Chen Yuejian, Xing Zongyi, Li Yifan, Yang Zhi
Appl Opt. 2017 Apr 1;56(10):2714-2723. doi: 10.1364/AO.56.002714.
Laser displacement sensor (LDS)-based online measuring of the wheel diameter has been widely adopted in engineering for advantages such as noncontact, high efficiency, and high precision. For almost all these online measuring systems, calibration is certainly needed in order to obtain the extrinsic parameters of sensors. A field-based easy-to-operate, economical, and efficient calibration method is proposed for an LDS-based wheel diameter online measuring system. Only one standard wheelset is used to build the 3D calibration target that is also the measurement target of the system. The extrinsic parameters for each LDS are obtained through minimizing the residual summation of squares. A multistart framework combining the generation of certain numbers of uniformly distributed starting points and a nonlinear programming solver is adopted to solve the minimizing function to obtain the global optimizer. Factors include the number of standard wheelset placement and sensor noises that will result in calibration error are analyzed. Field experiments are carried out, and the correctness of the calibration method is verified through comparisons with manual caliper-measuring results.
基于激光位移传感器（LDS）的轮径在线测量因其具有非接触、高效率和高精度等优点，在工程领域已被广泛采用。对于几乎所有这些在线测量系统而言，为了获取传感器的外部参数，校准是必不可少的。针对基于LDS的轮径在线测量系统，提出了一种基于现场的、易于操作、经济且高效的校准方法。仅使用一个标准轮对来构建3D校准靶标，该靶标同时也是系统的测量对象。通过最小化平方和残差来获取每个LDS的外部参数。采用结合生成一定数量均匀分布的起始点和非线性规划求解器的多起始框架来求解最小化函数，以获得全局最优解。分析了包括标准轮对放置数量和传感器噪声等会导致校准误差的因素。进行了现场实验，并通过与手动卡尺测量结果的比较验证了校准方法的正确性。
