Computer Vision-Based Structural Displacement Measurement Robust to Light-Induced Image Degradation for In-Service...

Area of Science:

• Civil Engineering
• Computer Vision
• Structural Health Monitoring

Background:

• Structural displacement monitoring is crucial for assessing civil engineering structure safety and serviceability.
• Conventional displacement sensors like linear variable differential transformers (LVDTs) present installation challenges in field testing.
• Computer vision offers a low-cost, non-contact alternative for displacement measurement, but is susceptible to light-induced errors.

Purpose of the Study:

• To develop a robust computer vision-based displacement measurement system for civil engineering structures.
• To enhance the system's resilience to light-induced measurement errors, particularly strong sunlight in field conditions.
• To validate the performance of the proposed system in both laboratory and real-world field experiments.

Main Methods:

• A computer vision approach was employed for non-contact displacement measurement.
• An adaptive region-of-interest (ROI) image-processing algorithm was developed to improve marker tracking accuracy.
• The algorithm was designed to mitigate errors caused by varying and strong sunlight conditions.

Main Results:

• The proposed computer vision system demonstrated reliable marker location determination even under unfavorable lighting.
• Experimental validation in laboratory and field settings confirmed the system's effectiveness.
• The approach showed enhanced robustness against light-induced measurement errors.

Conclusions:

• The developed computer vision system provides a viable and robust solution for measuring structural displacement in field conditions.
• The adaptive ROI algorithm effectively addresses challenges posed by sunlight, improving measurement accuracy.
• This non-contact method offers a practical alternative to traditional sensors for structural health monitoring.