DOI QR코드

DOI QR Code

Optimization of Input Features for Vegetation Classification Based on Random Forest and Sentinel-2 Image

랜덤포레스트와 Sentinel-2를 이용한 식생 분류의 입력특성 최적화

  • 이승민 (남서울대학교 공간정보공학과) ;
  • 정종철 (남서울대학교 공간정보공학과)
  • Received : 2020.08.29
  • Accepted : 2020.11.06
  • Published : 2020.12.31

Abstract

Recently, the Arctic has been exposed to snow-covered land due to melting permafrost every year, and the Korea Geographic Information Institute(NGII) provides polar spatial information service by establishing spatial information of the polar region. However, there is a lack of spatial information on vegetation sensitive to climate change. This research used a multi-temporal Sentinel-2 image to perform land cover classification of the Ny-Ålesund in Arctic Svalbard. In the pre-processing step, 10 bands and 6 vegetation spectral index were generated from multi-temporal Sentinel-2 images. In image-classification step is consisted of extracting the vegetation area through 8-class land cover classification and performing the vegetation species classification. The image classification algorithm used Random Forest to evaluate the accuracy and calculate feature importance through Out-Of-Bag(OOB). To identify the advantages of multi- temporary Sentinel-2 for vegetation classification, the overall accuracy was compared according to the number of images stacked and vegetation spectral index. Overall accuracy was 77% when using single-time Sentinel-2 images, but improved to 81% when using multi-time Sentinel-2 images. In addition, the overall accuracy improved to about 83% in learning when the vegetation index was used additionally. The most important spectral variables to distinguish between vegetation classes are located in the Red, Green, and short wave infrared-1(SWIR1). This research can be used as a basic study that optimizes input characteristics in performing the classification of vegetation in the polar regions.

최근 북극은 매년 영구 동토층이 녹아 눈으로 덮인 땅이 드러나고 있어 해당 지역 관리를 위한 공간정보가 필요하다. 한국의 국토지리정보원(NGII)은 극지방의 공간정보를 구축하여 극지공간정보 서비스를 제공하고 있으나, 식생 정보는 제공되지 않고 있으므로 식생 공간정보 구축을 위한 추가적인 연구가 필요하다. 본 연구에서는 북극 스발바르제도의 뉘올레순 지역에 대한 식생 분류를 수행하기 위해 다중 시기의 Sentinel-2 영상을 사용하였다. 전처리 단계에서는 다중 시기 Sentinel-2 영상으로부터 10개 밴드와 6가지 정규 지수식을 생성하였다. 영상 분류는 8개 속성에 대한 토지피복분류를 통해 전체 식생 영역을 추출하는 과정과 전체 식생 영역 내에서 다시 세분류를 수행하는 과정으로 이루어졌다. 영상 분류 알고리즘은 OOB(Out-Of-Bag)를 통해 정확도 평가 및 변수 중요도를 산정할 수 있는 랜덤포레스트를 사용하였다. 전체 정확도는 다시기 영상이 사용되었을 경우와 식생 지수가 추가되었을 경우의 이점을 확인하기 위해 사용된 영상 수에 따라 각각 정확도를 산정하였다. 단일시기의 Sentinel-2 영상은 전체 정확도가 77%였으나, 7개의 다중 시기 Sentinel-2 영상을 기반으로 학습하였을 때, 81%로 향상되었다. 또한, 식생 지수가 추가로 사용된 학습에서 전체 정확도가 약 83%로 향상되었다. 식생 분류 시 변수 중요도는 적색, 녹색, 단파적외선-1 밴드가 가장 높은 변수로 선정되었다. 본 연구는 극지방의 식생에 대한 분류를 수행할 시 입력특성을 최적화하는 기초 연구로 활용될 수 있을 것으로 판단된다.

Keywords

Acknowledgement

본 연구는 2019년 정부(국토교통부)의 재원으로 공간정보 융복합 핵심인재 양성 사업의 지원을 받아 수행된 연구임(2019-02-03)

