Spatial interpolation using conditional generative adversarial neural networks

Di Zhu, Ximeng Cheng, Fan Zhang, Xin Yao, Yong Gao, Yu Liu

Research output: Contribution to journalArticlepeer-review

62 Scopus citations


Spatial interpolation is a traditional geostatistical operation that aims at predicting the attribute values of unobserved locations given a sample of data defined on point supports. However, the continuity and heterogeneity underlying spatial data are too complex to be approximated by classic statistical models. Deep learning models, especially the idea of conditional generative adversarial networks (CGANs), provide us with a perspective for formalizing spatial interpolation as a conditional generative task. In this article, we design a novel deep learning architecture named conditional encoder-decoder generative adversarial neural networks (CEDGANs) for spatial interpolation, therein combining the encoder-decoder structure with adversarial learning to capture deep representations of sampled spatial data and their interactions with local structural patterns. A case study on elevations in China demonstrates the ability of our model to achieve outstanding interpolation results compared to benchmark methods. Further experiments uncover the learned spatial knowledge in the model’s hidden layers and test the potential to generalize our adversarial interpolation idea across domains. This work is an endeavor to investigate deep spatial knowledge using artificial intelligence. The proposed model can benefit practical scenarios and enlighten future research in various geographical applications related to spatial prediction.

Original languageEnglish (US)
Pages (from-to)735-758
Number of pages24
JournalInternational Journal of Geographical Information Science
Issue number4
StatePublished - Apr 2 2020
Externally publishedYes

Bibliographical note

Funding Information:
This research was supported by the National Natural Science Foundation of China [41625003 and 41830645] and the National Key Research and Development Program of China [2017YFB0503602] and the Open Project Fund of the institute for China Sustainable Urbanization, Tsinghua University (TUCSU-K-17026-01). The authors would like to thank Dr. Lei Dong, Dr. Michael Goodchild, Dr. Tao Cheng, Dr. Krzysztof Janowicz, Dr. May Yuan and the anonymous referees for their insightful comments.

Publisher Copyright:
© 2019, © 2019 Informa UK Limited, trading as Taylor & Francis Group.


  • deep learning
  • encoder-decoder
  • generative adversarial networks
  • Spatial interpolation
  • spatial prediction


Dive into the research topics of 'Spatial interpolation using conditional generative adversarial neural networks'. Together they form a unique fingerprint.

Cite this