视觉测量数据的灰色挖掘方法及其精度评定

Grey Mining Method of Vision Measurement Data and Its Precision Evaluation

在视觉测量的过程中,将由工程样机得到的有限数据进行灰色挖掘并且与曲线拟合相结合,充分地发挥2种不同方法各自的优势,以寻找位置精度与测量距离之间的关系.首先,利用少量的实验数据建立灰色预测模型,通过该模型对有限的测量数据进行挖掘以增加样本量.然后,利用灰色关联方法分析世界坐标系3个坐标轴上的位置精度与测量数据之间的关联性.最后,对灰色预测模型得到的数据进行多元、多阶的多项式拟合.实验结果表明:该方法预测数据的精度可以达到1级,并且位置精度与测量距离之间拟合的准确度优于99%.

In the process of vision measurement,the limited data obtained by engineering prototype is applied to grey data mining. The grey data mining method and curve fitting are combined to realize the advantages of the two methods and then to find out the relationship between the position error and the measurement distance. First,a grey forecasting model,which can be applied to data mining,is built to enlarge the sample size of the limited data. Second,the grey correlation analysis is used to deal with the correlations of position measurement error. Finally,with the results of the grey forecasting model,a polynomial fitting of multivariate and multi-order is done. Experimental results show the data mining precision has reached first degree and the precision of curve fitting between the position error and the measurement distance is superior to 99%.