References

  1. Atkinson, P.M. 2013. Downscaling in remote sensing. International Journal of Applied Earth Observation and Geoinformation. 22, 106-114. https://doi.org/10.1016/j.jag.2012.04.012
  2. Belgiu, M. and L. Dragut. 2016. Random forest in remote sensing: A review of applications and future directions. ISPRS Journal of Photogrammetry and Remote Sensing. 114:24-31. https://doi.org/10.1016/j.isprsjprs.2016.01.011
  3. Breiman, L. 2001. Random forests. Machine Learning 45(1):157-176. https://doi.org/10.1023/A:1010933404324
  4. Chen, Z., J. Pasher, J. Duffe and A. Behnamian. 2017. Mapping Arctic Coastal Ecosystems with High Resolution Optical Satellite Imagery Using a Hybrid Classification Approach. Canadian Journal of Remote Sensing. 43(6):513-527. https://doi.org/10.1080/07038992.2017.1370367
  5. Chi, J., C. Hyun, H. KIM, H. Joo, E. Yang, H. Park and S. Kang. 2017. Development of Web Based GIS for Polar Ocean Research. The Korean Association of Geographic Information Studies. 20(1):15-25.
  6. Epstein, H. Raynolds, M. Walker, D. Bhatt, U. Tucker, C. Pinzon, J. 2012. Dynamics of aboveground phytomass of the circumpolar Arctic tundra during the past three decades. Environmental Research Letters. 7:015506. https://doi.org/10.1088/1748-9326/7/1/015506
  7. ESA Standard Document. 2015. Sentinel 2 User Handbook. Issue 1, Rev 2.24. Available online: https://earth.esa.int/documents/247904/685211/Sentinel-2_User_Handbook.pdf (Accessed May 2, 2020).
  8. Ettehadi, P., S. Kaya, E. Sertel and U. Alganci. 2019. Separating Built-Up Areas from Bare Land in Mediterranean Cities Using Sentinel-2A Imagery. Remote Sensing. 11(3):345. https://doi.org/10.3390/rs11030345
  9. Frampton, W., J. Dash, G. Watmough and E. Milton. 2013. Evaluating the capabilities of Sentinel-2 for quantitative estimation of biophysical variables in vegetation. ISPRS Journal of Photogrammetry and Remote Sensing. 82:(83-92). https://doi.org/10.1016/j.isprsjprs.2013.04.007
  10. Gao, B. C. 1996. NDWI-A normalized difference water index for remote sensing of vegetation liquid water from space. Remote Sensing of Environment, 58(3): 257-266. https://doi.org/10.1016/S0034-4257(96)00067-3
  11. Gitelson, A.A., Y.J. Kaufman and M.N. Merzlyak. 1996. Use of a green channel in remote sensing of global vegetation from EOS-MODIS. Remote Sensing of Environment. 58(3):289-298. https://doi.org/10.1016/S0034-4257(96)00072-7
  12. Hawrylo, P., B. Bartlomiej, P. Wezyk and M. Szostak. 2018. Estimating defoliation of Scots pine stands using machine learning methods and vegetation indices of Sentinel-2. European Journal of Remote Sensing. 51(1):194-204. https://doi.org/10.1080/22797254.2017.1417745
  13. Hoscilo, A. and A. Lewandowska. 2019. Mapping Forest Type and Tree Species on a Regional Scale Using Multi-Temporal Sentinel-2 Data. Remote Sensing. 11(8): 929. https://doi.org/10.3390/rs11080929
  14. Huete, A. A. 1988. soil-adjusted vegetation index (SAVI). Remote Sens. Environ. 25, 295-309. https://doi.org/10.1016/0034-4257(88)90106-X
  15. Immitzer, M., M. Neuwirth, S. Bock, H. Brenner, F. Vuolo and C. Atzberger. 2019. Optimal Input Features for Tree Species Classification in Central Europe Based on Multi-Temporal Sentinel-2 Data. Remote Sensing. 11(22): 2599. https://doi.org/10.3390/rs11222599
  16. Jawak, S. and A. Luis. 2014. A Semiautomatic Extraction of Antarctic Lake Features Using Worldview-2 Imagery. Photogrammetric Engineering and Remote Sensing. 80(10):939-952. https://doi.org/10.14358/PERS.80.10.939
  17. Jeon, H., D. Kim, J. Kim, S. K. D. Vadivel, J. Kim, T. Kim and S. Jeong. 2020. Selection of Optimal Band Combination for Machine Learning-based Water Body Extraction using SAR Satellite Images. The Korean Association of Geographic Information Studies. 23(3):120-131.
  18. Johansen, B.E., H. Tommervik and S.R. Karlsen. 2012. Vegetation mapping of Svalbard utilising Landsat TM/ETM+ data. Polar Record. 48(244):47-63. https://doi.org/10.1017/S0032247411000647
