三维边缘轮廓线数据的识别

Recognition of Points on Three-Dimensional Edge Contour Lines

三维成像技术可快速获得被测表面的海量点云数据,如何从点云数据中识别出准确的边缘轮廓线数据是三维边缘轮廓线检测的首要工作。分析边缘轮廓线数据的位置特征,说明在边缘轮廓线的垂直截面线上识别边缘轮廓线数据点的重要意义;由于实际垂直截面线数据在边缘轮廓线数据点处存在倒角圆弧的情况,需通过倒角圆弧中间点求解出边缘轮廓线数据的位置;根据倒角圆弧数据的不同分布特征,提出3种倒角圆弧中间点的识别方法,即拟合抛物线求解顶点法、法向累积夹角半交角判别法、法向夹角引导的弧长均值插值法。在多条边缘轮廓线上验证了3种倒角圆弧中间点识别方法的有效性和精度。

Three-dimensional(3D)imaging technology obtains massive points(MPs)of a surface in a very short time.In this study,we recognize points on 3D edge contour lines(ECLs)from MPs of a surface.First,the position features of the points on the ECLs,which are called edge contour points(ECPs),are analyzed,and it is proposed to detect ECPs from the vertical section lines(VSLs)of the ECLs.Then,the position of an ECP needs to be calculated based on the middle point of the chamfered arc(MPCA)on a VSL.Considering the distribution character of points on the CA(PsCA),three methods are proposed here to recognize the MPCA:parabola fitting for its vertex method,calculating the accumulated normal angle(AcNA)of each PsCA for the point whose AcNA is closest to the half central angle of the CA,and calculating the approximate arc length(AAL)of each PsCA for the point whose AAL is closest to the half arc length of the CA.Finally,the effectiveness and accuracy of the three methods are verified on several ECLs.