한국환경생태학회지 (Korean Journal of Environment and Ecology)

  • 제32권6호
  • /
  • Pages.686-697
  • /
  • 2018
  • /
  • 1229-3857(pISSN)
  • /
  • 2288-131X(eISSN)
한국환경생태학회 (Korean Society of Environment and Ecology)
권호 표지
DOI QR코드

DOI QR Code

KOMPSAT-3A 위성영상과 토지피복도를 활용한 산림식생의 임상 분류법 개발

Development of a Classification Method for Forest Vegetation on the Stand Level, Using KOMPSAT-3A Imagery and Land Coverage Map

  • 송지용 (서울대학교 협동과정 조경학과) ;
  • 정종철 (남서울대학교 공간정보공학과) ;
  • 이상훈 (한양대학교 도시설계.경관생태조경학과)
  • Song, Ji-Yong (Dept. of Interdisciplinary Program in Landscape Architecture, Seoul National Univ.) ;
  • Jeong, Jong-Chul (Dept. of Geographic Information Science, Namseoul Univ.) ;
  • Lee, Peter Sang-Hoon (Dept. of Urban Design and Landscape Ecological Architecture, Hanyang Univ.)
  • 투고 : 2018.11.26
  • 심사 : 2018.12.14
  • 발행 : 2018.12.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

오늘날 원격탐지기술의 발달로 인해, 산림지역과 같이 피복 분류작업이 난해한 지역을 비롯한 광범위한 지역에서의 세밀한 변화탐지를 위한 고해상도 위성영상 취득이 가능해졌다. 하지만, 고해상도 영상에 대한 시계열분석의 과정에서 많은 양의 지상 관측 데이터가 요구된다. 본 연구에서는 토지피복도를 지상 관측데이터로 활용한 위성영상 분류 방법의 가능성을 시험하였다. 연구대상지는 강원도 원주시이며, 산림지역과 시가화지역이 공존하는 공간이다. 연구 자료는 2015년 3월에 촬영된 KOMPSAT-3A 영상과 2017년도 토지피복도를 이용하여 분류를 시도하였다. 서포트벡터머신(SVM)과 랜덤포레스트(RF)의 두 가지 상이한 화소기반 분류기법을 적용하여 대상지에 대한 피복분류의 분류정확도를 비교 분석하였으며, SVM 분석의 경우 다수 분석(Majority analysis)을 후속 진행하였다. 분석대상은 산림식생만 포함한 지역과 연구대상지 전지역으로 구분하였고, 대상 면적이 협소한 습지는 분석과정에서 제외하였다. 분류 결과는 오차 행렬의 전체 정확도가 두 가지 분류대상에 대해 RF 기법이 SVM 기법보다 더 나은 것으로 나타났다. 산림지역만을 대상으로 한 경우, RF 기법이 SVM 기법에 비해 18.3% 높은 값을 나타낸 반면, 전체지역을 대상으로 한 경우는 둘 사이의 간격이 5.5%로 줄어들었다. SVM 기법에 다수 분석 (Majority analysis)을 추가로 실시한 경우, 1% 정도의 정확도 향상이 나타났다. RF 기법은 산림지역의 활엽수를 분석해 내는데 상당히 효과적이었지만, 다른 대상에 대해서는 SVM 기법이 더 나은 결과를 나타내었다. 본 연구는 고해상도 단일시기 영상에 대한 화소 기반의 분류기법을 시험한 것으로, 추후 시계열분석 및 객체기반 분류기법의 추가적인 적용으로 향상된 정확도와 신뢰도를 얻을 수 있을 것으로 판단된다. 이 연구의 방법론은 시공간적으로 고해상도 분석결과를 제공함으로써, 대면적의 토지계획에 유용할 것으로 기대된다.

Due to the advance in remote sensing technology, it has become easier to more frequently obtain high resolution imagery to detect delicate changes in an extensive area, particularly including forest which is not readily sub-classified. Time-series analysis on high resolution images requires to collect extensive amount of ground truth data. In this study, the potential of land coverage mapas ground truth data was tested in classifying high-resolution imagery. The study site was Wonju-si at Gangwon-do, South Korea, having a mix of urban and natural areas. KOMPSAT-3A imagery taken on March 2015 and land coverage map published in 2017 were used as source data. Two pixel-based classification algorithms, Support Vector Machine (SVM) and Random Forest (RF), were selected for the analysis. Forest only classification was compared with that of the whole study area except wetland. Confusion matrixes from the classification presented that overall accuracies for both the targets were higher in RF algorithm than in SVM. While the overall accuracy in the forest only analysis by RF algorithm was higher by 18.3% than SVM, in the case of the whole region analysis, the difference was relatively smaller by 5.5%. For the SVM algorithm, adding the Majority analysis process indicated a marginal improvement of about 1% than the normal SVM analysis. It was found that the RF algorithm was more effective to identify the broad-leaved forest within the forest, but for the other classes the SVM algorithm was more effective. As the two pixel-based classification algorithms were tested here, it is expected that future classification will improve the overall accuracy and the reliability by introducing a time-series analysis and an object-based algorithm. It is considered that this approach will contribute to improving a large-scale land planning by providing an effective land classification method on higher spatial and temporal scales.