  19. Karlsen, S.R., A. Elvebakk, K.A. Hogda and T. Grydeland. 2014. Spatial and Temporal Variability in the Onset of the Growing Season on Svalbard, Arctic Norway - Measured by MODIS-NDVI Satellite Data. Remote Sensing. 6(9):8088-8106. https://doi.org/10.3390/rs6098088
  20. Kim, H.C. and T.B. Chae. 2019. Status of Korean Research Activity on Arctic Sea Ice Monitoring using KOMPSAT-series Satellite. Journal of Korean Earth Science Society. 40(4):329-339. https://doi.org/10.5467/JKESS.2019.40.4.329
  21. Kim, S.I., H.C. Kim, J.I. Shin and S.G. Hong. 2013. Land-Cover Classification of Barton Peninsular around King Sejong station located in the Antarctic using KOMPSAT-2 Satellite Imagery. Korean Journal of Remote Sensing. Vol.29. No.5:537-544. https://doi.org/10.7780/kjrs.2013.29.5.9
  22. Langford, Z.L., J. Kumar, F.M. Hoffman, A.L. Breen and C.M. Iversen. 2018. Arctic Vegetation Mapping Using Unsupervised Training Datasets and Convolutional Neural Networks. Remote Sensing. 11(1):69. https://doi.org/10.3390/rs11010069
  23. Lebourgeois, V., S. Dupuy, E. Vintrou, M. Ameline, S. Butler and A. Begue. 2017. A combined random forest and obia classification scheme for mapping smallholder agriculture at different nomenclature levels using multisource data (simulated sentinel-2 time series, vhrs and dem). Remote Sensing, 9(3): 259. https://doi.org/10.3390/rs9030259
  24. Lee, J., J. Im, K. Kim and J. Heo. 2015. Change Analysis of Aboveground Forest Carbon Stocks According to the Land Cover Change Using Multi-Temporal Landsat TM Images and Machine Learning Algorithms. The Korean Association of Geographic Information Studies. 18(4):81-99. https://doi.org/10.11108/kagis.2015.18.4.081
  25. Maglione, P., C. Parente and A. Vallario. 2014. Coastline extraction using high resolution WorldView-2 satellite imagery. European Journal of Remote Sensing. 47(1):685-699. https://doi.org/10.5721/EuJRS20144739
  26. National Geographic Information Institute. 2018. 2018 Arctic Spatial Information Establishment Project.
  27. Nguyen, H., T. Doan, E. Tomppo and R. McRoberts. 2020. Land Use/Land Cover Mapping Using Multitemporal Sentinel-2 Imagery and Four Classification Methods -A Case Study from Dak Nong, Vietnam. Remote Sensing. 12(9):1367. https://doi.org/10.3390/rs12091367
  28. Parnell, L.D., P. Lindenbaum, S. Khader, G. Dall'Olio, M. Swan, L. Jensen, S. Cockell, B. Pedersen, M. Mangan, C. Miller and I. Albert. 2011. BioStar: An Online Question & Answer Resource for the Bioinformatics Community, PLoS Computational Biology. 7(10), pp.e1002216. doi: 10.1371/journal.pcbi.1002216.
  29. Salvatori, R. R. Casacchia and M. Valt. 2005. Snow surface classification in the Western Svalbard Island. 31th International Symposium on Remote Sensing of Environment: Global Monitoring for Sustainability and Security. Saint Petersburg, Russia, Jun. 20-Jun. 24, 2005.
  30. Somvanshi, S. and M. Kumari. 2020. Comparative analysis of different vegetation indices with respect to atmospheric particulate pollution using Sentinel data. Applied Computing and Geosciences. 7:100032. https://doi.org/10.1016/j.acags.2020.100032
  31. The Local. 2017. Norway to boost climate change defences of 'doomsday' seed vault. May 15, https://www.thelocal.no/20170521/norway-to-boost-climate-change-defences-of-doomsday-seed-vault.(Accessed December 20, 2019).
  32. Tucker, C.J. 1979. Red and Photographic Infrared Linear Combinations for Monitoring Vegetation. Remote Sensing of Environment. 8(2):127-150. https://doi.org/10.1016/0034-4257(79)90013-0
  33. Vuolo, F., M. Neuwirth, M. Immitzer, C. Atzberger and W.T. Ng. 2018. How much does multi-temporal Sentinel-2 data improve crop type classification?. International Journal of Applied Earth Observation and Geoinformation. 72:122-130. https://doi.org/10.1016/j.jag.2018.06.007
  34. Zheng, H., P. Du, J. Chen, J. Xia, E. Li, Z. Xu, X. Li and N. Yokoya. 2017. Performance Evaluation of Downscaling Sentinel-2 Imagery for Land Use and Land Cover Classification by Spectral-Spatial Features. Remote Sensing. 9, 1274. https://doi.org/10.3390/rs9121274