Journal of the Korea Convergence Society (한국융합학회논문지)

Korea Convergence Society (한국융합학회)
권호 표지
DOI QR코드

DOI QR Code

Convergence of Artificial Intelligence Techniques and Domain Specific Knowledge for Generating Super-Resolution Meteorological Data

기상 자료 초해상화를 위한 인공지능 기술과 기상 전문 지식의 융합

  • Ha, Ji-Hun (IT Division, Korea Oceanic & Atmospheric System Technology) ;
  • Park, Kun-Woo (IT Division, Korea Oceanic & Atmospheric System Technology) ;
  • Im, Hyo-Hyuk (Korea Oceanic & Atmospheric System Technology) ;
  • Cho, Dong-Hee (Department of Computer Science, Kwangwoon University) ;
  • Kim, Yong-Hyuk (Department of Computer Science, Kwangwoon University)
  • 하지훈 ((주)한국해양기상기술 IT본부) ;
  • 박건우 ((주)한국해양기상기술 IT본부) ;
  • 임효혁 ((주)한국해양기상기술) ;
  • 조동희 (광운대학교 컴퓨터과학과) ;
  • 김용혁 (광운대학교 컴퓨터과학과)
  • Received : 2021.09.09
  • Accepted : 2021.10.20
  • Published : 2021.10.28
Download PDF
⟨ Previous Next ⟩

Abstract

Generating a super-resolution meteological data by using a high-resolution deep neural network can provide precise research and useful real-life services. We propose a new technique of generating improved training data for super-resolution deep neural networks. To generate high-resolution meteorological data with domain specific knowledge, Lambert conformal conic projection and objective analysis were applied based on observation data and ERA5 reanalysis field data of specialized institutions. As a result, temperature and humidity analysis data based on domain specific knowledge showed improved RMSE by up to 42% and 46%, respectively. Next, a super-resolution generative adversarial network (SRGAN) which is one of the aritifial intelligence techniques was used to automate the manual data generation technique using damain specific techniques as described above. Experiments were conducted to generate high-resolution data with 1 km resolution from global model data with 10 km resolution. Finally, the results generated with SRGAN have a higher resoltuion than the global model input data, and showed a similar analysis pattern to the manually generated high-resolution analysis data, but also showed a smooth boundary.

고해상도 심층신경망을 이용하여 기상데이터를 초해상화하면 보다 더 정밀한 연구와 실생활에 유용한 서비스를 제공할 수 있다. 본 논문에서는 고해상도 심층신경망 학습에 사용하기 위한 개선된 훈련자료 생산기술을 최초로 제안한다. 기상전문 지식으로 고해상도 기상 자료를 생성하기 위해, 전문 기관의 관측자료와 ERA5 재분석장 자료를 바탕으로 람베르트 정각원추도법과 객관분석을 적용했다. 그 결과, 기상 전문 지식 기반의 기온 및 습도 분석자료는 기존 배경장 대비 RMSE 값이 각각 최대 42%, 46% 개선되었다. 다음으로, 기상 전문 기술을 이용한 수동적인 데이터 생성 기법을 자동화하기 위해 인공지능 기술 중 하나인 SRGAN을 이용했고, 10 km 해상도를 가지는 전지구모델자료로부터 1 km 해상도를 가지는 고해상도 자료를 생성하는 실험을 진행했다. 최종적으로, SRGAN으로 생성한 결과는 전지구모델입력자료에 비해 높은 해상도를 가지며 수동으로 생성한 고해상도 분석자료와 유사한 분석 패턴을 보이면서도 부드러운 경계를 보였다.

Keywords

Acknowledgement

This research was supported by a grant (KMI2019-00310) from Development of Detailed Weather Data Production and Service Technology for Private Use funded by Korea Meteorological Institute (KMI).

