Abstract: Using deep learning to extract roads from satellite images is one of the most popular methods. However, the existing encoder-decoder-based deep networks usually produce fragmented roads, due ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results