线结构光多传感器三维测量系统误差校正方法  

Error Correction Method for Three-Dimensional Measurement System With Multi-Sensor and Linear-Structure Light

作　　者：张维光[1,2,3] 赵宏[1,2] 

机构地区：[1]西安交通大学机械制造系统工程国家重点实验室,西安710049 [2]西安交通大学苏州研究院,江苏苏州215021 [3]西安工业大学光电工程学院,西安710032

出　　处：《西安交通大学学报》2011年第6期75-80,共6页Journal of Xi'an Jiaotong University

基　　金：国家科技重大专项基金资助项目(2009ZX04001-061)

摘　　要：针对线结构光多传感器三维轮廓测量系统中多传感器坐标系统一误差及线结构光带特征平面方程求解误差的校正问题,应用一种可进行特征点多分辨率提取的平面靶标,选择两传感器共同测量范围内部分特征点作为参考点,应用迭代求解最近临点算法,求解两标定坐标系精确统一的参数,实现多传感器测量系统中两坐标系统一误差的校正.提出了一种带参数的线结构光带图像特征点亚像素提取算法,通过参数设置改变线结构光带特征平面的位置,对线结构光带特征平面方程求解误差进行校正.实验结果表明,误差校正算法精度高、重复性好,确保测量系统可以获得复杂型面物体高精度的截面测量配准数据.To correct the multi-sensor coordinate system＇s unifying error and the equation resolving error of the feature point plane of linear-structure light strip in multi-sensor 3-D measurement system development process, a plane calibration target with multi-resolution feature points extraction nature is chosen. Some feature points in two sensors common measuring region are selected as the reference points, and the iterative closest point algorithm（ICP） is adopted to resolve unifying parameters of two sensors coordinate system. A sub-pixel feature extraction algorithm with parameter setting is also presented for the image of linear-structure light strip to correct the equation resolving error of the feature point plane. The result shows that the correction methods are of high accuracy and good repeatability to obtain high accuracy measurement registration data for complex-profile objects.

关 键 词：三维测量 误差校正 平面靶标 迭代求解最近临点算法 多传感器 

分 类 号：TH741[机械工程—光学工程]