多尺度特征融合空洞卷积ResNet遥感图像建筑物分割

徐胜军; 欧阳朴衍; 郭学源; Taha Muthar Khan; 段中兴

doi:10.37188/OPE.20202807.1588

您当前的位置：

首页 >

文章列表页 >

多尺度特征融合空洞卷积ResNet遥感图像建筑物分割

信息科学 | 更新时间：2020-08-13

- 多尺度特征融合空洞卷积ResNet遥感图像建筑物分割
- Building segmentation in remote sensing image based on multiscale-feature fusion dilated convolution resnet
- 光学精密工程 2020年28卷第7期页码：1588-1599
- 作者机构：
  
  1.西安建筑科技大学信息与控制工程学院, 陕西西安 710055
  2.清华大学建筑设计院有限公司, 北京 100084
- 作者简介：
  
  [ "徐胜军(1976-), 男, 陕西西安人, 工学博士, 副教授, 硕士生导师, 主要从事图像处理、模式识别领域的研究。E-mail:duplin@sina.com" ]
  欧阳朴衍(1993-), 男, 山西长治人, 硕士研究生, 主要从事图像处理、遥感图像等方面的研究。E-mail:jokerman1993@foxmail.com OUYANG Pu-yan, E-mail:jokerman1993@foxmail.com
- 基金信息：
  
  国家自然科学基金资助项目(51678470);国家自然科学基金资助项目(61803293);陕西省教育厅专项科研项目资助(18JK0477);陕西省教育厅专项科研项目资助(2017JM6106);陕西省自然科学基础研究计划资助项目(2020JM-472);陕西省自然科学基础研究计划资助项目(2020JM-473);陕西省自然科学基础研究计划资助项目(2019JQ-760);西安建筑科技大学基础研究基金资助项目(JC1703);西安建筑科技大学基础研究基金资助项目(JC1706)
- DOI：10.37188/OPE.20202807.1588
  中图分类号： P236
- 收稿日期：2020-01-02，
  
  修回日期：2020-03-02，
  
  录用日期：2020-3-2，
  
  纸质出版日期：2020-07-15
- 稿件说明：
移动端阅览
徐胜军, 欧阳朴衍, 郭学源, 等. 多尺度特征融合空洞卷积ResNet遥感图像建筑物分割[J]. 光学精密工程, 2020,28(7):1588-1599.

Sheng-jun XU, Pu-yan OUYANG, Xue-yuan GUO, et al. Building segmentation in remote sensing image based on multiscale-feature fusion dilated convolution resnet[J]. Optics and precision engineering, 2020, 28(7): 1588-1599.
徐胜军, 欧阳朴衍, 郭学源, 等. 多尺度特征融合空洞卷积ResNet遥感图像建筑物分割[J]. 光学精密工程, 2020,28(7):1588-1599. DOI： 10.37188/OPE.20202807.1588.

Sheng-jun XU, Pu-yan OUYANG, Xue-yuan GUO, et al. Building segmentation in remote sensing image based on multiscale-feature fusion dilated convolution resnet[J]. Optics and precision engineering, 2020, 28(7): 1588-1599. DOI： 10.37188/OPE.20202807.1588.

摘要

针对传统建筑物提取方法难以有效描述遥感图像细节特征，导致复杂场景下道路、树木及建筑物之间分割边界不清晰等问题，提出了一种基于多尺度特征融合空洞卷积ResNet(Multiscale-feature Fusion Dilated Convolution ResNet，MFDC-ResNet)模型。首先，为了获取遥感图像建筑物更大范围的特征信息，在深度残差网络中引入空洞卷积增大特征提取的感受野，以捕捉更丰富的多尺度细节特征；其次，为了增强空洞卷积中心点对图像局部区域特征的表达能力，利用3×3卷积核提取遥感图像的中心点区域特征，引入更多的中心点空间先验信息；最后，利用空间金字塔池化模型对不同尺度空洞卷积特征进行融合，获取不同尺度的遥感图像建筑物的上下文信息。在WHU遥感图像数据集上的实验表明，平均交并比mIoU达到0.820，召回率Recall达到0.882。提出算法不仅提高了分割精度，而且有效克服了道路、树木等因素的干扰，得到了较清晰的建筑物边界。

