【摘要】 测量机直线度误差是测量数据误差的主要组成部分,一般可采用测量标准件的方法进行标定。将最小截平方法引入移动最小二乘法,对标准件测量数据进行拟合处理,采用最小截平方法对移动最小二乘法的局部拟合系数进行估计,在支持域内实现对粗大误差的剔除,构造稳健的回归方法。与传统的移动最小二乘法相比较,该方法能够有效控制数据中粗大误差的影响,准确反映测量数据包含的信息。为了验证该算法的有效性,对坐标测量机标准件测量数据进行处理。实验结果表明,针对包含粗大误差的坐标测量数据,采用改进后的拟合算法,更能真实地反映测量机的直线度误差分布情况。更多还原

【Abstract】 The straightness error of Coordinate Measuring Machine(CMM)is the main resource of the measurement error,and the straightness error of machine is usually calibrated by measuring the standard optical flat. In this paper,the Least Trimmed Square(LTS)method is introduced to the Moving Least Square(MLS)method which is used for fitting of measurement data.The LTS method is used for determining the local fitting coefficients of MLS method and constructing a robust regression method for eliminating the outliers of measurement data. Compared with the conventional MLS method,the proposed method can effectively control the influence of outliers and correctly reflect the information in measurement data. To verify the performance of the proposed method,it is used for dealing with the measurement data of standard optical flat measured by CMM. The result shows that the proposed method more correctly reflects the straightness of CMM which can be used for fitting the measurement data with outliers.更多还原