光点定位中的曲面拟合迭代算法

Iteration algorithm of surface fitting in the detection of light-spot position

在光电检测系统中,需要对光点信号的位置实现亚像素定位。光纤从一端照明,由光纤另一端出射光点,由面阵CCD摄像机获取光点的灰度分布。基于最小二乘法,采用高斯曲面拟合算法拟合其分布,用高斯拟合函数的极值点作为光点的定位点。实验结果表明,采用多次迭代算法可以提高定位精度;拟合算法的稳定性(标准差)为0.7μm.高斯曲面拟合算法具有重复精度高、稳定性好的特点。

Many photoelectric detection systems need to get the position of a light-spot with sub-pixel accuracy. The light source illuminates the optical fiber from one end. The optical fiber emits a light spot from the other end. The gray distribution of the light spot is acquired by Matrix CCD camera. The algorithm of Gaussian distribution fitting to fit its gray distribution based on least square method is presented. The extreme point of the Gaussian fitting function represents the position of the light spot. Experimental data show that the iteration algorithm can improve the position precision. The stability of the fitting algorithm (standard deviation) is 0.7μm. Experimental results indicate that the algorithm has the virtue of high repeated accuracy and stability.