Road Extraction from UAV Images Using a Deep ResDCLnet Architecture
Obtaining near real-time road features is very important in emergent situations like flood and geological disaster cases. Remote sensing images with very high spatial resolution usually have many details in land use and land cover, which complicate the detection and extraction of road features. In t...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Taylor & Francis Group
2021-05-01
|
Series: | Canadian Journal of Remote Sensing |
Online Access: | http://dx.doi.org/10.1080/07038992.2021.1913046 |
_version_ | 1797661130682793984 |
---|---|
author | Wuttichai Boonpook Yumin Tan Bingxin Bai Bo Xu |
author_facet | Wuttichai Boonpook Yumin Tan Bingxin Bai Bo Xu |
author_sort | Wuttichai Boonpook |
collection | DOAJ |
description | Obtaining near real-time road features is very important in emergent situations like flood and geological disaster cases. Remote sensing images with very high spatial resolution usually have many details in land use and land cover, which complicate the detection and extraction of road features. In this paper, we propose a deep residual deconvolutional network (Deep ResDCLnet), to extract road features from unmanned aerial vehicle (UAV) images. This proposed network is based on the deep neural network from SegNet architecture, the rich skip connection in a residual bottleneck, and the direct relationship among intermediate feature maps from the pixel deconvolution algorithm. It can improve the performance of a supervised learning model by differentiating and extracting complex road features on aerial photographs and UAV imagery. The proposed network is evaluated with the standard public Massachusetts road dataset and the UAV dataset collected alongside Yangtze River, and is compared with four state-of-art network architectures. The results show that the Deep ResDCLnet outperforms all four networks in terms of extraction accuracy, which demonstrates the effectiveness of the network in road extraction from very high spatial resolution imagery. |
first_indexed | 2024-03-11T18:40:37Z |
format | Article |
id | doaj.art-a12e532742b94d37bafc276726dac41b |
institution | Directory Open Access Journal |
issn | 1712-7971 |
language | English |
last_indexed | 2024-03-11T18:40:37Z |
publishDate | 2021-05-01 |
publisher | Taylor & Francis Group |
record_format | Article |
series | Canadian Journal of Remote Sensing |
spelling | doaj.art-a12e532742b94d37bafc276726dac41b2023-10-12T13:36:23ZengTaylor & Francis GroupCanadian Journal of Remote Sensing1712-79712021-05-0147345046410.1080/07038992.2021.19130461913046Road Extraction from UAV Images Using a Deep ResDCLnet ArchitectureWuttichai Boonpook0Yumin Tan1Bingxin Bai2Bo Xu3School of Transportation Science and Engineering, Beihang UniversitySchool of Transportation Science and Engineering, Beihang UniversitySchool of Transportation Science and Engineering, Beihang UniversityDepartment of Geography and Environmental Studies, California State UniversityObtaining near real-time road features is very important in emergent situations like flood and geological disaster cases. Remote sensing images with very high spatial resolution usually have many details in land use and land cover, which complicate the detection and extraction of road features. In this paper, we propose a deep residual deconvolutional network (Deep ResDCLnet), to extract road features from unmanned aerial vehicle (UAV) images. This proposed network is based on the deep neural network from SegNet architecture, the rich skip connection in a residual bottleneck, and the direct relationship among intermediate feature maps from the pixel deconvolution algorithm. It can improve the performance of a supervised learning model by differentiating and extracting complex road features on aerial photographs and UAV imagery. The proposed network is evaluated with the standard public Massachusetts road dataset and the UAV dataset collected alongside Yangtze River, and is compared with four state-of-art network architectures. The results show that the Deep ResDCLnet outperforms all four networks in terms of extraction accuracy, which demonstrates the effectiveness of the network in road extraction from very high spatial resolution imagery.http://dx.doi.org/10.1080/07038992.2021.1913046 |
spellingShingle | Wuttichai Boonpook Yumin Tan Bingxin Bai Bo Xu Road Extraction from UAV Images Using a Deep ResDCLnet Architecture Canadian Journal of Remote Sensing |
title | Road Extraction from UAV Images Using a Deep ResDCLnet Architecture |
title_full | Road Extraction from UAV Images Using a Deep ResDCLnet Architecture |
title_fullStr | Road Extraction from UAV Images Using a Deep ResDCLnet Architecture |
title_full_unstemmed | Road Extraction from UAV Images Using a Deep ResDCLnet Architecture |
title_short | Road Extraction from UAV Images Using a Deep ResDCLnet Architecture |
title_sort | road extraction from uav images using a deep resdclnet architecture |
url | http://dx.doi.org/10.1080/07038992.2021.1913046 |
work_keys_str_mv | AT wuttichaiboonpook roadextractionfromuavimagesusingadeepresdclnetarchitecture AT yumintan roadextractionfromuavimagesusingadeepresdclnetarchitecture AT bingxinbai roadextractionfromuavimagesusingadeepresdclnetarchitecture AT boxu roadextractionfromuavimagesusingadeepresdclnetarchitecture |