线结构光测头外参数的精确标定方法

ACCURATE CALIBRATION OF STRUCTURED-LIGHT SENSORS FOR SOLVING THE EXTRINSIC PARAMETERS

为了利用线结构光测头实现三维扫描测量,首先建立了将线结构光测头的二维测量数据转换为三维数据的变换矩阵;通过测量标准球确定球面上两圆弧之间的距离与测量机扫描轴的移动距离之间存在线性关系,使得线结构光测头能测量空间一固定点(标准球球心),从而确定了求解变换矩阵的“共轭对”;最后提出搜索寻优法,将3个旋转角度值在设定范围内以微小步长递增,利用每组具体角度将测量标准球的二维数据转换为三维数据并拟合球,以离散点到拟合球面的距离之和最小为寻优条件,与之相对应的三个旋转角为寻优目标。该方法标定过程简单,便于实际应用,试验结果表明,变换矩阵的求解具有很高的精度。

In order to implement 3D measurement utilizing structured-light sensors, the matrix for transforming the 2D data from the sensor into 3D data is firstly established. By measuring a ball using the sensor, the distance between two circles on the ball and the distance between two positions of the CMM scanning axis is found possessing linear relationship. This enables the sensor to measure a fixed point （ball center）, thus some ＂conjugate pairs＂ for solving the matrix arc identified. Finally, a method for searching optimal solution of the three rotational angles is proposed. With the increment of the three rotational angles in a range, the 2D data picked up on the ball is transformed into 3D data using each set of angles, at the meantime a sphere is fitted using the 3D data. The summation of distance between the 3D data and the fitted sphere surface is calculated, the angles corresponding to the smallest summation arc the optimal solution. The procedure of calibration is simple and facilitative for practical application, the solution is proved to be accurate by experiment.