3D打印中数据精简处理技术

Data reduction processing technology in 3D printing

目前流行的3D打印方法是把模型数据切割成切片数据进行打印,但是切割后的轮廓上存在很多冗余数据点。针对该问题,文中提出一种在角度弦高法的基础上融合二分法和筛选法的简化处理算法。该方法将轮廓上的数据分割成一个一个的簇数据集,再根据簇内每个点的偏角度和弦高依次进行判断筛选;然后通过二分迭代的方法对簇数据进行重新聚类,以此来确定需要去除哪些数据点。选取1001个原始数据点的梅花形状轮廓为实验对象,当设定偏角阈值Δα=10°,弦高度阈值为Δh=0.1 mm时,所提算法精简率为80.82%,且用时仅为0.0049 s,重构曲线的平均误差仅为0.0237 mm。所提算法精简的效果比较理想,不仅能够在轮廓的曲率变化较高的地方保留足够的数据点,而且在轮廓的曲率变化较低处也能把细节的特征点很好地保留下来,保留的数据能够跟随曲率特征实现合理的分布。实测结果表明,要使精简的效果良好且快速稳定,需要把每次处理的数据簇的个数设定在10~20。

The current popular 3D printing method is to cut the model data into slice data for printing,but there are many redundant data points on the cut contour.Address this problem,a simplified processing algorithm that integrates binary and screening methods is proposed on the basis of angle chord-height method.In this method,the data on the outline is divided into a cluster data set one by one,and then judge and filter according to the offset angle and chord height of each point in the cluster.The cluster data is reclustered by means of the binary iteration method to determine which data points need to be removed.The plum blossom shaped contour of 1001 original data points is selected as the experimental object,when the deviation threshold is set to 10°and the chord height threshold is set to 0.1 mm,the proposed algorithm can achieve a reduction rate of 80.82%,and the time is only 0.0049 s.The average error of the reconstructed curve is only 0.0237 mm.The simplification effect of the proposed algorithm is relatively ideal,as it can not only retain sufficient data points in areas with high curvature changes of the contour,but also retain detailed feature points in areas with low curvature changes of the contour.The retained data can follow the curvature features to achieve a reasonable distribution.The actual measurement results indicate that it is necessary to set the number of data clusters processed each time to 10~20,so as to achieve good and fast simplification.