改进DeepLabV3+网络的遥感影像道路综合提取方法

doi:10.13474/j.cnki.11-2246.2022.0171.

测绘通报 ›› 2022, Vol. 0 ›› Issue (6): 55-61.doi: 10.13474/j.cnki.11-2246.2022.0171.

改进DeepLabV3+网络的遥感影像道路综合提取方法

任月娟, 葛小三

河南理工大学测绘与国土信息工程学院, 河南焦作 454003

收稿日期:2021-07-02 发布日期:2022-06-30
通讯作者: 葛小三。E-mail:gexiaosan@163.com
作者简介:任月娟(1996-),女,硕士生,主要从事遥感、地理信息服务技术方面的研究。E-mail:renyuejuan2020@163.com
基金资助:
河南省自然科学基金(222300420450);国家自然科学基金(41572341);河南省高等教育教学改革研究与实践项目(学位与研究生教育)(2021SJGLX100Y)

An road synthesis extraction method of remote sensing image based on improved DeepLabV3+ network

REN Yuejuan, GE Xiaosan

School of Surveying and Mapping and Land Information Engineering, Henan Polytechnic University, Jiaozuo 454003, China

Received:2021-07-02 Published:2022-06-30

摘要/Abstract

摘要： 遥感图像复杂场景道路提取过程受树木和建筑物阴影,以及荒地、空地等因素干扰较多。针对利用DeepLabV3+网络模型进行道路提取时存在的道路信息不完整和细节信息丢失的问题,本文提出了一种改进DeepLabV3+网络的遥感影像道路提取方法。该方法以轻量级的MobileNetV2作为骨干网络进行特征提取,采用空间金字塔池化模块获得多尺度道路信息特征,从而减少道路遥感图像细节的损失,并提高网络模型的道路提取精度。在DeepGlobe数据集上进行道路提取试验的结果表明,该方法在提升准确率的基础上,有效降低了计算的复杂度;像素准确率和交并比分别达79.7%、64.3%,均优于DeepLabV3+网络及其他经典网络模型,表现出更优异的道路提取能力。

关键词: 道路提取, 改进DeepLabV3+, MobileNetV2, 空间金字塔池化

Abstract: In the process of road extraction of complex scenes from remote sensing images, the shadows of trees and buildings as well as wasteland and open space are often interfered by many factors. In view of the problems of incomplete road information and loss of detail information in road extraction from DeepLabV3+ network model, this paper proposes a road extraction method of remote sensing image based on DeepLabV3 + network, which utilizes lightweight MobileNetV2 as the backbone network for feature extraction. The spatial pyramid pooling module is used to obtain multi-scale road information features to reduce the loss of details of road remote sensing images and improve the accuracy of road extraction. Experimental results of road extraction on the DeepGlobe dataset show that the proposed method can effectively reduce the computational complexity while ensuring that the accuracy is improved. In terms of pixel accuracy and intersection ratio, it reaches 79.7% and 64.3%, respectively, which are superior to DeepLabV3+ network and other classical network models, showing better road extraction ability.

Key words: road extraction, improved DeepLabV3+, MobileNetV2, space pyramid pool

中图分类号:

P237

任月娟, 葛小三. 改进DeepLabV3+网络的遥感影像道路综合提取方法[J]. 测绘通报, 2022, 0(6): 55-61.

REN Yuejuan, GE Xiaosan. An road synthesis extraction method of remote sensing image based on improved DeepLabV3+ network[J]. Bulletin of Surveying and Mapping, 2022, 0(6): 55-61.

