Sequence Image Interpolation via Separable Convolution Network

Remote-sensing time-series data are significant for global environmental change research and a better understanding of the Earth. However, remote-sensing acquisitions often provide sparse time series due to sensor resolution limitations factors, such as cloud noise optical data. Image interpolation is method that used deal with this issue. This paper considers deep learning learn complex mapping an interpolated intermediate image from predecessor successor images, called separable convolution network sequence interpolation. The uses 1D kernel instead 2D kernels capture spatial characteristics input images then trained end-to-end using images. Our experiments, which were performed unmanned aerial vehicle (UAV) Landsat-8 datasets, show effective produce high-quality data-driven model can simulate diverse nonlinear information.