Noncontact full-field spatial-temporal deformation measurement of rockfall prevention structure with deep learning-based optic flow algorithm under complex Environments: Full-scale experiments and field tests
•A noncontact vision method was developed for full-field deformation measurement of the flexible barrier system.•A lightweight deep learning-based optical flow method and camera disturbances compensation algorithm was developed.•A dual-threshold mechanism was developed to extract dynamic displacemen...
Gespeichert in:
| Veröffentlicht in: | Mechanical systems and signal processing Jg. 234; S. 112820 |
|---|---|
| Hauptverfasser: | , , , , |
| Format: | Journal Article |
| Sprache: | Englisch |
| Veröffentlicht: |
Elsevier Ltd
01.07.2025
|
| Schlagworte: | |
| ISSN: | 0888-3270 |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Zusammenfassung: | •A noncontact vision method was developed for full-field deformation measurement of the flexible barrier system.•A lightweight deep learning-based optical flow method and camera disturbances compensation algorithm was developed.•A dual-threshold mechanism was developed to extract dynamic displacement curves.•The proposed method is applicable to deformation monitoring of any type of flexible barrier system.•Full-scale experiments and field tests verify the proposed method’s accuracy, robustness, and generalizability.
Flexible barrier systems play a crucial role in rockfall hazard mitigation due to their exceptional energy dissipation capabilities. While deformation patterns of flexible barrier systems during rockfall impacts constitute data for post-event structural assessment and emergency response planning, traditional contact sensor-based approaches face significant limitations in installation complexity and vulnerability to impact-induced damage. Therefore, this paper proposes a non-contact full-field spatial–temporal deformation monitoring framework for flexible barrier systems, integrating a lightweight optical flow algorithm with camera motion compensation technology. The main contributions include three aspects: (1) Development of a computationally efficient architecture integrating lightweight optical flow algorithm with camera motion correction, ensuring sub-pixel accuracy in dynamic deformation measurement; (2) Implementation of a dual-threshold mechanism that combines motion amplitude analysis with boundary gradient detection, effectively resolving the precision-stability paradox in real-time monitoring under complex environmental interference; (3) Comprehensive experimental validation through full-scale impact tests (750 kJ and 1500 kJ energy levels) and in-situ field tests of flexible barrier systems in rockfall disaster region, systematically demonstrating its high accuracy (error matching reduction to 65%∼77.3%), computational efficiency (time-saving of 83.1% ∼89.3%), and environmental robustness. The methodology yields deformation datasets that provide critical technical support for rapid performance evaluation of flexible barrier systems after rockfall events. |
|---|---|
| ISSN: | 0888-3270 |
| DOI: | 10.1016/j.ymssp.2025.112820 |