키워드

참고문헌

  1. Breiman, L.(1999) Random forests-random features, Technical Report 567, Statistics Department, University of California, Berkeley.
  2. Chapelle, O., P. Haffner and V.N. Vapnik(1999) Support vector machines for histogram-based image classification. IEEE Transactions on Neural Networks 10(5): 1055-1064. https://doi.org/10.1109/72.788646
  3. Choi, J.Y., S.H. Lee, S.A. Lee, S.Y. Ji and P.S.H. Lee(2015) Applicability of supervised classification for subdividing forested areas using SPOT-5 and KOMPSAT-2 data. The Korea Society of Environmental Restoration Technology 18(2): 89-104. (in Korean with English abstract) https://doi.org/10.13087/kosert.2015.18.2.89
  4. Ghamisi, P., J.A. Benediktsson and S. Phinn(2015) Land-cover classification using both hyperspectral and LiDAR data. International Journal of Image and Data Fusion 6: 189-215. https://doi.org/10.1080/19479832.2015.1055833
  5. Hayes, M., S.N. Miller and M.A. Murphy(2014) High-resolution landcover classification using Random Forest. Remote Sensing Letters 5: 112-121. https://doi.org/10.1080/2150704X.2014.882526
  6. Jung, M., S.H. Lee, E. Chang and S. Hong(2012) Method of Monitoring Forest Vegetation Change based on Change of MODIS NDVI Time Series Pattern. Journal of Korea Spatial Information Society 20(4): 47-55. (in Korean with English abstract)
  7. Kang, N.Y., S.Y. Go and G. S. Cho(2013) A comparative study on suitable SVM kernel function of land cover classification using KOMPSAT-2 imagery. The Korea Society for Geospatial Information System 21(2): 19-25. (in Korean with English abstract)
  8. Kim, H.O., J.M. Yeom and Y. S. Kim(2011) The multi-temporal characteristics of spectral vegetation indices for agricultural land use on RapidEye satellite imagery. Korea Aerospace Research Institute 10(1): 149-155. (in Korean with English abstract)
  9. Kim, S.J. and Y.G. Lee(2018) Present Status and Future Prospect of Satellite Image Uses in Water Resources Area. Korean Journal of Ecology and Environment 51(1): 105-123. (in Korean with English abstract) https://doi.org/10.11614/KSL.2018.51.1.105
  10. Kim, T.H., K.H. Kim, G.B. Nam, J.H. Shim, W.J. Choi and M. H. Cho(2010) Development of Natural Disaster Damage Investigation System using High Resolution Spatial Images. Korea Spatial Information Society 12(1): 57-65. (in Korean with English abstract)
  11. Ku, C.Y.(2011) A Study on Managing and processing the Images with Different Spatial Resolution for the Systemic Land Cover Classification. The Korean Association of Professional Geographers 45(3): 375-386. (in Korean with English abstract)
  12. Kulkarni, A.D. and B. Lowe(2016) Random Forest Algorithm for Land Cover Classification. International Journal on Recent and Innovation Trends. In Computing and Communication 4(3): 58-63.
  13. Landis, J.R. and G.G. Koch(1977) The measurement of observer agreement for categorical data. Biometrics 33: 159-174. https://doi.org/10.2307/2529310
  14. Lee, . G., J.H. You,Y.G. Yu and H. J. Lee(2018) Establish of Parallel Object Based Classification Process and Extraction of Spatial Objects Using KOMPSAT-3A Satellite Image, Proceeding of 2018 Symposium of the Korean Society for Geospatial Information Science, pp. 93-94. (in Korean with English abstract)
  15. Lee, J., J.H. Ru and Y.G. Yu(2010) Extracting High Quality Thematic Information by Using High-Resolution Satellite Imagery. The Korea Society for Geospatial Information System 18(1): 73-81. (in Korean with English abstract)
  16. Lee, S.H.(2009) Adaptive Reconstruction of NDVI Image Time Series for Monitoring Vegetation Changes. The Korean Society of Remote Sensing 25(2): 95-105. (in Korean with English abstract)
