Deformation Reconstruction of Injection Molded Part Based on Stratified ICP Registration and B-spline Fitting

Abstract

Injection molding is a multi-field interacted process with high pressure, high speed and high temperature, which leads to that the deformation and its microcosmic distribution of molded part are intractable to measure and describe accurately. In this paper, a deformation reconstruction method based on the stratified Iterative Closed Point (ICP) registration and B-spline fitting was proposed. In the method, an adaptive smoothing filtering algorithm was presented to process the measured point cloud, and the registration considering the impact of the deformation was applied to align the measured to the ideal point clouds to get accurate deformation, and the cubic B-spline surface interpolation algorithm was utilized for the deformation reconstruction. The experiment results showed that the developed adaptive filtering could better smooth the data than the traditional mean, median, and Gaussian filtering algorithms, and the error of the stratified ICP was about 13% less than that of the traditional ICP and the transformation reliabilities were improved, furthermore, the deformation values, corresponding positions, and microcosmic distribution of deformation were completely supplemented.