Answers are grounded in this paper and cite their sources.
Local displacements of large-scale bridges bring important information for structural health monitoring. Remote sensing through satellites and synthetic aperture radar (SAR) imagery displays significant opportunities to provide these displacements. However, challenges such as restricted access to images, inability to perform real-time monitoring akin to vibration-based SHM, and difficulties posed by large sizes and speckle noise of SAR images complicate their use in long-term monitoring programs. Environmental variability, especially seasonal temperature changes, can also influence displacements masking actual impacts of damage. To address these challenges, an innovative stacking ensemble regression method is proposed to simultaneously forecast and normalize small datasets of displacement responses retrieved from limited SAR images. This method comprises two levels of non-parametric base regressors and a parametric meta regressor trained by the predictions of the base learners along with the original response data. The base regressors are univariate, robust, and Bayesian linear regression models, while the meta regressor is developed from the ridge regression. The effectiveness and practicality of the proposed method are demonstrated through limited temperature and displacement samples of two large-scale steel bridges. Results show high prediction accuracy and successful normalization capabilities of the proposed method.
No takes yet. Share an insight, caveat, or question.
Entezami et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: