{"title":"基于轻量级学习的水下桥梁检测声纳点云语义分割","authors":"Zelin Huang , Yanjie Zhu , Wen Xiong , Shuaihui Zhang","doi":"10.1016/j.autcon.2025.106387","DOIUrl":null,"url":null,"abstract":"<div><div>Point cloud semantic segmentation of bridge foundations and underwater terrain is essential for structure inspection and scour monitoring. However, underwater environmental noise and rotational disturbances present significant challenges. The existing algorithms demonstrate limited robustness and lack a lightweight model. To address these limitations, a lightweight semantic segmentation network based on the dynamic graph convolutional neural network is proposed. The network employs edge convolutions for feature extraction, a Squeeze and Excitation block for feature refinement, and a densely connected structure enhancing the robustness. A bridge sonar point cloud dataset, BrSPCD, was created for training and evaluation. The experimental results demonstrate superior segmentation performance and robustness, with an F1 score decline of 4.87 % under maximum noise and a decline of 4.11 % under maximum rotational impact. Ablation studies validate the effectiveness of the attention block and dense connectivity. Additionally, complexity analysis shows an 85.02 % reduction in average runtime compared to existing methods.</div></div>","PeriodicalId":8660,"journal":{"name":"Automation in Construction","volume":"178 ","pages":"Article 106387"},"PeriodicalIF":11.5000,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Lightweight learning-based sonar point cloud semantic segmentation for underwater bridge inspection\",\"authors\":\"Zelin Huang , Yanjie Zhu , Wen Xiong , Shuaihui Zhang\",\"doi\":\"10.1016/j.autcon.2025.106387\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Point cloud semantic segmentation of bridge foundations and underwater terrain is essential for structure inspection and scour monitoring. However, underwater environmental noise and rotational disturbances present significant challenges. The existing algorithms demonstrate limited robustness and lack a lightweight model. To address these limitations, a lightweight semantic segmentation network based on the dynamic graph convolutional neural network is proposed. The network employs edge convolutions for feature extraction, a Squeeze and Excitation block for feature refinement, and a densely connected structure enhancing the robustness. A bridge sonar point cloud dataset, BrSPCD, was created for training and evaluation. The experimental results demonstrate superior segmentation performance and robustness, with an F1 score decline of 4.87 % under maximum noise and a decline of 4.11 % under maximum rotational impact. Ablation studies validate the effectiveness of the attention block and dense connectivity. Additionally, complexity analysis shows an 85.02 % reduction in average runtime compared to existing methods.</div></div>\",\"PeriodicalId\":8660,\"journal\":{\"name\":\"Automation in Construction\",\"volume\":\"178 \",\"pages\":\"Article 106387\"},\"PeriodicalIF\":11.5000,\"publicationDate\":\"2025-07-09\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Automation in Construction\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S0926580525004273\",\"RegionNum\":1,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"CONSTRUCTION & BUILDING TECHNOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Automation in Construction","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0926580525004273","RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CONSTRUCTION & BUILDING TECHNOLOGY","Score":null,"Total":0}
Lightweight learning-based sonar point cloud semantic segmentation for underwater bridge inspection
Point cloud semantic segmentation of bridge foundations and underwater terrain is essential for structure inspection and scour monitoring. However, underwater environmental noise and rotational disturbances present significant challenges. The existing algorithms demonstrate limited robustness and lack a lightweight model. To address these limitations, a lightweight semantic segmentation network based on the dynamic graph convolutional neural network is proposed. The network employs edge convolutions for feature extraction, a Squeeze and Excitation block for feature refinement, and a densely connected structure enhancing the robustness. A bridge sonar point cloud dataset, BrSPCD, was created for training and evaluation. The experimental results demonstrate superior segmentation performance and robustness, with an F1 score decline of 4.87 % under maximum noise and a decline of 4.11 % under maximum rotational impact. Ablation studies validate the effectiveness of the attention block and dense connectivity. Additionally, complexity analysis shows an 85.02 % reduction in average runtime compared to existing methods.
期刊介绍:
Automation in Construction is an international journal that focuses on publishing original research papers related to the use of Information Technologies in various aspects of the construction industry. The journal covers topics such as design, engineering, construction technologies, and the maintenance and management of constructed facilities.
The scope of Automation in Construction is extensive and covers all stages of the construction life cycle. This includes initial planning and design, construction of the facility, operation and maintenance, as well as the eventual dismantling and recycling of buildings and engineering structures.