Extracting River Illegal Buildings from UAV Image Based on Deeplabv3+
- 47 Downloads
At present, the area extraction and contour identification of illegal buildings in rivers is generally a combination of manual identification and professional software. This method identifies illegal houses with low efficiency, large workload, huge human resource consumption and high requirements for the overall quality of staff. Aiming at the above problems, this paper proposes a method for extracting and identifying the area of illegal buildings in rivers based on deep learning. Identify the Pixel Accuracy (PA) and the Mean Intersection over Union (MIoU) of illegal house method reaching 94.71% and 89.09%. After the deeplabv3+network learns the illegal building features, it automatically detects and identifies the building and generates the building outline shp file. The shp file and the auxiliary arcgis software can be used to extract the illegal area and contour of the building. Based on the method of this paper, the contour marking of river house is shortened from 0.5–1 h to 2.5–5 min compared with the manual identification time. Compared with the manual method for extracting the illegal building area, the area extraction rate is basically above 90%. The results of this method are reliable and in line with actual needs.
KeywordsBuilding area extraction Contour recognition Deep learning Deeplabv3+ Arcgis
This study was jointly supported by China Southern Power Grid Guangzhou Power Supply Bureau Co., Ltd. Key Technology Project (080000KK52190001); Guangdong Provincial Science and Technology Program (2017B010117008); Guangzhou Science and Technology Program (201806010106, 201902010033); the National Natural Science Foundation of China (41976189, 41976190); the Guangdong Innovative and Entrepreneurial Research Team Program (2016ZT06D336); the Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (GML2019ZD0301); the GDAS’s Project of Science and Technology Development (2016GDASRC-0211, 2018GDASCX-0403, 2019GDASYL-0301001, 2017GDASCX-0101, 2018GDAS CX-0101).
- 1.Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNet classification with deep convolutional neural networks. In: Advances in Neural Information Processing Systems, pp. 1097–1105 (2012)Google Scholar
- 2.Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556 (2014)
- 3.Redmon, J., Divvala, S., Girshick, R., et al.: You only look once: unified, real-time object detection, pp. 779–788 (2015)Google Scholar
- 4.Long, J., Shelhamer, E., Darrell, T.: Fully convolutional networks for semantic segmentation. IEEE Trans. Pattern Anal. Mach. Intell. PP(99), 1 (2014)Google Scholar
- 5.Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-24574-4_28CrossRefGoogle Scholar
- 6.Badrinarayanan, V., Kendall, A., Cipolla, R.: SegNet: a deep convolutional encoder-decoder architecture for image segmentation. arXiv preprint arXiv:1511.00561 (2015)
- 7.Lin, G., Milan, A., Shen, C., et al.: RefineNet: multi-path refinement networks for high-resolution semantic segmentation arXiv preprint arXiv:1611.06612 (2016)
- 8.Chen, L.C., Papandreou, G., Kokkinos, I., et al.: Semantic image segmentation with deep convolutional nets and fully connected CRFs. Comput. Sci. 4, 357–361 (2014)Google Scholar
- 10.Chen, L.C., Papandreou, G., Schroff, F., et al.: Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587 (2017)
- 11.Chen, L.C., Zhu, Y., Papandreou, G., et al.: Encoder-decoder with atrous separable convolution for semantic image segmentation. arXiv preprint arXiv:1802.02611 (2018)
- 12.He, K., Zhang, X., Ren, S., et al.: Deep residual learning for image recognition. arXiv preprint arXiv:1512.03385 (2015)