References

  1. K. Jia & S. Gong. (2008). Generalized face super-resolution. Journal of IEEE Transactions on Image Processing, 17(6), 873-886. DOI : 10.1109/TIP.2008.922421
  2. I, Sanchez & V. Vilaplana. (2018). Brain MRI super-resolution using 3D generative adversarial networks. arXiv preprint arXiv:1812.11440.
  3. Y. S. Chung, R. K. Park & J. M. Kim. (2014). Study on predictive modeling of incidence of traffic accidents caused by weather conditions. Journal of the Korea Contents Association, 5(1), 9-15. DOI : 10.15207/JKCS.2014.5.1.009
  4. H. J. Yoon. (2016). Development of contents on the marine meteorology sevice by meterology and climate big data. Journal of the Korea Institue of Electronic Communication Sciences, 11(2), 125-138. DOI : 10.13067/JKIECS.2016.11.2.125
  5. H. S. Hwang et al. (2018). Design and implementation of an urban safety service system using realtime weather and atmosphere data. Journal of Korea Multimedia Society, 21(5), 599-608. DOI : 10.9717/KMMS.2018.21.5.599
  6. Y. W. Kim et al. (2017). Design and implementation of a flood disaster safety system using realtime weather big data. Journal of the Korea Contents Association, 17(1), 351-362. DOI : 10.5392/JKCA.2017.17.01.351
  7. M. S. Won et al. (2015). A study on prediction improvement of forest fire danger rating fused with mountain meteorology information. The Korean Society of Agricultural and Forest Meteorology Conference, (pp. 126-130).
  8. J. H. Jeong & Y. T. Chae. (2018). Improvement for forecasting of photovoltaic power output using real time weather data based on machine learning. Journal of the Korean Society of Living Environmental System, 25(1), 119-125. DOI : 10.21086/KSLES.2018.02.25.1.119
  9. J. Kim, J. K. Lee & K. M. Lee. (2016). Accurate image super-resolution using very deep convolutional networks. IEEE Conference on Computer Vision and Pattern Recognition. (pp. 1646-1654). DOI : 10.1109/CVPR.2016.182
  10. K. Stengel et al. (2020). Adversarial super-resolution of climatological wind and solar data. National Acadamy of Sciences, 117(29), 16805-16815. DOI : 10.1073/PNAS.1918964117
  11. H. S. Tark et al. (2017). A new prediction model for power consumption with local weather information. Journal of the Korea Contents Association, 16(11), 488-498. DOI : 10.5392/JKCA.2016.16.11.488
  12. Y. H. Kim et al. (2020). Construction of super-resolution convolutional neural network model for super-resolution of temperature data. Journal of the Korea Convergence Society, 11(8), 7-13. DOI : 10.15207/JKCS.2020.11.8.007
  13. C. Dong et al. (2014). Learning a deep convolution network for image super-resolution. Europen Conference on Computer Vision. (pp. 184-199). DOI : 10.1007/978-3-319-10593-2_13
  14. C. Dong et al. (2015). Image super-resolution using deep convolutional networks. Journal of IEEE Transactions on Pattern Analysis and Machine Intelligence. 38(2). 295-307. DOI : 10.1109/TPAMI.2015.2439281
  15. H. Ren, M. EI-Khamy & J. Lee. (2017). Image super resolution based on fusing multiple convolution neural networks. IEEE Conference on Computer Vision and Pattern Recognition Workshops. (pp. 54-61). DOI : 10.1109/CVPRW.2017.142
  16. M. Ishii et al. (2005). Objective analyses of sea-surface temperature and marine meteorological varialbes for the 20th century using ICOADS and the Kobe collection. Journal of the Royal Meteorological Society, 25(7), 865-879. DOI : 10.1002/JOC.1169
  17. C. Ledig et al. (2017). Photo-realistic single image super-resolution using a generative adversarial network. IEEE Conference on Computer Vision and Pattern Recognition. (pp. 4681-4690). DOI : 10.1109/CVPR.2017.19
  18. M. K. Lee et al. (2018). Detecting anomalies in meteorologoical data using support vector regression. Journal of Advances in Meteorolgy, 2018. DOI : 10.1155/2018/5439256
  19. J. S. Lee et al. (2012). Objective analysis of temperature using the elevation-dependent weighting function. Journal of Atmosphere, 22(2), 233-243. DOI : 10.14191/ATMOS.2012.22.2.233