Due to the complex environment surrounding buildings and the varying scales of buildings,smaller buildings are prone to issues such as incomplete or missed segmentation,which leads to a reduction in the accuracy of extraction results. To address these challenges,an improved model based on the Unet++ architecture (REUnet++) was proposed. This model enhanced performance by incorporating a ResNet34 residual network as the encoder structure. Additionally,an attention module,SE,was integrated into the model to strengthen its ability to extract important features from the dataset. Experiments were conducted on the public xBD dataset. The experimental results demonstrate that the REUnet++ model outperforms existing models in both feature extraction and segmentation accuracy in complex scenarios. Compared to the Unet++ model,the F1 score increases by 3.08%,and the intersection over union score improves by 4.68%,with significant improvements observe in other relevant performance metrics. Finally,the model’s generalization ability was further validated using the WHU building dataset.
YU Shuang-shuang, KANG Shuai, ZHANG Jian-jun, JIN Man, YU Ye.
Remote Sensing Image Building Segmentation Method Based on Improved Unet++ Network[J].
Science Technology and Engineering, 2026, 26(4): 1607-1615 DOI:10.12404/j.issn.1671-1815.2500782
为了验证所提方法的有效性,基于预测值与标签之间的比较应用5个评价指标,即精确率(precision,记为P),召回率(recall,记为R),F1分数(F1-score,记为F1),交并比(intersection over union,IoU)和准确率(accuracy,记为Acc)。P反映模型在处理整个数据集时的准确性;R主要验证模型识别出的正类样本的准确性;F1分数适用于类别不平衡情况下的性能评估,能够平衡召回率和准确率的影响,提供一个更为全面的评价;IoU是用来衡量预测边界框与真实边界框之间重叠程度的指标,IoU的值越高,说明模型对对象的定位越精确;Acc是衡量模型预测正确性的重要标准。这5个评估指标可描述为
YangMingwang, ZhaoLike, YeLinfeng, et al. A review of building extraction methods for remote sensing images based on convolutional neural network[J]. Journal of Earth Information Science, 2024, 26(6): 1500-1516.
LiPengsong, LiJunda, WuLiangwu, et al. Image recognition algorithm based on thresholdsegmentation method and convolutional neural network[J]. Journal of Jilin University (Science Edition), 2020, 58(6): 1436-1442.
Ma YanGulimira·Kezierbieke. A review of research on image semantic segmentation methods in high resolution remote sensing image decoding[J]. Computer Science and Exploration, 2023, 17(7): 1526-1548.
MaXiangyue, XuJindong, NiMengying. Remote sensing image segmentation based on mult-iscale feature fuzzy convolutional neural network[J]. Journal of Shandong University (Engineering Edition), 2024, 54(3): 44-54.
[19]
LongJ, ShwlhamerE, DarrelT. Fully convolutional networks for semantic segmentation[C]// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2015: 3431-3440.
[20]
ShengH, ZhangY, ZhangW, et al. ME-FCN: a multi-scale feature-enhanced fully convolutional network for building footprint extraction[J]. Remote Sensing, 2024, 16(22): 4305.
ZhuJunbin, DuBin, XuShimin. Crack detect-ion using deep supervised networks with multi-scale and multi-channel feature fusion[J]. Science Technology and Engineering, 2024, 24(13): 5595-5603.
[23]
ChenG, TanX, GuoB, et al. SDFCNv2: an improved FCN framework for remote sensing images semantic segmentation[J]. Remote Sensing, 2021, 13(23): 4902.
[24]
TianL, ZhongX, ChenM. Semantic segmentation of remote sensing image based on GAN and FCN network model[J]. Scientific Programming, 2021, 11: 9491376.
XiYubin, ZhaoLiangjun, NingFeng, et al. Remote sensing image segmentation method based on improved DeepLabV3+[J]. Modern Electronic Technique, 2019, 47(11): 51-58.
SuZhipeng, LiJingwen, JiangJianwu, et al. Semantie segmentation method for remote sensing images based on improved DeepLabV3+[J]. Laser & Optoelectronics Progress, 2023, 60(6): 359-366.
[29]
ChenJ W, LiY, ChenJ, et al. Building extraction method based on MFF-Deeplabv3+ network for high-resolution remote sensing images[J]. Journal of University of Chinese Academy of Sciences, 2024, 41(5): 654-664.
[30]
LiuJ, GuH, LiZ, et al. Multi-scale feature fusion attention network for building extraction in remote sensing images[J]. Electronics, 2024, 13(5): 923.
SheXiangyang, MaYijun. Improved semantic segmentation algorithm for remote sensing images[J]. Computer Engineering and Science, 2023, 45(3): 504-511.
[33]
ChenT, MaoM, WuJ, et al. SAU-Net: a novel network for building extraction from high resolution remote sensing images by reconstructing fine-grained semantic features[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 9: 6747-6761.
YueYuchao, WangYingmei, QinJiachuan. CADCNet: an improved algorithm for retinal vessel segmentation[J]. Science Technology and Engineering, 2025, 25(3): 962-968.
[36]
YuM, ChenX, ZhangW, et al. AGs-Unet: building extraction model for high resolution remote sensing images based on attention gates U network[J]. Sensors, 2022, 22(8): 2932.
[37]
HanR, FanX, LiuJ. EUNet: edge-UNet for accurate building extraction and edge emphais in Gaofen-7 images[J]. Remote Sensing, 2024, 16(13): 2397.
[38]
ZhouZ W, RanmanS M M, TajbakhshN, et al. Unet++: a nested U-Net architecture for medical image segmentation[C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support: 4th International Workshop and 8th International Workshop. Granada: Springer, 2018: 1-9.
[39]
HuJ, ShenL, SunG. Squeeze-and-excitation networks[C]// IEEE Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2018: 7132-7141.
[40]
RonnebrgerO, FischerP, BroxT. U-Net: convolutional networks for biomedical image segmentation[C]// Medical Image Computing and Computer-Assisted Intervention. Berlin: Springer, 2015: 234-241.
[41]
HeK M, ZhangX Y, RENS Q, et al. Deep residual learning for image recognition[C]// IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV: IEEE, 2016: 770-778.
[42]
GuptaR, GoodmanB, RichardH, et al. XBD: a dataset for assessing building damage from satellite imagery[J]. arXiv Preprint, arXiv. 1911.09296, 2019.
SunShukui, FanJing, SunZhongqiang, et al. A review of research on image data enhancement based on deep learning[J]. Computer Science, 2024, 51(1): 150-167.
[45]
JiS P, WeiS Q, LuM. Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(1): 574-586.