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

The Korean Association of Geographic Information Studies (한국지리정보학회)
권호 표지
DOI QR코드

DOI QR Code

Land Cover Classification Based on High Resolution KOMPSAT-3 Satellite Imagery Using Deep Neural Network Model

심층신경망 모델을 이용한 고해상도 KOMPSAT-3 위성영상 기반 토지피복분류

  • 문갑수 ((주)지오씨엔아이 공간정보기술연구소) ;
  • 김경섭 ((주)지오씨엔아이 공간정보기술연구소) ;
  • 정윤재 ((주)지오씨엔아이 공간정보기술연구소)
  • Received : 2020.09.04
  • Accepted : 2020.09.18
  • Published : 2020.09.30
Download PDF
⟨ Previous Next ⟩

Abstract

In Remote Sensing, a machine learning based SVM model is typically utilized for land cover classification. And study using neural network models is also being carried out continuously. But study using high-resolution imagery of KOMPSAT is insufficient. Therefore, the purpose of this study is to assess the accuracy of land cover classification by neural network models using high-resolution KOMPSAT-3 satellite imagery. After acquiring satellite imagery of coastal areas near Gyeongju City, training data were produced. And land cover was classified with the SVM, ANN and DNN models for the three items of water, vegetation and land. Then, the accuracy of the classification results was quantitatively assessed through error matrix: the result using DNN model showed the best with 92.0% accuracy. It is necessary to supplement the training data through future multi-temporal satellite imagery, and to carry out classifications for various items.

원격탐사 분야에서 토지피복분류에는 머신러닝 기반의 SVM 모델이 대표적으로 활용되고 있는 한편, 신경망 모델을 이용한 연구도 지속적으로 수행되고 있다. 다목적실용위성의 고해상도 영상을 이용한 연구는 미흡한 실정이며, 따라서 본 연구에서는 고해상도 KOMPSAT-3 위성영상을 이용하여 신경망 모델의 토지피복분류 정확도를 평가하고자 하였다. 경주시 인근 해안지역의 위성영상을 취득하여 훈련자료를 제작하고, 물과 식생 및 육지의 세 항목에 대해 SVM, ANN 및 DNN 모델로 토지피복을 분류하였다. 분류 결과의 정확도를 오차 행렬을 통해 정량적으로 평가한 결과 DNN 모델을 활용한 토지피복분류가 92.0%의 정확도로 가장 우수한 결과를 나타냈다. 향후 다중 시기의 위성영상을 통해 훈련자료를 보완하고, 다양한 항목에 대한 분류를 수행 및 검증한다면 연구의 신뢰성을 높일 수 있을 것으로 판단된다.

Keywords

References

  1. Chakraborty, T., M.S. Alam and M.D. Islam. 2019. Land cover classification from multispectral remote sensing image using k-nearest neighbor, deep neural network, decision tree and SVM Algorithms: a machine learning-based comparison approach. Conference paper of 1st International Conference on Urban and Regional Planning 2019. Dhaka, Bangladesh, Oct. 5, 2019. pp.1-10.
  2. Elhag, M. and S. Boteva. 2016. Mediterranean land use and land cover classification assessment using high spatial resolution data. World Multidisciplinary Earth Sciences Symposium, IOP Conference Series: Earth and Environmental Science. Prague, Czech Republic, Sep. 5-9, 2016. Vol. 44(4), 042032.
  3. Heo, S.Y., J.Y. Kim and T.H. Moon. 2018. Predicting crime risky area using machine learning. Journal of the Korean Association of Geographic Information Studies 21(4):64-80. https://doi.org/10.11108/KAGIS.2018.21.4.064
  4. Jeon, H.S. and G.S. Cho. 2000. A comparison of neural networks and maximum likelihood classifier for the classification of land-cover. Journal of Korean Society for Geospatial Information System 8(2):23-33.
  5. Jo, W.H., Y.H. Lim and K.H. Park. 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.
  6. Kang, M.S., S.W. Park and K.S. Yoon. 2006. Land cover classification of image data using artificial neural networks. Journal of Korean Society of Rural Planning 12(1):75-83.
  7. Kang, N.Y., J.G. Pak, G.S. Cho and Y. Yeu. 2012. An analysis of land cover classification methods using IKONOS satellite image. Journal of Korean Society for Geospatial Information System 20(3):65-71. https://doi.org/10.7319/kogsis.2012.20.3.065
  8. Kim, D.H., K.Y. Hwang and Y. Yoon. 2019. Prediction of traffic congestion in seoul by deep neural network. The Journal of The Korea Institute of Intelligent Transport Systems 18(4):44-57.
  9. Kim, H.O. and J.M. Yeom. 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. https://doi.org/10.11108/kagis.2012.15.4.026
  10. Kim, Y.K. 2020. A study on urban land cover classification using object-based image analysis(OBIA) techniques. Journal of The Korean Cadastre Information Association 22(1):122-144. https://doi.org/10.46416/JKCIA.2020.04.22.1.122
  11. Ko, S.Y., G. Jun and J.H. Won. 2016. HyperConv: spatio-spectral classication of hyperspectral images with deep convolutional neural networks. The Korean Journal of Applied Statistics 29(5):859-872. https://doi.org/10.5351/KJAS.2016.29.5.859
  12. Ku. C.Y. 2007. Generation of the land cover map with high resolution satellite image. The Geographical Journal of Korea 41(1):83-94.
  13. Lee, H.J., J.H. Ru and Y.G. Yu. 2010. Extracting high quality thematic information by using high-resolution satellite imagery. Journal of Korean Society for Geospatial Information Science 18(1):73-81.
  14. Lee, M.B., N.S. Kim, G.R. Lee and U. Han. 2005. A study on the environmental changes of coastal area in oncheon gun of pyeongnam province by neural network classification using satellite images, West Coast of North Korea. Journal of the Korean association of regional geographers 11(2):278-290.
  15. Lee, M.J., S.R. Lee and S.W. Jeon. 2012. Landslide hazard mapping and verification using probability rainfall and artificial neural networks. Journal of the Korean Association of Geographic Information Studies 15(2):57-70. https://doi.org/10.11108/kagis.2012.15.2.057
  16. Lim, H.Y., H.S. Kim, J.S. Choi and S.H. Song. 2005. A study on efficient topography classification of high resolution satelite image. Journal of Korean Society for Geospatial Information System 13(3):33-40.
  17. Moon, T.H. 2000. Integrated simulation system of GIS and ANN for land price appraisal. Journal of the Korean Association of Geographic Information Studies 3(2):1-10.
  18. Na, H.S. and J.S. Lee. 2014. Analysis of land cover characteristics with objectbased classfication method -focusing on the DMZ in Inje-gun, Gangwon-do-. Journal of the Korean Association of Geographic Information Studies 17(2):121-135. https://doi.org/10.11108/kagis.2014.17.2.121
  19. Song, A.R. and Y.G. Kim. 2017. Deep learning-based hyperspectral image classification with application to environmental geographic information systems. Korean Journal of Remote Sensing 33(6):1061-1073.
  20. Yang, I.T., S.M. Han and J.K. Park. 2002. An application of artificial intelligence system for accuracy improvement in classification of remotely sensed images. Journal of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography 20(1):21-31.
  21. Yoon, Y.S., Y.S. Kim and Y.S. Kim. 2007. A study on the object-based classification method for wildfire fuel type map. Aerospace Engineering and Technology 6(1):213-221.
  22. Xu, C.X., J.H. Lim, X.M. Jin and H.C. Yun. 2018. Land cover mapping and availability evaluation based on drone images with multi-spectral camera. Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography 36(6):589-599. https://doi.org/10.7848/KSGPC.2018.36.6.589

Cited by

  1. Landsat-8 위성영상 기반 수분지수 및 기계학습을 활용한 대구광역시의 지표수 탐지 vol.24, pp.1, 2021, https://doi.org/10.11108/kagis.2021.24.1.001
  2. U-Net 기반 딥러닝 모델을 이용한 다중시기 계절학적 토지피복 분류 정확도 분석 - 서울지역을 중심으로 - vol.37, pp.3, 2020, https://doi.org/10.7780/kjrs.2021.37.3.4