Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention Mechanism

The segmentation and recognition of power lines are crucial for the UAV-based inspection of overhead power lines. To address the issues of class imbalance, low sample quantity, and long-range dependency in images, a specialized semantic segmentation network for power line segmentation called Axial-U...

Full description

Saved in:
Bibliographic Details
Main Authors: Ding Hu, Zihao Zheng, Yafei Liu, Chengkang Liu, Xiaoguo Zhang
Format: Article
Language:English
Published: MDPI AG 2024-12-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/16/23/4585
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849220463178285056
author Ding Hu
Zihao Zheng
Yafei Liu
Chengkang Liu
Xiaoguo Zhang
author_facet Ding Hu
Zihao Zheng
Yafei Liu
Chengkang Liu
Xiaoguo Zhang
author_sort Ding Hu
collection DOAJ
description The segmentation and recognition of power lines are crucial for the UAV-based inspection of overhead power lines. To address the issues of class imbalance, low sample quantity, and long-range dependency in images, a specialized semantic segmentation network for power line segmentation called Axial-UNet++ is proposed. Firstly, to tackle the issue of long-range dependencies in images and low sample quantity, a gated axial attention mechanism is introduced to expand the receptive field and improve the capture of relative positional biases in small datasets, thereby proposing a novel feature extraction module termed axial-channel local normalization module. Secondly, to address the imbalance in training samples, a new loss function is developed by combining traditional binary cross-entropy loss with focal loss, enhancing the precision of image semantic segmentation. Lastly, ablation and comparative experiments on the PLDU and Mendeley datasets demonstrate that the proposed model achieves 54.7% IoU and 80.1% recall on the PLDU dataset, and 79.3% IoU and 93.1% recall on the Mendeley dataset, outperforming other listed models. Additionally, robustness experiments show the adaptability of the Axial-UNet++ model under extreme conditions and the augmented image dataset used in this study has been open sourced.
format Article
id doaj-art-63b54c4cb6a44b61ad7ffd416fb8edd1
institution Kabale University
issn 2072-4292
language English
publishDate 2024-12-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-63b54c4cb6a44b61ad7ffd416fb8edd12024-12-13T16:31:21ZengMDPI AGRemote Sensing2072-42922024-12-011623458510.3390/rs16234585Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention MechanismDing Hu0Zihao Zheng1Yafei Liu2Chengkang Liu3Xiaoguo Zhang4School of Instrument Science and Engineering, Southeast University, Nanjing 210096, ChinaSchool of Software, Southeast University, Suzhou 215123, ChinaSchool of Instrument Science and Engineering, Southeast University, Nanjing 210096, ChinaSchool of Instrument Science and Engineering, Southeast University, Nanjing 210096, ChinaSchool of Instrument Science and Engineering, Southeast University, Nanjing 210096, ChinaThe segmentation and recognition of power lines are crucial for the UAV-based inspection of overhead power lines. To address the issues of class imbalance, low sample quantity, and long-range dependency in images, a specialized semantic segmentation network for power line segmentation called Axial-UNet++ is proposed. Firstly, to tackle the issue of long-range dependencies in images and low sample quantity, a gated axial attention mechanism is introduced to expand the receptive field and improve the capture of relative positional biases in small datasets, thereby proposing a novel feature extraction module termed axial-channel local normalization module. Secondly, to address the imbalance in training samples, a new loss function is developed by combining traditional binary cross-entropy loss with focal loss, enhancing the precision of image semantic segmentation. Lastly, ablation and comparative experiments on the PLDU and Mendeley datasets demonstrate that the proposed model achieves 54.7% IoU and 80.1% recall on the PLDU dataset, and 79.3% IoU and 93.1% recall on the Mendeley dataset, outperforming other listed models. Additionally, robustness experiments show the adaptability of the Axial-UNet++ model under extreme conditions and the augmented image dataset used in this study has been open sourced.https://www.mdpi.com/2072-4292/16/23/4585gated axial attention mechanismsemantic segmentationpower line detectionUNet++loss function
spellingShingle Ding Hu
Zihao Zheng
Yafei Liu
Chengkang Liu
Xiaoguo Zhang
Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention Mechanism
Remote Sensing
gated axial attention mechanism
semantic segmentation
power line detection
UNet++
loss function
title Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention Mechanism
title_full Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention Mechanism
title_fullStr Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention Mechanism
title_full_unstemmed Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention Mechanism
title_short Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention Mechanism
title_sort axial unet power line detection network based on gated axial attention mechanism
topic gated axial attention mechanism
semantic segmentation
power line detection
UNet++
loss function
url https://www.mdpi.com/2072-4292/16/23/4585
work_keys_str_mv AT dinghu axialunetpowerlinedetectionnetworkbasedongatedaxialattentionmechanism
AT zihaozheng axialunetpowerlinedetectionnetworkbasedongatedaxialattentionmechanism
AT yafeiliu axialunetpowerlinedetectionnetworkbasedongatedaxialattentionmechanism
AT chengkangliu axialunetpowerlinedetectionnetworkbasedongatedaxialattentionmechanism
AT xiaoguozhang axialunetpowerlinedetectionnetworkbasedongatedaxialattentionmechanism