A bridge point cloud databank for digital bridge understanding DOI Creative Commons
Hongwei Zhang, Yanjie Zhu, Wen Xiong

и другие.

Computer-Aided Civil and Infrastructure Engineering, Год журнала: 2024, Номер unknown

Опубликована: Ноя. 25, 2024

Abstract Despite progress in automated bridge point cloud segmentation based on deep learning, challenges persist. For instance, the absence of a public dataset specifically designed for instances, and existing datasets display lack diversity types inconsistency component labeling. These factors may hinder further improvement accuracy segmentation. In this paper, universal multi‐type databank, named BrPCD, consisting total 98 data (PCD; 10 them are obtained from scanning, rest is by augmentation) small to long‐span bridges, established. Additionally, method augmenting PCD proposed, significantly enriching spatial feature information bridges within dataset. Furthermore, introduced annotation rules, uniform categorization semantic labels components implemented, enhancing applicability our across various tasks different bridges. A benchmark testing was conducted BrPCD using PointNet model. The results indicate that parameters learned through enable accurate at level components. other words, can function as dataset, applicable networks aimed

Язык: Английский

A BIM-based virtual trial assembly approach for long-span steel structures DOI Creative Commons
Hui Gao, Donglin Wang,

Shida Wu

и другие.

Alexandria Engineering Journal, Год журнала: 2025, Номер 116, С. 644 - 659

Опубликована: Янв. 9, 2025

Язык: Английский

Процитировано

0
A bridge point cloud databank for digital bridge understanding DOI Creative Commons
Hongwei Zhang, Yanjie Zhu, Wen Xiong

и другие.

Computer-Aided Civil and Infrastructure Engineering, Год журнала: 2024, Номер unknown

Опубликована: Ноя. 25, 2024

Abstract Despite progress in automated bridge point cloud segmentation based on deep learning, challenges persist. For instance, the absence of a public dataset specifically designed for instances, and existing datasets display lack diversity types inconsistency component labeling. These factors may hinder further improvement accuracy segmentation. In this paper, universal multi‐type databank, named BrPCD, consisting total 98 data (PCD; 10 them are obtained from scanning, rest is by augmentation) small to long‐span bridges, established. Additionally, method augmenting PCD proposed, significantly enriching spatial feature information bridges within dataset. Furthermore, introduced annotation rules, uniform categorization semantic labels components implemented, enhancing applicability our across various tasks different bridges. A benchmark testing was conducted BrPCD using PointNet model. The results indicate that parameters learned through enable accurate at level components. other words, can function as dataset, applicable networks aimed

Язык: Английский

Процитировано

1