Multi-temporal crack segmentation in concrete structure using deep learning approaches

2024-11-07

ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences

AI Summary

To address the insufficient robustness of single-frame image segmentation for early crack detection in concrete structures, this work introduces multi-temporal sequence modeling—first applied to crack segmentation—alongside a lightweight Swim UNETR architecture that achieves significant performance and temporal consistency gains while reducing parameters by ~50%. Leveraging 32-frame temporal crack imagery, the model explicitly captures cross-frame spatiotemporal dependencies, enhancing resilience to noise and geometric deformation. Experiments demonstrate state-of-the-art results: 82.72% IoU and 90.54% F1-score—outperforming the best single-temporal baseline by +6.03% and +4.36%, respectively—yielding more stable predictions and sharper crack boundaries. This study establishes a novel paradigm for temporal visual analysis in structural health monitoring.

Technology Category

• Computer Vision: Segmentation
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Bridging structured and unstructured data
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Abstract. Cracks are among the earliest indicators of deterioration in concrete structures. Early automatic detection of these cracks can significantly extend the lifespan of critical infrastructures, such as bridges, buildings, and tunnels, while simultaneously reducing maintenance costs and facilitating efficient structural health monitoring. This study investigates whether leveraging multi-temporal data for crack segmentation can enhance segmentation quality. Therefore, we compare a Swin UNETR trained on multi-temporal data with a U-Net trained on mono-temporal data to assess the effect of temporal information compared with conventional single-epoch approaches. To this end, a multi-temporal dataset comprising 1356 images, each with 32 sequential crack propagation images, was created. After training the models, experiments were conducted to analyze their generalization ability, temporal consistency, and segmentation quality. The multi-temporal approach consistently outperformed its mono-temporal counterpart, achieving an IoU of 82.72% and a F1-score of 90.54%, representing a significant improvement over the mono-temporal model’s IoU of 76.69% and F1-score of 86.18%, despite requiring only half of the trainable parameters. The multi-temporal model also displayed a more consistent segmentation quality, with reduced noise and fewer errors. These results suggest that temporal information significantly improves the performance of segmentation models, offering a promising solution for improved crack identification and long-term monitoring of concrete structures, even with limited sequential data.

Problem

• Investigating multi-temporal data for enhanced concrete crack segmentation quality
• Comparing deep learning models for temporal versus mono-temporal crack detection
• Improving structural health monitoring through sequential crack propagation analysis

Innovation

• Multi-temporal crack segmentation using deep learning
• Swin UNETR model trained on sequential crack images
• Enhanced segmentation quality with temporal information