Abstract

To solve the problem in which a traditional ResNet101 model cannot effectively describe the detailed features of remote sensing images

leading to the unclear segmentation boundary between roads

trees

and buildings in complex scenes

a Multiscale-feature Fusion Dilated Convolution ResNet (MFDC-ResNet) was proposed. First

to obtain large-scale building feature information of remote sensing images

a dilated convolution was introduced in the deep residual network to capture richer multi-scale details. Second

to enhance the expression ability of the center point of dilated convolution on the building of feature images

a 3×3 convolution kernel was proposed to extract features in the local area of remote sensing images. Finally

a spatial pyramid pool model of multi-scale feature fusion was proposed to fuse the multi-scale features

obtain the building contextual information of different scales of remote sensing images

and complete the accurate segmentation of buildings. The results of the experiments show that the mean Intersection over Union (mIoU) of building segmentation in WHU is 0.820 and the recall rate is 0.882. The developed method can effectively overcome the influence of roads

trees

and other factors. Moreover

the building boundary can be extracted clearly and smoothly from the remote sensing images and the segmentation accuracy is improved.

关键词

Keywords

references

T G ANDRES , P GANCARSKI , L BERTI-EQUILLE . Remote sensing image analysis by aggregation of segmentation-classification collaborative agents . Pattern Recognition , 2018 . 73 259 - 274 . DOI: 10.1016/j.patcog.2017.08.030 http://doi.org/10.1016/j.patcog.2017.08.030 .

M WANG , Q CUI , J WANG , 等 . Raftcultivation area extraction from high resolution remote sensing imagery by fusing multi-scale region-line primitive association features . Isprs Journal of Photogrammetry & Remote Sensing , 2017 . 123 ( 9 ): 104 - 113 .

J DENG , Y H BAI , F J LYU , 等 . Remote sensing alteration information extraction and threshold determination by band ratio method . Journal of Geology , 2017 . 41 ( 3 ): 504 - 510 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsdz201703020 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsdz201703020 .

M ZHAO , X ZHANG , Z SHI , 等 . Restoration of motion blurred images based on rich edge region extraction using a gray-level co-occurrence matrix . IEEE Access , 2018 . 6 15532 - 15540 . DOI: 10.1109/ACCESS.2018.2815608 http://doi.org/10.1109/ACCESS.2018.2815608 .

闫敬文 , 陈宏达 , 刘蕾 . 高光谱图像分类的研究进展 . 光学精密工程 , 2019 . 27 ( 3 ): 175 - 188 . http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201903020 http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201903020 .

J W YAN , H D CHEN , L LIU . Overview of hyperspectral image classification . Opt.Precision Eng. , 2019 . 27 ( 3 ): 175 - 188 . http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201903020 http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201903020 .

KRIZHEVSKY A, SUTSKEVER I, HINTON G. ImageNet classification with deep convolutional neural networks[C]. NIPS. Curran Associates Inc, 2012.

SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[J]. arXiv: 1409.1556 , 2014.

HE K, ZHANG X, REN S, et al.. Deep residual learning for image recognition[C]. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE Computer Society , 2016.

F YE , X LI , X ZHANG . FusionCNN: a rmote sensing image fusion algorithm based on deep convolutional neural networks . Multimedia Tools and Applications , 2019 . 78 ( 11 ): 14683 - 14703 . DOI: 10.1007/s11042-018-6850-3 http://doi.org/10.1007/s11042-018-6850-3 .

LIUH, ZENG D, TIAN Q. Super-pixel cloud detection using hierarchical fusion CNN[C]. 2018 IEEE Fourth International Conference on Multimedia Big Data (BigMM), IEEE , 2018.

