DOI QR코드

DOI QR Code

A Study on the Deep Learning-based Tree Species Classification by using High-resolution Orthophoto Images

고해상도 정사영상을 이용한 딥러닝 기반의 산림수종 분류에 관한 연구

  • Received : 2021.06.16
  • Accepted : 2021.07.19
  • Published : 2021.09.30

Abstract

In this study, we evaluated the accuracy of deep learning-based tree species classification model trained by using high-resolution images. We selected five species classed, i.e., pine, birch, larch, korean pine, mongolian oak for classification. We created 5,000 datasets using high-resolution orthophoto and forest type map. CNN deep learning model is used to tree species classification. We divided training data, verification data, and test data by a 5:3:2 ratio of the datasets and used it for the learning and evaluation of the model. The overall accuracy of the model was 89%. The accuracy of each species were pine 95%, birch 89%, larch 80%, korean pine 86% and mongolian oak 98%.

본 연구에서는 드론으로 취득한 고해상도 정사영상 자료를 이용하여, 컨볼루션 신경망(Convolution Neural Network, CNN)을 이용한 딥러닝 기법을 통해 수종에 대한 자동분류 가능성을 분석해 보고자 하였다. 수종판독을 위한 분류항목을 소나무, 자작나무, 낙엽송, 잣나무 그리고 신갈나무 5개 수종으로 선정하였다. 고해상도 정사영상과 임상도를 이용하여 총 5,000개의 데이터셋을 구축하였다. 수종분류를 위한 학습모델로 CNN 기법을 적용하였고, 데이터셋을 5:3:2의 비율로 훈련데이터, 검증테이터, 테스트데이터를 구분하여 모델의 학습 및 평가에 사용하였다. 모델의 전체 정확도는 89%로 나타났으며, 수종별 정확도는 소나무 95%, 자작나무 89%, 낙엽송 80%, 잣나무 86%, 신갈나무 98%로 나타났다.

Keywords

Acknowledgement

본 연구는 산림청(한국임업진흥원) 산림과학기술 연구개발사업'(2019145A00-2121-AB01)'의 지원에 의하여 이루어진 것입니다. 연구 수행을 위해 드론 촬영에 협조해 주신 홍천북방 선도산림경영단지 관계자분들께 감사드립니다.

References

  1. Carleer, A and Wolff E. 2004. Exploitation of very high resolution satellite data for tree species identification. Photogrammetric Engineering & Remote Sensing 70(1):135-140. https://doi.org/10.14358/pers.70.1.135
  2. Cha, S.Y., Pi, U.H., Yi, J.H and Park, C.H. 2011. Identification of two common types of forest cover, Pinus densiflora(PD) and Querqus mongolica(Qm), using the 1st harmonics of a discrete fourier transform, Korean Journal of Remote sensing 27(3):329-338. https://doi.org/10.7780/kjrs.2011.27.3.329
  3. Cho, H.G and Lee, K.S. 2014. Comparision between hyperspectral and multispectral images for the classification of coniferous species. Korean Journal of Remote Sensing 30(1):25-36. https://doi.org/10.7780/kjrs.2014.30.1.3
  4. Dhillon, A and Verma, G. 2020. Convolutional neural network: a review of models, methodologies and applications to object detection, Prog Artif Intell 9:85-112. https://doi.org/10.1007/s13748-019-00203-0
  5. Kim, B.M., Woo, H.S and Park, J.W. 2020. A study of the performance of deep learning-based automatic classification of forest plants: A comparison of data collection methods. Journal of Korean Society of Forest Science 109(1):23-30.
  6. Lee, J.Y. 2020. The Trend of AI image recognition technology. TTA Journal 187:44-51.
  7. Li, Z., Yang, W., Peng, S., and Liu, F. 2020. A survey of convolutional neural networks: Analysis applications and prospects. IEEE Transactions on Neural Networks and Learning Systems 99:1-21.
  8. Lim, J.B., Kim, K.M and Kim, M.K. 2019. The development of major tree species classification model using different satellite images and machine learning in Gwangneung area. Korean Journal of Remote Sensing 35(6):1037-1052.
  9. Nezami, S., Khoramshashi, E., Nevalainen, O. Polonen I and Honkavaara, E. 2020. Tree species classification of drone hype rspectral and RGB imagery with deep lea rning convolutional neural networks. Rem ote Sens 12(7):1070. https://doi.org/10.3390/rs12071070
  10. Park, J.G., Go, S.Y and Cho, G.S. 2013. Vegetation classification using KOMPSAT-2 imagery and high-resolution airborne imagery in urban area. Journal of the Korean Society for Geospatial Information System. 21(4):21-27. https://doi.org/10.7319/KOGSIS.2013.21.4.021
  11. Yang, G., Zhao, Y., Li, B., Ma, Y., Li, R., Jing, J and Dian, Y. 2019. Tree species classification by employing multiple features acquired from integrated sensors. Journal of Sensors. 2019:1-12.

Cited by

  1. 심층신경망을 이용한 KOMPSAT-3/3A/5 영상으로부터 자연림과 인공림의 분류 vol.37, pp.6, 2021, https://doi.org/10.7780/kjrs.2021.37.6.3.5