Building extraction in complex scenes based on the fusion of multi-feature improved PSPNet model

doi:10.13474/j.cnki.11-2246.2021.0170

Abstract

Abstract: Aiming at the problem of low accuracy of building extraction in complex scenes of high-resolution remote sensing images, this paper proposes an improved PSPNet model which integrates multiple features. On the basis of PSPNet network, the expansion convolution module is added and the shallow features of the image are fused. The results show that the overall prediction accuracy by the improved PSPNet model is 95.90%, and the average building extraction accuracy is 77.77%, which is higher than other models. It varies in performance from scene to scene. In the first scene that is complex the prediction accuracy is as high as 80.35%; in the second scene with village buildings in the city, the prediction accuracy is 75%; in the third scene with high-rise buildings, the prediction accuracy is 78.11%. This model can effectively improve the extraction accuracy of buildings in complex scenes of high-resolution remote sensing images.

Key words: semantic segmentation, building extraction, PSPNet, dilated convolution, pyramid pooling moduel

CLC Number:

P237

WU Hua, ZHANG Xinchang, SUN Ying, CAI Weinan, YAN Jun, DENG Jianwen, ZHANG Jianguo. Building extraction in complex scenes based on the fusion of multi-feature improved PSPNet model[J]. Bulletin of Surveying and Mapping, 2021, 0(6): 21-27.

References

[1] 谢嘉丽,李永树,李何超,等.利用灰度共生矩阵纹理特征识别空心村损毁建筑物的方法[J].测绘通报,2017(12):90-93,102.
[2] 吴剑,陈鹏,刘耀林,等.震害损毁建筑物高分辨率遥感信息提取方法[J].地理与地理信息科学,2013,29(3):35-38,47,2.
[3] HUANG X,ZHANG L.An SVM ensemble approach combining spectral,structural, and semantic features for the classification of high-resolution remotely sensed imagery[J].IEEE Transactions on Geoscience and Remote. Sensing,2013,51(1):257-272.
[4] GHANEA M,MOALLEM P,MOMENI M.Automatic building extraction in dense urban areas through GeoEye multispectral imagery[J].International Journal of Remote Sensing,2014,35(13):5094-5119.
[5] WANG J,YANG X C,QIN X B,et al.An efficient approach for automatic rectangular building extraction from very high-resolution optical satellite imagery[J].IEEE Geoscience and Remote Sensing Letters, 2015,12(3):487-491.
[6] HUANG X, ZHANG L P, A multidirectional and multiscale morphological index for automatic building extraction form multispectral GeoEye-1 imagery[J]. Photogrammetric Engineering and Remote Sensing, 2011,77(7):721-732.
[7] 胡荣明,黄小兵,黄远程.增强形态学建筑物指数应用于高分辨率遥感影像中建筑物提取[J].测绘学报,2014,43(5):514-520.
[8] FAN J L,XU W, WU Y, et al. Human tracking using convolution neural networks[J].IEEE Transactions on Neural Networks, 2010,21(10):1610-1623.
[9] LONG J,SHELHAMER E,DARRELL T,et al. Fully convolu-tional networks for semantic segmentation[C]//Proceeding of IEEE Conference on Computer Vision and Pattern Recognition. Washington D C:IEEE Computer Society,2015.
[10] RONNEBERGER O,FISCHER P,BROX T. U-Net:convolu-tional networks for biomedical image segmentation[C] //Proceedings of the 18th International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham:Springer,2015.
[11] HU J,SHEN L,ALBANIE S,et al. Squeeze-and-excitation networks[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2020,42(8):2011-2023.
[12] ZHAO H S,SHI J P,QI X J,et al.Pyramid scene parsing network[C]//Proceeding of the 2017 IEEE Conference on Computer Vision and Pattern Recognition.New York: IEEE, 2017.
[13] HE K M,ZHANG X Y,REN S Q,et al. Deep residual learning for image recognition[C]//Proceeding of IEEE Conference on Computer Vision and Pattern Recognition. Washington D C:IEEE Computer Society, 2016.
[14] 陈凯强,高鑫,闫梦龙,等.基于编解码网络的航空影像像素级建筑物提取[J].遥感学报,2020,24(9):1134-1142.
[15] 张浩然,赵江洪,张晓光.利用U-net网络的高分遥感影像建筑提取方法[J].遥感信息,2020,35(3):143-150.
[16] 刘钊,廖斐凡,赵桐.基于PSPNet的遥感影像城市建成区提取及其优化方法[J].国土资源遥感,2020,32(4): 84-89.
[17] 刘亦凡,张秋昭,王光辉,等.利用深度残差网络的遥感影像建筑物提取[J].遥感信息,2020,35(2):59-64.
[18] SHAO Z F, TANG P H, WANG Z Y,et al. BRRNet: A fully convolutional neural network for automatic building extraction from high-resolution remote sensing images[J], Remote Sensing, 2020, 12(6): 1050.