Building Extraction from UAV Images Based on Attention Enhancement

Chao Fang; Yunmao Liao

doi:10.62051/ijcsit.v2n1.24

Authors

Chao Fang
Yunmao Liao

DOI:

https://doi.org/10.62051/ijcsit.v2n1.24

Keywords:

Building Extraction; UAV Images; Mask R-CNN; CBAM

Abstract

Automatic extraction of buildings from remote sensing images using deep learning methods is crucial for urban and rural construction and management. However, the existing models are affected by the background noise and the complexity of building types during the extraction process, resulting in poor building extraction. In order to solve this problem, this paper proposes an improved model based on Mask R-CNN, which establishes a building instance extraction model for UAV imagery by adding a CBAM attention mechanism module to the backbone residual neural network ResNet. The traditional villages around Beijing are selected as the study area, and the comprehensive experimental results on the homemade building dataset show that the mAP value of the improved model is 86.4%, which is 2.1% higher than that of the original Mask R-CNN model, indicating that the improved model is more effective in building extraction.

Downloads

Download data is not yet available.

References

Y.Z. Liu, B.M. Zhang, J.F. Xu, K. Hou and X. Zhou. Building Extraction from High Resolution Remote Sensing Imagery with Multi-feature and Multi-scale[J]. Bulletin of Surveying and Mapping, 2017, (12):53-57.

F.F. Zhu, Z.Q. Li, S.W. Yang and M. Yang. Urban building object-oriented extraction method based on feature component[J]. Science of Surveying and Mapping,2020,45(1):84-91.

X. Li, J.N. Cao. Obiect-oriented Building Extraction Based on Feature Optimization[J]. Computer Systems & Applications,2022,31(09):360-367.

Q. Liu, X.Y. Hu, X.T. Li and X.L. Qin. Building Recognition Method in Forest Districts Combining the Pixel-level and Object-level[J]. Remote Sensing Technology and Application, 2021,36(06):1350-1357.

C. Fan, H. Jiang. Precise Extraction of Buildings' lnformation in WorldView2 lmages Based on Random Forests[J]. Geospatial Information,2016,14(01):58-62+5.

Z.H. Pan, J. Jin, S.L. Chen, T. Su, S. Liu and N.P. Gao. Automatic Building Extraction Based on Airborne LiDAR Point Cloud Data[J]. Geospatial Information,2022,20(05):57-59+101.

L.Y. Chen, H. Lin, J.H. Wu. Building extraction based on random forest and superpixel segmentation[J]. Bulletin of Surveying and Mapping,2021(02):49-53.

Y.Q. Chen, Y.F. Luo, J.S. Long and X.Y. Guo. Contextual Extraction of Urban Buildings from Airborne LiDAR Point Cloud[J]. Journal of Spatio-temporal Information,2019,26(05): 58-63.

Y.M. Zhang, H.T. Fu, H.Y. Sun, R. Zhang, L.C. Chen and L.H. Pan. A Building Recognition Method for Multispectral lmage Based on lmproved FCN[J]. Computer Engineering,2019, 45(01):239-245.

S. Jin, M. Guan, Y.C. Bian and S.L. Wang. Building Extraction from Remote Sensing lmages Based on lmproved U-Net[J]. Laser & Optoelectronics Progress,2023,60(04):59-65.

Y.J. Ren, X.S. Ge.An road synthesis extraction method of remote sensing image based on improved DeeplabV3+ network[J].Bulletin of Surveying and Mapping, 2022(06):55-61.

Y. Lin, Q.H. Zhao, Z.Y Shen and Y. Li. A building segmentation method for remote sensing with improved SegNet and transfer learning[J]. Science of Surveying and Mapping,2022, 47(06):78-89.

K.M. He, G. Gkioxari, P. Dollár, R. Girshick. Mask R-CNN. In Proceedings of the IEEE International Conference on Computer Vision;2017; pp. 2961–2969.

D.Y. He, W.Z. Shi, Z.B. Lin, X.X. Qiao, Y.X. Liu and Y.H. Lin. Building Extraction from Remote Sensing Image Based on Improved MaskR-CNN[J]. Computer Systems & Applications,2020,29(9):156−163.

N. Lin, T. Huang, P.L. Sun and Y.Y. Wang. Building Extraction of High-resolution Remote Sensing lmagery on Optimized Mask-RCNN[J]. Remote Sensing Information,2022,37(3):1-6.

M.J. Hu, D.J. Feng, Q. Li. Automatic extraction of buildings based on instance segmentation model[J]. Bulletin of Surveying and Mapping,2020,( 4):16-20.

S.Woo, J. Park, J.Y. Lee and I.S. Kweon. CBAM: Convolutional Block Attention Module. In Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany,8–14September 2018, pp,3–19.