鞋楦特征识别与数据光顺

Feature recognition and data smoothing of shoe-last

光顺的数据是鞋楦曲面建模的基础,必须对测量得到的含有噪音成分的原始数据进行识别和光顺处理.本文分析了尖点的分布规律,给出了基于曲率的尖点识别方法,并得到了两种求解曲率的方法.针对尖点识别过程中尖点阈值大小的确定问题,提出了一种自动调整的阈值确定方法,提高了算法的效率,同时有效的避免了尖点的漏判和多判问题.在尖点识别的基础上,对鞋楦数据尖点曲线和截面曲线分别进行了最小二乘和多点求均值的光顺处理,得到符合要求的光顺的鞋楦数据.

Data smoothing is the basis for the surface modeling of shoe-last, and the recognition and smoothing treatment of the initial data that is measured inclu paper analyses the distribution rule of sharp points and recognition metho curvature, at the same time two ways of calculating the curvature are obt determination prob value by autocondi em of threshold value of recognizing, a determinatio it is necessary to ding noise points. d of points base ained. Aimi n method of is put forward, which improves the efficiency of algorithm and avoids estimating mistakes. On the basis of points curve are smoothed using least square and average make This ng at the threshold effectively recognizing ,the points curve and the cross section value of multipoints.