eduzhai > Applied Sciences > Engineering >

SmaAt-UNet Precipitation Nowcasting using a Small Attention-UNet Architecture

  • king
  • (0) Download
  • 20210505
  • Save

... pages left unread,continue reading

Document pages: 10 pages

Abstract: Weather forecasting is dominated by numerical weather prediction that triesto model accurately the physical properties of the atmosphere. A downside ofnumerical weather prediction is that it is lacking the ability for short-termforecasts using the latest available information. By using a data-driven neuralnetwork approach we show that it is possible to produce an accurateprecipitation nowcast. To this end, we propose SmaAt-UNet, an efficientconvolutional neural networks-based on the well known UNet architectureequipped with attention modules and depthwise-separable convolutions. Weevaluate our approaches on a real-life datasets using precipitation maps fromthe region of the Netherlands and binary images of cloud coverage of France.The experimental results show that in terms of prediction performance, theproposed model is comparable to other examined models while only using aquarter of the trainable parameters.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×