.png){kind=icon}
.png){kind=icon}
Return
An adaptive phase feedback algorithm for displacement tracking from video frames
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ymssp.2025.113521.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Computer vision-based non-contact displacement measurement techniques have demonstrated significant potential for structural health monitoring. A new category of Phase Motion Estimation (PME) method has been developed recently extracting spatial motion information of structures by analyzing the phase information in image sequences. It has sub-pixel accuracy without the need of artificial markers. However, conventional PME methods suffer from a subjectively selected wavelength parameter that may lead to limited computation efficiency if it is too large, or wrapping of phase information if it is too small. This study proposes an adaptive wavelength for the phase-feedback to address this issue. It adaptively adjusts the wavelength of Gabor wavelet in the PME method in real time to ensure non-wrapping phase information. The proposed algorithm is systematically evaluated with numerical simulations and experiment in the laboratory. Simulation results demonstrate that it can effectively eliminate phase wrapping, with a Mean Absolute Error (MAE) of less than 0.06 pixels and Root Mean Squares Error (RMSE) of less than 0.08 pixels in the estimated displacement which is similar to that from conventional PME. The computation time is also reduced by 38.5 %. Experiment with a steel beam shows that the displacement extracted is closely matching that from a laser displacement sensor with a correlation coefficient of 0.998 and a MAE and RMSE of 0.17 mm and 0.27 mm, respectively. The computational time has been significantly improved by 65 % compared to that from the traditional PME approach.
Journal
IF:
8.9
Papers:
1.3W
Citations:
6.6W