Korean Journal of Remote Sensing (대한원격탐사학회지)

Korean Society of Remote Sensing (대한원격탐사학회)
권호 표지
DOI QR코드

DOI QR Code

Deep Learning-based Hyperspectral Image Classification with Application to Environmental Geographic Information Systems

딥러닝 기반의 초분광영상 분류를 사용한 환경공간정보시스템 활용

  • Song, Ahram (Department of Civil Environmental Engineering, Seoul National University) ;
  • Kim, Yongil (Department of Civil Environmental Engineering, Seoul National University)
  • 송아람 (서울대학교 건설환경공학부) ;
  • 김용일 (서울대학교 건설환경공학부)
  • Received : 2017.11.14
  • Accepted : 2017.11.28
  • Published : 2017.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

In this study, images were classified using convolutional neural network (CNN) - a deep learning technique - to investigate the feasibility of information production through a combination of artificial intelligence and spatial data. CNN determines kernel attributes based on a classification criterion and extracts information from feature maps to classify each pixel. In this study, a CNN network was constructed to classify materials with similar spectral characteristics and attribute information; this is difficult to achieve by conventional image processing techniques. A Compact Airborne Spectrographic Imager(CASI) and an Airborne Imaging Spectrometer for Application (AISA) were used on the following three study sites to test this method: Site 1, Site 2, and Site 3. Site 1 and Site 2 were agricultural lands covered in various crops,such as potato, onion, and rice. Site 3 included different buildings,such as single and joint residential facilities. Results indicated that the classification of crop species at Site 1 and Site 2 using this method yielded accuracies of 96% and 99%, respectively. At Site 3, the designation of buildings according to their purpose yielded an accuracy of 96%. Using a combination of existing land cover maps and spatial data, we propose a thematic environmental map that provides seasonal crop types and facilitates the creation of a land cover map.

본 연구는 4차 산업의 핵심기술인 인공지능과 환경공간정보의 융합을 통한 정보생산 및 활용가능성을 제시하고자 대표적인 딥러닝(deep-learning) 기법인 CNN(Convolutional Neural Network)을 이용한 영상분류를 수행하였다. CNN은 학습을 통해 스스로 분류기준에 따른 커널의 속성을 결정하며, 최적의 특징영상(feature map)을 추출하여 화소를 분류한다. 본 연구에서는 CNN network를 구성하여 기존의 영상처리 기법으로 해결이 어려웠던 분광특성이 유사한 물질간의 분류 및 GIS속성정보에 따른 분류를 수행하였으며, 항공초분광센서인 CASI(Compact Airborne Spectrographic imager)와 AISA(Airborne Imaging Spectrometer for Application)로 취득된 영상을 이용하였다. 실험대상지역은 총 3곳이며, Site 1과 Site 2는 감자, 양파, 벼 등의 다양한 농작물을 포함하며, Site 3는 단독주거시설, 공동주거시설 등 세분류 토지피복도의 분류 항목으로 구성된 건물을 포함한다. 실험결과, 분류 정확도 96%, 99%로 농작물을 종류에 따라분류하였으며, 96%의 정확도로 건물을 용도에 따라 분류하였다. 본 연구의 결과를 환경공간정보 서비스에 활용하기 위하여 계절별 농작물의 종류를 제공할 수 있는 환경주제도를 제안하였으며, 기존의 토지피복도와 최신 GIS자료를 이용한 세분류 토피지복도 제작 및 갱신 가능성을 확인하였다.

Keywords

References

  1. Cao, X., F. Zhou, L. Xu, D. Meng, Z. Xu, and J. Paisley, 2017. Hyperspectral Image Segmentation with Markov Random Fields and a Convolutional Neural Network, Computer Vision and Pattern Recognition, arXiv preprint arXiv:1705.00727.
  2. Duro, D., S. Franklin, and M. Dube, 2012. A comparison of pixel-based and object-based image analysis with selected machine learning algorithms for the classification of agricultural landscapes using SPOT-5 HRG Imagery, Remote Sensing of Environment, 118: 259-272. https://doi.org/10.1016/j.rse.2011.11.020
  3. LeCun, Y., Y. Bengio, and G. Hinton, 2015. Deep learning, Nature, 521(7553): 436-444. https://doi.org/10.1038/nature14539
  4. Lee, H., R. Kang, K. Kim, G. Nam, M. Kwon, H. Song, S. Cheon, J. Lee, J. Yoon, I. Lee, and H. Lee, 2013. Estimating temporal and spatial variation of chlorophyll-a concentration from multi-spectral imagery in Nak-dong River basin, Water Quality Control Center, NEIR-RP2013-296 (in Korean with English abstract).
  5. Li, Y., H. Zhang, and Q. Shen, 2017. Spectral-Spatial Classification of Hyperspectral Imagery with 3D Convolutional Neural Network, Remote Sensing, 9(1): 67. https://doi.org/10.3390/rs9010067
  6. Krizhevsky, A., I. Sutskever, and G. Hinton, 2012. Imagenet classification with deep convolutional neural networks, In Advances in neural information processing systems, 1097-1105.
  7. Kim, H. and J. 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 (in Korean with English abstract). https://doi.org/10.11108/kagis.2012.15.4.026
  8. Makantasis, K., K. Karantzalos, A. Doulamis, and N. Doulamis, 2015. Deep supervised learning for hyperspectral data classification through convolutional neural networks, Proc. of 2015 In Geoscience and Remote Sensing Symposium (IGARSS), Milan, Italy, Jul. 26-31, pp. 4959-4962.
  9. Ministry of Environment, 2016. Build a land cover map [7th] and improve the national environmental guidance system (in Korean with English abstract).
  10. Petersson, H., D. Gustafsson, and D. Bergstrom, 2016. Hyperspectral image analysis using deep learning -A review. In Image Processing Theory Tools and Applications (IPTA), Proc. of 2016 6th International Conference, Oulu, Finland, Dec. 12-15, pp. 1-6.
  11. Shine, J., T. Lee, P. Jung, and H. Kwon, 2015. A study on land cover map of UAV imagery using an object-based classification method, Journal of the Korean Society for Geospatial Information Science, 23(4): 25-33 (in Korean with English abstract). https://doi.org/10.7319/kogsis.2015.23.4.025
  12. Yu, S., S. Jia, and C. Xu, 2017. Convolutional neural networks for hyperspectral image classification, Neurocomputing, 219: 88-98. https://doi.org/10.1016/j.neucom.2016.09.010

Cited by

  1. 유·무인 항공영상을 이용한 심층학습 기반 녹피율 산정 vol.37, pp.6, 2017, https://doi.org/10.7780/kjrs.2021.37.6.1.22