ZHU B, GAO H, WANG X, et al.. Change detection based on the combination of improved SegNet neural network and morphology[C]. 2018 IEEE 3rd International Conference on Image, Vision and Computing (ICIVC), IEEE, 2018.

BADRINARAYANANV , A KENDALL , R CIPOLLA . SegNet: A deep convolutional encoder-decoder architecture for image segmentation . IEEE Transactions on Pattern Analysis and Machine Intelligence , 39 ( 12 ): 2481 - 2495 . DOI: 10.1109/TPAMI.2016.2644615 http://doi.org/10.1109/TPAMI.2016.2644615 .

J Y YANG , Z X ZHOU , Z R DU , 等 . Rural construction land extraction from high spatial resolution remote sensing image based on SegNet semantic segmentation model . Transactions of the Chinese Society of Agricultural Engineering , 2019 . 35 ( 5 ): 251 - 258 . http://d.old.wanfangdata.com.cn/Periodical/nygcxb201905031 http://d.old.wanfangdata.com.cn/Periodical/nygcxb201905031 .

PIRAMANAYAGAM S, SCHWARTZKOPF W, KOEHLER F W, et al.. Classification of remote sensed images using random forests and deep learning framework[C]. SPIE Remote Sensing. Proceedings of the SPIE , 2016, 10004(10): 1-8.

SAITO S, AOKI Y. Building and road detection from large aerial imagery[C]. Image Processing: Machine Vision Applications VⅢ, 2015: 1814-1821.

BITTNER K, CUI S Y, REINARLZ P. Building extraction from remote sensing date using fully convolutional network[C]. ISPRS Hannover Workshop: Hrigi , 2017: 481-486.

L JIAO , M LIANG , H CHEN , 等 . Deep fully convolutional network-based spatial distribution predicton for hypersaectral image classification . IEEE Transaction on Geoscience and Remote Sensing , 2017 . 55 ( 10 ): 5585 - 5599 . DOI: 10.1109/TGRS.2017.2710079 http://doi.org/10.1109/TGRS.2017.2710079 .

Y L XU , M M ZHU , P XIN , 等 . Rapid airplane detection in remote sensing images based on multilayer feature fusion in fully convolutional neural networks . Sensors , 2018 . 18 ( 7 ): 2335 - 2344 . DOI: 10.3390/s18072335 http://doi.org/10.3390/s18072335 .

王舒洋 , 慕晓冬 , 杨东方 , 等 . 融合高阶信息的遥感影像建筑物自动提取 . 光学精密工程 , 2019 . 27 ( 11 ): 2474 - 2483 . http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201911023 http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201911023 .

SH Y WANG , X D MU , D F YANG , 等 . High-order statistics integration method for automatic building extraction of remote sensing images . Opt.Precision Eng. , 2019 . 27 ( 11 ): 2474 - 2483 . http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201911023 http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201911023 .

朱福珍 , 刘越 , 黄鑫 , 等 . 改进的稀疏表示遥感图像超分辨重建 . 光学精密工程 , 2019 . 27 ( 3 ): 213 - 220 . http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201903023 http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201903023 .

F Z ZHU , Y LIU , X HUANG , 等 . Remote sensing image super-resolution based on improved sparse representation . Opt.Precision Eng. , 2019 . 27 ( 3 ): 213 - 220 . http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201903023 http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201903023 .

Y SHI , OTTO CHARLES , K J ANIL . Face clustering: representation and pairwise constraints . IEEE Transactions on Information Forensics & Security , 2017 . 13 ( 7 ): 1626 - 1640 . http://d.old.wanfangdata.com.cn/Periodical/hljkx201411005 http://d.old.wanfangdata.com.cn/Periodical/hljkx201411005 .

浏览量

274

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

融合分形几何特征Resnet遥感图像建筑物分割

面向单幅遥感图像超分辨率的空间自适应及频率融合网络

位置敏感Transformer航拍图像目标检测模型

基于特征融合的改进型PointPillar点云目标检测

融合边缘增强与非局部模块的遥感图像超分辨率重建生成对抗网络