基于单视图三维重建的凹凸制造特征识别

Concave-convex Manufacturing Features Recognition Based on 3D Reconstruction of Single View

为实现凹凸制造特征机器人的自动识别,文中提出了一种不依赖于CAD设计模型的自动特征识别新方法。该方法以零件的单幅图像为识别线索,首先采用改进的SFS算法对零件表面进行三维曲面重建;然后对重建模型表面的形状指数进行分析以计算特征分割线,利用特征线将曲面进行分割以获得相应的特征区域;最后基于特征识别规则实现对零件凹凸制造特征的有效识别。该方法能够在缺少CAD模型时有效地实现制造特征的自动识别,从而为来料加工以及二次装配过程中机器人的自动特征识别提供重要的方法。通过实例零件验证了该方法的有效性和准确性。

This paper proposed a new method of automatic feature recognition without relying on the CAD design model to achieve robot automatic recognition for concave-convex manufacturing features.The method takes the single image of the part as the identification clue.Firstly,the surface of the part is reconstructed by using the improved method of shape from shading.Then,the features segmentation lines which are used to segment the surface are calculated by analyzing the surface shape indexes of the reconstructed model.The surface is segmented by feature lines to obtain the corresponding classification region.Finally,based on the feature recognition rules,the feature of concave and convex manufacturing is identified effectively.This algorithm can effectively solve the problem of automatic identification of manufacturing features in the absence of CAD model,providing important method for the automatic feature recognition of the robot in the processing of incoming materials and two assembly.The validity and accuracy of the proposed method were verified by an example.