://docs.sentinel-hub.com/api/latest/data/planet/planet-scope/ (Accessed August 9, 2023).
  12. Sun, H., Wang, W., Lin, R., Zhang, Z. and Zhang, B. 2021. Mapping Plastic Greenhouses with Two-Temporal Sentinel-2 Images and 1D-CNN Deep Learning. Remote Sensing, 13(14):2820.
  13. V7. YOLO: Algorithm for Object Detection Explained [+Examples]. https://www.v7labs.com/blog/yolo-object-detection (Accessed August 9, 2023)
  14. Wu, X., Sahoo, D. and Hoi, S.C.H. 2020. Recent advances in deep learning for object detection. Neurocomputing, 396(5):39-64. https://doi.org/10.1016/j.neucom.2020.01.085
  15. Yoon, B., Seong, S. and Choi, J. 2023a. Detection of Plastic Greenhouses by Using Deep Learning Model for Aerial Orthoimages. Korean Journal of Remote Sensing, 39 (2):183-192. https://doi.org/10.7780/KJRS.2023.39.2.5
  16. Yoon, B.H., Jang, A.., Seong, S.K.. and Choi, J.W. 2023b. Analysis of Agricultural Plastic Greenhouses Detection Results Based on the Seasonal Characteristics of Satellite Imagery. Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography, 41(2):127-134.