SCL-GCN: Stratified Contrastive Learning Graph Convolution Network for pavement crack detection from mobile LiDAR point clouds
Accurate pavement crack detection is important for routine maintenance of pavements and reduction of possible traffic accidents. Most existing rule- or learning-based point-level approaches cannot achieve high detection accuracy and efficiency owing to the disorderly arrangement, scattered intensiti...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Elsevier
2023-04-01
|
Series: | International Journal of Applied Earth Observations and Geoinformation |
Subjects: | |
Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843223000705 |
_version_ | 1797842878441979904 |
---|---|
author | Huifang Feng Lingfei Ma Yongtao Yu Yiping Chen Jonathan Li |
author_facet | Huifang Feng Lingfei Ma Yongtao Yu Yiping Chen Jonathan Li |
author_sort | Huifang Feng |
collection | DOAJ |
description | Accurate pavement crack detection is important for routine maintenance of pavements and reduction of possible traffic accidents. Most existing rule- or learning-based point-level approaches cannot achieve high detection accuracy and efficiency owing to the disorderly arrangement, scattered intensities, diverse crack structures, large data volumes, and complex annotation of mobile laser scanning (MLS) point clouds. To address these issues, we developed SCL-GCN, a Stratified Contrastive Learning Graph Convolution Network with a novel dual-branch architecture for MLS-point cloud-based pavement crack detection. First, a multi-scale graph representation construction module was designed based on a stratification strategy. This module creates strengthened spaces for the raw pavement point cloud and its downsampled subset, from which adjacency matrices and initial representations are generated. The stratification strategy samples neighbors densely in the raw point clouds and sparsely in the downsampled subset to form the neighborhood for each point, utilizing long-range contexts to increase the effective receptive field while lowing the extra computation. Next, a graph feature contrastive learning module is proposed to take advantage of stratified features. This module supervises the learning process of the two branches to avoid learning bias caused by an imbalanced data distribution, promoting convergence and improving performance. The experimental results show that the developed SCL-GCN model outperforms state-of-the-art methods. With a training/testing ratio of only 1:6 and an overall training time of less than 70 min, the average precision, recall, and F1-score of the SCL-GCN reached 75.7%, 75.1%, and 75.2%, respectively. |
first_indexed | 2024-04-09T16:55:17Z |
format | Article |
id | doaj.art-dbcb2b88a2304ca8ae6edafb65ecfcb8 |
institution | Directory Open Access Journal |
issn | 1569-8432 |
language | English |
last_indexed | 2024-04-09T16:55:17Z |
publishDate | 2023-04-01 |
publisher | Elsevier |
record_format | Article |
series | International Journal of Applied Earth Observations and Geoinformation |
spelling | doaj.art-dbcb2b88a2304ca8ae6edafb65ecfcb82023-04-21T06:41:10ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322023-04-01118103248SCL-GCN: Stratified Contrastive Learning Graph Convolution Network for pavement crack detection from mobile LiDAR point cloudsHuifang Feng0Lingfei Ma1Yongtao Yu2Yiping Chen3Jonathan Li4Fujian Key Laboratory of Sensing and Computing for Smart Cities, School of Informatics, Xiamen University, Xiamen 361005, ChinaSchool of Statistics and Mathematics, Central University of Finance and Economics, Beijing 102206, China; Corresponding authors.Faculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huaian 223003, ChinaSchool of Geospatial Engineering and Science, Sun Yat-sen University, 519082 Zhuhai, China; Corresponding authors.Department of Geography and Environmental Management, University of Waterloo, Waterloo, ON N2L 3G1, Canada; Department of Systems Design Engineering, University of Waterloo, Waterloo, ON N2L 3G1, CanadaAccurate pavement crack detection is important for routine maintenance of pavements and reduction of possible traffic accidents. Most existing rule- or learning-based point-level approaches cannot achieve high detection accuracy and efficiency owing to the disorderly arrangement, scattered intensities, diverse crack structures, large data volumes, and complex annotation of mobile laser scanning (MLS) point clouds. To address these issues, we developed SCL-GCN, a Stratified Contrastive Learning Graph Convolution Network with a novel dual-branch architecture for MLS-point cloud-based pavement crack detection. First, a multi-scale graph representation construction module was designed based on a stratification strategy. This module creates strengthened spaces for the raw pavement point cloud and its downsampled subset, from which adjacency matrices and initial representations are generated. The stratification strategy samples neighbors densely in the raw point clouds and sparsely in the downsampled subset to form the neighborhood for each point, utilizing long-range contexts to increase the effective receptive field while lowing the extra computation. Next, a graph feature contrastive learning module is proposed to take advantage of stratified features. This module supervises the learning process of the two branches to avoid learning bias caused by an imbalanced data distribution, promoting convergence and improving performance. The experimental results show that the developed SCL-GCN model outperforms state-of-the-art methods. With a training/testing ratio of only 1:6 and an overall training time of less than 70 min, the average precision, recall, and F1-score of the SCL-GCN reached 75.7%, 75.1%, and 75.2%, respectively.http://www.sciencedirect.com/science/article/pii/S1569843223000705Point cloudsPavement crack detectionStratify strategyContrastive learningGraph convolution |
spellingShingle | Huifang Feng Lingfei Ma Yongtao Yu Yiping Chen Jonathan Li SCL-GCN: Stratified Contrastive Learning Graph Convolution Network for pavement crack detection from mobile LiDAR point clouds International Journal of Applied Earth Observations and Geoinformation Point clouds Pavement crack detection Stratify strategy Contrastive learning Graph convolution |
title | SCL-GCN: Stratified Contrastive Learning Graph Convolution Network for pavement crack detection from mobile LiDAR point clouds |
title_full | SCL-GCN: Stratified Contrastive Learning Graph Convolution Network for pavement crack detection from mobile LiDAR point clouds |
title_fullStr | SCL-GCN: Stratified Contrastive Learning Graph Convolution Network for pavement crack detection from mobile LiDAR point clouds |
title_full_unstemmed | SCL-GCN: Stratified Contrastive Learning Graph Convolution Network for pavement crack detection from mobile LiDAR point clouds |
title_short | SCL-GCN: Stratified Contrastive Learning Graph Convolution Network for pavement crack detection from mobile LiDAR point clouds |
title_sort | scl gcn stratified contrastive learning graph convolution network for pavement crack detection from mobile lidar point clouds |
topic | Point clouds Pavement crack detection Stratify strategy Contrastive learning Graph convolution |
url | http://www.sciencedirect.com/science/article/pii/S1569843223000705 |
work_keys_str_mv | AT huifangfeng sclgcnstratifiedcontrastivelearninggraphconvolutionnetworkforpavementcrackdetectionfrommobilelidarpointclouds AT lingfeima sclgcnstratifiedcontrastivelearninggraphconvolutionnetworkforpavementcrackdetectionfrommobilelidarpointclouds AT yongtaoyu sclgcnstratifiedcontrastivelearninggraphconvolutionnetworkforpavementcrackdetectionfrommobilelidarpointclouds AT yipingchen sclgcnstratifiedcontrastivelearninggraphconvolutionnetworkforpavementcrackdetectionfrommobilelidarpointclouds AT jonathanli sclgcnstratifiedcontrastivelearninggraphconvolutionnetworkforpavementcrackdetectionfrommobilelidarpointclouds |