Consistency Self-Training Semi-Supervised Method for Road Extraction from Remote Sensing Images

Road extraction techniques based on remote sensing image have significantly advanced. Currently, fully supervised road segmentation neural networks based on remote sensing images require a significant number of densely labeled road samples, limiting their applicability in large-scale scenarios. Cons...

Full description

Saved in:
Bibliographic Details
Main Authors: Xingjian Gu, Supeng Yu, Fen Huang, Shougang Ren, Chengcheng Fan
Format: Article
Language:English
Published: MDPI AG 2024-10-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/16/21/3945
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850062336723779584
author Xingjian Gu
Supeng Yu
Fen Huang
Shougang Ren
Chengcheng Fan
author_facet Xingjian Gu
Supeng Yu
Fen Huang
Shougang Ren
Chengcheng Fan
author_sort Xingjian Gu
collection DOAJ
description Road extraction techniques based on remote sensing image have significantly advanced. Currently, fully supervised road segmentation neural networks based on remote sensing images require a significant number of densely labeled road samples, limiting their applicability in large-scale scenarios. Consequently, semi-supervised methods that utilize fewer labeled data have gained increasing attention. However, the imbalance between a small quantity of labeled data and a large volume of unlabeled data leads to local detail errors and overall cognitive mistakes in semi-supervised road extraction. To address this challenge, this paper proposes a novel consistency self-training semi-supervised method (CSSnet), which effectively learns from a limited number of labeled data samples and a large amount of unlabeled data. This method integrates self-training semi-supervised segmentation with semi-supervised classification. The semi-supervised segmentation component relies on an enhanced generative adversarial network for semantic segmentation, which significantly reduces local detail errors. The semi-supervised classification component relies on an upgraded mean-teacher network to handle overall cognitive errors. Our method exhibits excellent performance with a modest amount of labeled data. This study was validated on three separate road datasets comprising high-resolution remote sensing satellite images and UAV photographs. Experimental findings showed that our method consistently outperformed state-of-the-art semi-supervised methods and several classic fully supervised methods.
format Article
id doaj-art-ffd07efe1e404cfeaa9a04a97a5a0022
institution DOAJ
issn 2072-4292
language English
publishDate 2024-10-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-ffd07efe1e404cfeaa9a04a97a5a00222025-08-20T02:49:56ZengMDPI AGRemote Sensing2072-42922024-10-011621394510.3390/rs16213945Consistency Self-Training Semi-Supervised Method for Road Extraction from Remote Sensing ImagesXingjian Gu0Supeng Yu1Fen Huang2Shougang Ren3Chengcheng Fan4College of Artificial Intelligence, Nanjing Agricultural University, Nanjing 210095, ChinaCollege of Artificial Intelligence, Nanjing Agricultural University, Nanjing 210095, ChinaCollege of Artificial Intelligence, Nanjing Agricultural University, Nanjing 210095, ChinaCollege of Artificial Intelligence, Nanjing Agricultural University, Nanjing 210095, ChinaInnovation Academy for Microsatellites of CAS, Shanghai 201210, ChinaRoad extraction techniques based on remote sensing image have significantly advanced. Currently, fully supervised road segmentation neural networks based on remote sensing images require a significant number of densely labeled road samples, limiting their applicability in large-scale scenarios. Consequently, semi-supervised methods that utilize fewer labeled data have gained increasing attention. However, the imbalance between a small quantity of labeled data and a large volume of unlabeled data leads to local detail errors and overall cognitive mistakes in semi-supervised road extraction. To address this challenge, this paper proposes a novel consistency self-training semi-supervised method (CSSnet), which effectively learns from a limited number of labeled data samples and a large amount of unlabeled data. This method integrates self-training semi-supervised segmentation with semi-supervised classification. The semi-supervised segmentation component relies on an enhanced generative adversarial network for semantic segmentation, which significantly reduces local detail errors. The semi-supervised classification component relies on an upgraded mean-teacher network to handle overall cognitive errors. Our method exhibits excellent performance with a modest amount of labeled data. This study was validated on three separate road datasets comprising high-resolution remote sensing satellite images and UAV photographs. Experimental findings showed that our method consistently outperformed state-of-the-art semi-supervised methods and several classic fully supervised methods.https://www.mdpi.com/2072-4292/16/21/3945semi-supervisedsemantic segmentationgenerative adversarial networkroad extractionremote sensing image
spellingShingle Xingjian Gu
Supeng Yu
Fen Huang
Shougang Ren
Chengcheng Fan
Consistency Self-Training Semi-Supervised Method for Road Extraction from Remote Sensing Images
Remote Sensing
semi-supervised
semantic segmentation
generative adversarial network
road extraction
remote sensing image
title Consistency Self-Training Semi-Supervised Method for Road Extraction from Remote Sensing Images
title_full Consistency Self-Training Semi-Supervised Method for Road Extraction from Remote Sensing Images
title_fullStr Consistency Self-Training Semi-Supervised Method for Road Extraction from Remote Sensing Images
title_full_unstemmed Consistency Self-Training Semi-Supervised Method for Road Extraction from Remote Sensing Images
title_short Consistency Self-Training Semi-Supervised Method for Road Extraction from Remote Sensing Images
title_sort consistency self training semi supervised method for road extraction from remote sensing images
topic semi-supervised
semantic segmentation
generative adversarial network
road extraction
remote sensing image
url https://www.mdpi.com/2072-4292/16/21/3945
work_keys_str_mv AT xingjiangu consistencyselftrainingsemisupervisedmethodforroadextractionfromremotesensingimages
AT supengyu consistencyselftrainingsemisupervisedmethodforroadextractionfromremotesensingimages
AT fenhuang consistencyselftrainingsemisupervisedmethodforroadextractionfromremotesensingimages
AT shougangren consistencyselftrainingsemisupervisedmethodforroadextractionfromremotesensingimages
AT chengchengfan consistencyselftrainingsemisupervisedmethodforroadextractionfromremotesensingimages