  17. Li, M., L. Ma, T. Blaschke, L. Cheng and D. Tiede(2016) A systematic comparison of different object-based classification techniques using high spatial resolution imagery in agricultural environments. International Journal of Applied Earth Observation and Geoinformation 49: 87-98. https://doi.org/10.1016/j.jag.2016.01.011
  18. Lim, Y.S., Y.D. Eo and M.W. Pyeon(2016) Experiments of Individual Tree and Crown Width Extraction by Band Combination Using Monthly Drone Images. The Korean Society for Geospatial Information Science 24(4): 67-74. (in Korean with English abstract)
  19. Mittal, V., D. Singh and L.M. Saini(2015) A critical analysis of EM based fusion of different polarization data for effect on land cover classification. Advances in Space Research 56(6): 1094-1105. https://doi.org/10.1016/j.asr.2015.06.004
  20. Myint, S.W., P. Gober, A. Brazel, S. Grossman-Clarke and Q. Weng(2011) Per-pixel vs. object-based classification of urban land cover extraction using high spatial resolution imagery. Remote Sensing of Environment 115(5): 1145-1161. https://doi.org/10.1016/j.rse.2010.12.017
  21. Oh, C.Y., S.Y. Park, H.S. Kim, Y.W. Lee and C. U. Choi(2010) Comparison of Landcover Map Accuracy Using High Resolution Satellite Imagery. The Korean Association of Geographic Information Studies 13(1): 89-100. (in Korean with English abstract)
  22. Ortiz, M.J., A.R. Formaggio and J.C.N. Epiphanio(1997) Classification of croplands through integration of remote sensing GIS and historical database. International Journal of Remote Sensing 18(1): 95-105. https://doi.org/10.1080/014311697219295
  23. Park, J.G., S.Y. Go and G.S. Cho(2013) Vegetation Classification using KOMPSAT-2 Imagery and High-resolution airborne imagery in Urban Area. Journal of the Korean Society for Geospatial Information Science 21(4): 21-27. (in Korean with English abstract) https://doi.org/10.12672/ksis.2013.21.3.021
  24. Park, J.J., S.W. Oh, K.A. Park, M.S. Lee, J.C. Jang and M.J. Lee(2018) A methodology of ship detection using high-resolution satellite optical image. Journal of the Korean Earth Science Society 39(3): 241-249. (in Korean with English abstract) https://doi.org/10.5467/JKESS.2018.39.3.241
  25. Scholkopf, B., S. Mika, C.J.C. Burges, P. Knirsch, K.-R. Muller, G. Ratsch and A. J. Smola(1999) Input space versus feature space in Kernel-based methods. IEEE Transactions on Neural Networks 10(55): 1000-1016. https://doi.org/10.1109/72.788641
  26. Sim, W.S., S.Y. Sung and C. K. Cheng(2009) Performance Analysis of Support Vector Machine Using Kernel Function. Proceeding of 2009 Symposium of the Institute of Electronics Engineers of Korea, pp. 796-797. (in Korean with English abstract)
  27. Tomppo, E. and M. Katila(1991) Satellite image-based national forest inventory of Finland, Proceeding of IGARSS 1991 Symposium of Geoscience and Remote Sensing Symposium - Remote Sensing: 3rd Global Monitoring for Earth Management, pp. 1141-1144.
  28. Voigt, S., T. Kemper, T. Riedlinger, R. Keifl, K. Scholte and H. Mehl(2007) Satellite image analysis for disaster and crisis-management support. IEEE Transactions on Geoscience and Remote Sensing 45(6): 1520-1528. https://doi.org/10.1109/TGRS.2007.895830
  29. Wu, T.F., C.J. Lin and R.C. Weng(2004) Probability estimates for multi-class classification by pairwise coupling. Journal of Machine Learning Research 5: 975-1005.
  30. Yong, S.S., G.S. Kang and H.P. Heo(2016) Current Status and Future Prospects of Satellite Payloads Technology. Korean Society for Aeronautical and Space Sciences 44(8): 710-717. (in Korean with English abstract) https://doi.org/10.5139/JKSAS.2016.44.8.710
  31. Zhu, G. and D. G. Blumberg(2002) Classification using ASTER data and SVM algorithms; The case study of Beer Sheva, Israel. Remote Sensing of Environment 80: 233-240. https://doi.org/10.1016/S0034-4257(01)00305-4