3维数据的置信区间及异常数据的修复

3维测量是逆向工程设计的数据来源，为此提出了用弦矢量法确定3维数据的置信区间，并对异常数据进行修复。利用3维曲面上某点相邻的点，求出该点的两个互相垂直的弦矢量，由此，曲面上的一个矩形区域的数据就构成了两个弦矢量阵列。对不等间距的弦矢量阵列进行了校正。对两个等效弦矢量阵列进行概率统计分析，即可确定该测量数据的置信区间。根据该置信区间对异常数据进行了修正。最后对两个目标实物的测量数据进行了分析，结果表明该方法对异常数据的修复是非常有效的。

Determination of confidence interval of 3-D data and repair of abnormal data

A Chord Vector Method (CVM) was proposed to determine the Confidence Interval (CI) of the 3-D data and repair abnormal data. Using the contiguous points of the chosen point in the 3-D geometric curve surface, the two Chord Vectors (CV), perpendicular to each other were obtained, meanwhile the two Chord Vector Arrays (CVAs) were obtained by selecting a rectangle area of the curve surface. The unequal intervals CVAs were calibrated. Then, the CI of the measured data was determined by performing the statistic analysis for the equivalent CVAs. The repair of abnormal data was performed according to the CI. Finally, after the analysis of measured data of two objects, the effectiveness of the method for repair of abnormal data had been verified.