大口径光学元件面形检测中重叠激光光斑的分离

Segmentation for Overlapped Laser Spot in Large Optical Components Topography Measurement System

在大口径光学元件面形检测系统中,激光光斑的检测及后续处理是基础,直接影响面形检测的精度;针对光学元件面形检测中激光光斑重叠的问题,提出了一种激光光斑重叠分离算法;该算法首先用一种减背景与乘法滤波相结合的方法对光斑图像进行预处理,再用两次阈值法分割图像以获取光斑目标,最后采用距离变换、重叠判别及估算圆心相结合来分离重叠的光斑;实验中,对3种重叠类型不同的光斑进行分离,结果表明,算法能有效地分离重叠的激光光斑;目前,该算法通过长时间的实验验证与改进,已成功运用于实际项目中。

In the large optical components topography measurement system, the accuracy of the topography measurement is based on the measurement and treatment of laser spot. To aim at the problem of overlapped laser spot in optical component topography measurement, a overlapped laser spot segmentation is proposed. First, a method combine background subtraction with multiplicative filtration is presented to preprocess the laser spot image; then, a twice--segmentation method is used to segment the image to acquire laser spot; finally, a method combine distance transform with overlapping detection and estimate the centre of a circle is proposed to segment overlapped laser spot. In the experiment, three different overlapped laser spots are segmented, results demonstrate the segmention of overlapped laser spot is reached efficiently. At present, the algorithm is used in actual project successfully through a long time of verifieafion and improvement.