TL;DR
This paper introduces a novel visually servoed paired structured light system that overcomes the limited measurement range of previous systems. By incorporating a 2-DOF manipulator, the system actively keeps projected laser beams on target screens, enabling accurate 6-DOF displacement measurement for large structures with an expanded range.
What Problem Does This Paper Address?
Accurately measuring the displacement of large structures, such as bridges and buildings, is crucial for structural health monitoring (SHM). Existing sensors like accelerometers, strain gauges, and GPS have drawbacks: they indirectly measure displacement, are difficult to install, or are expensive to maintain. Vision-based systems offer direct measurement and lower cost, but traditional approaches rely on fixed cameras and targets, making them susceptible to external factors and limited by the line-of-sight. Previous paired structured light (SL) systems could estimate 6-DOF displacement but were constrained by the physical size of their screens, limiting the measurable range. This paper addresses this limitation by proposing a system that actively maintains the laser-target alignment, thereby expanding the measurement range.
Methods
The core innovation is the integration of a visually servoed 2-DOF manipulator with a paired structured light system. Each module of the system consists of two screens facing each other, each equipped with lasers and a camera. The manipulator, controlled by visual feedback from the cameras, actively adjusts the pose of the lasers to ensure the projected beams consistently hit their respective target screens. This visual servoing mechanism is key to overcoming the range limitations of static systems.
The process involves:
- Image Capture: Cameras on each side capture images of the screens.
- Calibration: Lens distortion is corrected, and homography is calculated to determine screen boundaries.
- Laser Point Detection: The positions of the projected laser beams on the screens are identified.
- Manipulator Control: The system calculates the required rotation angles for the 2-DOF manipulators to minimize the difference between the desired and current laser beam positions.
- 6-DOF Displacement Estimation: The 6-DOF displacement is then calculated using the manipulator's rotation angles and the detected laser beam positions.
- Initial Offset Calibration: The estimated displacement is calibrated against an initial offset to provide relative measurements.
The paper also details the kinematic relationships and transformation matrices used to derive the system's equations, employing a Newton-Raphson method for displacement estimation.
Datasets
Not discussed in the paper. The experiments were conducted using a custom-built prototype.
Key Insights and Ideas
- Active Range Extension: The most significant insight is that by actively controlling the laser projection using visual servoing and a manipulator, the system can overcome the inherent range limitations imposed by fixed screen sizes. This is a crucial step towards practical, large-scale structural monitoring.
- Robustness to External Changes: While not explicitly quantified, the paper implies that by keeping the laser beams on the screen, the system becomes less susceptible to external environmental changes (like weather or illumination) that could disrupt line-of-sight in passive vision systems.
- Scalability for Massive Structures: The ability to cascade multiple modules suggests a clear path for applying this technology to very large structures like long-span bridges and high-rise buildings, where traditional sensors are impractical.
Key Results
The experimental results demonstrate that the proposed visually servoed paired SL system can accurately estimate 6-DOF displacement. The paper shows that the system achieves high accuracy and a significantly expanded measurement range compared to previous static paired SL systems. Figures 6 and 7 present experimental data showing the estimated translational and rotational displacements closely matching the ground truth, indicating the system's effectiveness.
Limitations and Future Work
Limitations: * The paper acknowledges that the measurable range of the previous paired SL system was limited by screen size. While the new system expands this range, the exact quantitative expansion is not detailed. * The current prototype uses a 2-DOF manipulator. For more complex movements or larger structures, additional degrees of freedom might be necessary.
Future Work: * The authors plan to build multiple modules and test the system's applicability to real large structures, such as long-span bridges and high-rise buildings. * Further investigation into optimizing the manipulator control and the overall system for even greater accuracy and range is implied.
Summary updated