**5. Conclusions**

We propose an e fficient road extraction method based on a convolution neural network for high-resolution remote sensing images. The model combines the virtue of dense connection mode and U-Net and solves the problem of tree and shadow occlusion to a certain extent, which we call DenseUNet. In particular, we use a U-Net architecture combined with a suitable weighted loss function to place more emphasis on foreground pixels. Following simple connection rules (fractal extensions), DenseUNet naturally integrates deep supervision, the properties of identity mappings, and diversified depth attributes. The dense connections within dense units and the skip connections between the encoding and decoding paths of the network will help to transfer information and accelerate computation, so they can learn more compactly and ge<sup>t</sup> more accurate models.

Although deep neural networks have acquired remarkable success in many fields, there are no sophisticated theories yet. However, one of the critical disadvantages of deep learning models is their limited interpretability, and often these models are described as "black boxes" that do not provide insight into their inner workings. On the other hand, it will be challenging to create a general model through theoretical guidance. Hence, the results obtained from such specific planning problem are di fficult to apply to other problems in the same field. We plan to use the trained DenseUNet model to transfer knowledge to improve new tasks in future work.

**Author Contributions:** J.X. and X.Z. conceived and designed the method; J.X. performed the experiments and wrote the manuscript; Z.Z. provided the Conghua dataset; W.F. revised the manuscript.

**Funding:** This research was funded by National Key R&D Program of China (Grant No. 2018YFB2100702), National Natural Science Foundation of China (Grant Nos. 41875122, 41431178, 41801351 and 41671453), Natural Science Foundation of Guangdong Province: 2016A030311016, Research Institute of Henan Spatio-Temporal Big Data Industrial Technology: 2017DJA001, National Administration of Surveying, Mapping and Geoinformation of China (Grant No. GZIT2016-A5-147), Hunan Botong Information Co.,ltd.: BTZH2018001.

**Acknowledgments:** The authors would like to thank the anonymous reviewers for their constructive comments. The authors are also grateful to Hinton et al. for providing the Massachusetts dataset.

**Conflicts of Interest:** The authors declare no conflicts of interest.