参考文献

[1] 吴永静,吴锦超,林超,等.基于深度学习的高分辨率遥感影像光伏用地提取[J].测绘通报,2021(5):96-101.
[2] WANG Shuyang,MU Xiaodong,YANG Dongfang,et al.Road extraction from remote sensing images using the inner convolution integrated encoder-decoder network and directional conditional random fields[J].Remote Sensing,2021,13(3):465-466.
[3] 王斌,陈占龙,吴亮,等.兼顾连通性的U-Net网络高分辨率遥感影像道路提取[J].遥感学报,2020,24(12):1488-1499.
[4] LAPTEV I,MAYER H,LINDEBERG T,et al.Automatic extraction of roads from aerial images based on scale space and snakes[J].Machine Vision and Applications,2000,12(1):23-31.
[5] BENZ U C,HOFMANN P,WILLHAUCK G,et al.Multi-resolution,object-oriented fuzzy analysis of remote sensing data for GIS-ready information[J].ISPRS Journal of Photogrammetry and Remote Sensing,2004,58(3/4):239-258.
[6] ZHAO Shuhe,FENG Xuezhi,KANG Guoding,et al.Multi-source remote sensing image fusion based on support vector machine[J].Chinese Geographical Science,2002,12(3):244-248.
[7] LOPEZ J,BRANCH J W,CHEN Gang.Line-based image segmentation method:a new approach to segment VHSR remote sensing images automatically[J].European Journal of Remote Sensing,2019,52(1):613-631.
[8] 万子健,李连营,杨敏,等.车辆轨迹数据提取道路交叉口特征的决策树模型[J].测绘学报,2019,48(11):1391-1403.
[9] 方莉娜,卢丽靖,赵志远,等.车载激光点云道路边界提取的Snake方法[J].测绘学报,2020,49(11):1438-1450.
[10] LI Peikang,ZANG Yu,WANG Cheng,et al.Road network extraction via deep learning and line integral convolution[C]//Proceedings of 2016 IEEE International Geoscience and Remote Sensing Symposium.Beijing:IEEE,2016:1599-1602.
[11] AMERI F,VALADAN ZOEJ M J.Road vectorisation from high-resolution imagery based on dynamic clustering using particle swarm optimisation[J].The Photogrammetric Record,2015,30(152):363-386.
[12] 贺浩,王仕成,杨东方,等.基于Encoder-Decoder网络的遥感影像道路提取方法[J].测绘学报,2019,48(3):330-338.
[13] MOSTAFA Y.A new shape descriptor for road network separation from parking lots and intersection detection on VHR remote sensing images[J].International Journal of Remote Sensing,2020,41(11):4226-4237.
[14] CHEN Ziyi,LUO Ruixiang,LI J,et al.U-net based road area guidance for crosswalks detection from remote sensing images[J].Canadian Journal of Remote Sensing,2021,47(1):83-99.
[15] CHEN L C,PAPANDREOU G,KOKKINOS I,et al.DeepLab:semantic image segmentation with deep convolutional nets,atrous convolution,and fully connected CRFs[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2018,40(4):834-848.
[16] CHEN L C,ZHU Yukun,PAPANDREOU G,et al.Encoder-decoder with atrous separable convolution for semantic image segmentation[M]//Computer Vision-ECCV 2018.Cham:Springer International Publishing,2018:833-851.
[17] SANDLER M,HOWARD A,ZHU Menglong,et al.MobileNetV2:inverted residuals and linear bottlenecks[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City:IEEE,2018:4510-4520.
[18] CHOLLET F.Xception:deep learning with depthwise separable convolutions[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Honolulu:IEEE,2017:1800-1807.
[19] YANG T J,CHEN Y H,SZE V.Designing energy-efficient convolutional neural networks using energy-aware pruning[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Honolulu:IEEE,2017.

改进DeepLabV3+网络的遥感影像道路综合提取方法

An road synthesis extraction method of remote sensing image based on improved DeepLabV3+ network

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 12

编辑推荐

Metrics

本文评价

[1]	刘茂华, 李曼雯. 融合高分影像和LiDAR数据的城市道路提取[J]. 测绘通报, 2021, 0(6): 1-5,60.
[2]	彭赤, 毛先胤, 杨柳青, 马晓红, 吕乾勇, 杜昊, 王迪, 余容. 多山地区输电通道内道路提取方法[J]. 测绘通报, 2021, 0(4): 105-110,125.
[3]	郑恬静, 黄金彩, 周宝定, 张德津. 基于众源轨迹数据的行人路网提取[J]. 测绘通报, 2021, 0(3): 69-74.
[4]	程效猛, 郑浩, 眭海刚, 冯文卿. 结合矢量引导的高分辨率遥感影像道路自动提取[J]. 测绘通报, 2018, 0(9): 19-23.
[5]	李洪川, 楚恒, 霍英海. 城市道路的多特征多核SVM提取方法[J]. 测绘通报, 2018, 0(2): 72-77.
[6]	陈光, 薛梅, 陈良超, 眭海刚. 自适应聚类学习的道路网提取方法[J]. 测绘通报, 2018, 0(11): 30-35.
[7]	安瑶军, 陈晓轩, 隋立春, 周荣荣. 一种依据点云法向量相似度的结构化道路提取方法[J]. 测绘通报, 2018, 0(11): 69-72.
[8]	魏国武, 王琦, 张阳阳, 陈永生. 道路综合特征下高分辨率遥感影像的提取[J]. 测绘通报, 2017, 0(8): 31-35.
[9]	楚恒, 李洪川, 刘红彬. 基于对象直方图G统计量的遥感影像道路提取[J]. 测绘通报, 2017, 0(12): 63-67.
[10]	张达, 李霖, 李游. 基于车载激光扫描的城市道路提取方法[J]. 测绘通报, 2016, 0(7): 30-34,47.
[11]	张采芳田岩郑毓勇. 基于最优尺度的高分辨率遥感影像道路提取[J]. 测绘通报, 2015, 0(1): 90-94.
[12]	谭仁龙万幼川袁芳李刚. 基于圆形模板的高分辨率遥感影像道路半自动提取[J]. 测绘通报, 2014, 0(10): 63-66.