摘要
采用激光三角测距原理设计并制作了高集成度的激光三角位移传感器测头LDSP—100。第一版机型的软件算法基于曲线拟合中的多项式拟合和高斯拟合,进行光斑定位。由于一体化设计对光斑定位稳定性要求极高,单纯的拟合算法定位无法满足。经进一步优化设计,实现了基于平方加权灰度质心的分段平均算法。实验、对比和分析表明:优化后的基于平方加权灰度质心的分段平均算法,重复性误差达到±0.003 7 mm(100 mm量程),即±0.003 7%F.S.,高于基于灰度质心的分段平均算法60.4%;高于高斯拟合算法22.9%。
Displacement sensors based on laser triangulation ranging principle are widely used because of advantages of small volume and high precision. Highly integrated laser triangulation displacement sensor named LDSP-100 is developed. The algorithm utilizes in the first vision of LDSP-100 is based on polynomial fitting and Gauss fitting method ,light spot localization is carried out ,but stabilization is highly required in integrated design, single fitting method can not meets the needs. An optimized algorithm called segment average algorithm based on square weighted gray centroid is designed and realized. And this is proved to be better than Gauss fitting method. According to a series of experiment and analysis, it is found that the repeatability precision can achieve ±0.0037 mm,which equals to ±0. 003 7 % F. S. with segment average algorithm based on square weighted gray centroid. It is 60.4 % higher than gray centroid segment average algorithm and 22.9 % higher than Gauss fitting algorithm.
出处
《传感器与微系统》
CSCD
2016年第9期62-65,共4页
Transducer and Microsystem Technologies
基金
上海市闵行区科委产学研项目(2014MH116)
关键词
激光三角测量
重复性精度
高斯拟合法
灰度质心
平方加权灰度质心
分段平均
laser triangulation measurement
repeatability precision
Gauss fitting method
grey centroid
square weighted grey centroid
segment average
