The Added Value of Cycle-GAN for Agriculture Studies

Creative Commons License

Şener E., Çolak E., Erten E., Taşkın Kaya G.

2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, Brussels, Belgium, 11 - 16 July 2021, pp.7039-7042

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/igarss47720.2021.9553876
  • City: Brussels
  • Country: Belgium
  • Page Numbers: pp.7039-7042
  • Istanbul Technical University Affiliated: Yes


It is significant to monitor the phenological stages of agri- cultural crops with accurate and up-to-date information. In monitoring the phenological phases of some crops, optical remote sensing data offers significant spectral information and outstanding feature identification. However, a continu- ous time series of optical remote sensing data is difficult to obtain due to the weather dependency of optical acquisitions. In this paper, the feasibility of transfer learning between the features of Sentinel-1 and Sentinel-2 is evaluated to reduce these difficulties. A feature translation based on deep learning (DL) method, namely Cycle-Consistent Generative Adversar- ial Networks (cycle-GAN), was applied between Sentinel-1 and Sentinel-2 data. In order to evaluate the effect of the cycle-GAN method on crop type mapping and identification, Random Forest classification was applied to four different cases (Real SAR, Fake Optical + Real SAR, Real Optical, and Real Optical + Real SAR).