eduzhai > Applied Sciences > Engineering >

Deep Neural Networks for automatic extraction of features in time series satellite images

  • king
  • (0) Download
  • 20210507
  • Save

... pages left unread,continue reading

Document pages: 7 pages

Abstract: Many earth observation programs such as Landsat, Sentinel, SPOT, and Pleiadesproduce huge volume of medium to high resolution multi-spectral images everyday that can be organized in time series. In this work, we exploit bothtemporal and spatial information provided by these images to generate landcover maps. For this purpose, we combine a fully convolutional neural networkwith a convolutional long short-term memory. Implementation details of theproposed spatio-temporal neural network architecture are provided. Experimentalresults show that the temporal information provided by time series imagesallows increasing the accuracy of land cover classification, thus producingup-to-date maps that can help in identifying changes on earth.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×