An UNet3+ Network based on global pyramid aggregation for change detection in optical remote-sensing imagesGosNIIASLEarning, VIsion and Remote sensing laboratory

Change detection (CD) is a meaningful and challenging task for remote sensing (RS) image analysis. Deep learning (DL) based methods have shown great potential in change detection tasks, there are still two problems with existing deep learning methods such as CNN and Transformer: (1) They do not targ...

Full description

Saved in:
Bibliographic Details
Main Authors: Yanbo Sun, Wenxing Bao, Wei Feng, Kewen Qu, Xuan Ma, Xiaowu Zhang
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:Applied Computing and Geosciences
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2590197424000570
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846117759502516224
author Yanbo Sun
Wenxing Bao
Wei Feng
Kewen Qu
Xuan Ma
Xiaowu Zhang
author_facet Yanbo Sun
Wenxing Bao
Wei Feng
Kewen Qu
Xuan Ma
Xiaowu Zhang
author_sort Yanbo Sun
collection DOAJ
description Change detection (CD) is a meaningful and challenging task for remote sensing (RS) image analysis. Deep learning (DL) based methods have shown great potential in change detection tasks, there are still two problems with existing deep learning methods such as CNN and Transformer: (1) They do not target different depths to extract global semantics in the network; (2) The increase in network depth will lead to uncertainty in the edge pixels of changing targets and the absence of small targets. First, to address this challenge and address these issues, this work proposes a global pyramid aggregation UNet3+ (GPA-UNet3+) change detection model, that uses UNet3+ as the backbone network and connects the encoder and decoder with a pyramid structure. Secondly, a Global Atrous Spatial Pooling Pyramid Module (GASPPM) is proposed. Refined features at different depths and aggregated them to enhance the network’s ability to extract global semantics. Finally, the Edge Enhancement Channel Attention Module (EECAM) is specifically proposed to alleviate the edge pixel uncertainty and spatial position information loss caused by the increase in network depth. Multiple experiments are conducted on two common change detection datasets and a real dataset. Extensive experimental results show that the proposed method outperforms other state-of-the-art methods, achieving the highest F1-score of 90.95%, 95.31%, and 88.32% on the LEVIR-CD dataset, SVCD dataset and Shizuishan Mining Area dataset, respectively.
format Article
id doaj-art-efeb1ba2e10945dfbed8c1c96f0584ab
institution Kabale University
issn 2590-1974
language English
publishDate 2024-12-01
publisher Elsevier
record_format Article
series Applied Computing and Geosciences
spelling doaj-art-efeb1ba2e10945dfbed8c1c96f0584ab2024-12-18T08:51:50ZengElsevierApplied Computing and Geosciences2590-19742024-12-0124100210An UNet3+ Network based on global pyramid aggregation for change detection in optical remote-sensing imagesGosNIIASLEarning, VIsion and Remote sensing laboratoryYanbo Sun0Wenxing Bao1Wei Feng2Kewen Qu3Xuan Ma4Xiaowu Zhang5School of Computer Science and Engineering, North Minzu University, YinChuan, 750021, ChinaSchool of Computer Science and Engineering, North Minzu University, YinChuan, 750021, China; Corresponding author.School of Electronic Engineering, Xidian University, Xi’an, 710071, ChinaSchool of Computer Science and Engineering, North Minzu University, YinChuan, 750021, ChinaSchool of Computer Science and Engineering, North Minzu University, YinChuan, 750021, ChinaSchool of Computer Science and Engineering, North Minzu University, YinChuan, 750021, ChinaChange detection (CD) is a meaningful and challenging task for remote sensing (RS) image analysis. Deep learning (DL) based methods have shown great potential in change detection tasks, there are still two problems with existing deep learning methods such as CNN and Transformer: (1) They do not target different depths to extract global semantics in the network; (2) The increase in network depth will lead to uncertainty in the edge pixels of changing targets and the absence of small targets. First, to address this challenge and address these issues, this work proposes a global pyramid aggregation UNet3+ (GPA-UNet3+) change detection model, that uses UNet3+ as the backbone network and connects the encoder and decoder with a pyramid structure. Secondly, a Global Atrous Spatial Pooling Pyramid Module (GASPPM) is proposed. Refined features at different depths and aggregated them to enhance the network’s ability to extract global semantics. Finally, the Edge Enhancement Channel Attention Module (EECAM) is specifically proposed to alleviate the edge pixel uncertainty and spatial position information loss caused by the increase in network depth. Multiple experiments are conducted on two common change detection datasets and a real dataset. Extensive experimental results show that the proposed method outperforms other state-of-the-art methods, achieving the highest F1-score of 90.95%, 95.31%, and 88.32% on the LEVIR-CD dataset, SVCD dataset and Shizuishan Mining Area dataset, respectively.http://www.sciencedirect.com/science/article/pii/S2590197424000570Change detectionConvolutional neural networksUNet3+Global pyramid aggregationEdge enhancementRemote sensing
spellingShingle Yanbo Sun
Wenxing Bao
Wei Feng
Kewen Qu
Xuan Ma
Xiaowu Zhang
An UNet3+ Network based on global pyramid aggregation for change detection in optical remote-sensing imagesGosNIIASLEarning, VIsion and Remote sensing laboratory
Applied Computing and Geosciences
Change detection
Convolutional neural networks
UNet3+
Global pyramid aggregation
Edge enhancement
Remote sensing
title An UNet3+ Network based on global pyramid aggregation for change detection in optical remote-sensing imagesGosNIIASLEarning, VIsion and Remote sensing laboratory
title_full An UNet3+ Network based on global pyramid aggregation for change detection in optical remote-sensing imagesGosNIIASLEarning, VIsion and Remote sensing laboratory
title_fullStr An UNet3+ Network based on global pyramid aggregation for change detection in optical remote-sensing imagesGosNIIASLEarning, VIsion and Remote sensing laboratory
title_full_unstemmed An UNet3+ Network based on global pyramid aggregation for change detection in optical remote-sensing imagesGosNIIASLEarning, VIsion and Remote sensing laboratory
title_short An UNet3+ Network based on global pyramid aggregation for change detection in optical remote-sensing imagesGosNIIASLEarning, VIsion and Remote sensing laboratory
title_sort unet3 network based on global pyramid aggregation for change detection in optical remote sensing imagesgosniiaslearning vision and remote sensing laboratory
topic Change detection
Convolutional neural networks
UNet3+
Global pyramid aggregation
Edge enhancement
Remote sensing
url http://www.sciencedirect.com/science/article/pii/S2590197424000570
work_keys_str_mv AT yanbosun anunet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT wenxingbao anunet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT weifeng anunet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT kewenqu anunet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT xuanma anunet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT xiaowuzhang anunet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT yanbosun unet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT wenxingbao unet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT weifeng unet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT kewenqu unet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT xuanma unet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory
AT xiaowuzhang unet3networkbasedonglobalpyramidaggregationforchangedetectioninopticalremotesensingimagesgosniiaslearningvisionandremotesensinglaboratory