研究目的
To propose a new static calibration method for low-cost laser triangulation systems that does not require precise movement or high precision gauge objects, aiming to reduce manufacturing costs and enable widespread applications.
研究成果
The proposed static calibration method for low-cost laser triangulation systems avoids the need for precise movements of both camera and reference object, significantly reducing costs. Laboratory prototypes validated the method, achieving sub-millimeter precision. The method supports the design of multi-point forming systems and has potential for widespread applications in sheet-metal forming and panel manufacturing.
研究不足
The precision of the final result depends on camera resolution, laser spot size, mechanical errors, and system components and assembly. The method's precision is currently sufficient for manufacturing large panels but could be improved with additional work.
1:Experimental Design and Method Selection:
The study proposes a static calibration method for laser triangulation systems, utilizing a geometric estimator as a quality indicator. The method involves preparing data on a scalar compound target without the need for motion or high precision gauge objects.
2:Sample Selection and Data Sources:
A piece of paper with grids of known dimensions is used to provide readings for points from the xOy plane. The grid paper is aligned with the camera to ensure the y0 axis overlays on the x axis.
3:List of Experimental Equipment and Materials:
Prototypes built include a webcam (Logitech C310) and a CCD camera (SONY XCG-C130), laser pointers, and laser LED diodes. A rotating structure for angular adjustment and a marked base plate for positioning the grid paper are also used.
4:Experimental Procedures and Operational Workflow:
The calibration process involves four steps: preparing the grid paper, finding xd with the camera's help, finding several points on the xOy plane, and conducting the calibration method to find xc, zc, d.
5:Data Analysis Methods:
Nonlinear regression is conducted to find hidden variables xc, zc, d, with the Levenberg-Marquardt method suggested for fitting.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容