三种用于加工特征识别的神经网络方法综述

加工特征自动识别技术是智能化设计与制造的关键支撑,已有的实用性算法普遍存在学习能力差、识别范围有限和识别速度慢等共性问题。神经网络方法在计算机视觉和模式识别领域获得了巨大成功,其自学习与自适应能力和高速计算等优势也已在加工特征识别中得到初步的展现。对加工特征识别中具有应用潜力的三种不同的神经网络方法进行了研究,剖析了神经网络识别加工特征中的预处理与编码和神经网络结构设计等关键性问题,分析了不同神经网络方法的异同点,总结了当前神经网络识别加工特征的发展方向,为相关领域的研究提供一定的理论指导与技术支持。

A review of three neural network methods for manufacturing feature recognition

Automatic manufacturing feature recognition is a crucial support for intelligent design and its manufacturing. However, existing practical algorithms expose common problems such as poor learning ability, limited recognition range, and slow recognition speed. The neural network method has achieved great success in the fields of computer vision and pattern recognition. At the same time, its self-learning and self-adaptive capabilities and high-speed computing have also been preliminarily demonstrated in manufacturing feature recognition. Three different neural network methods with potential applications in manufacturing feature recognition are studied in this paper. The key issues of feature preprocessing and coding and architecture design of neural networks in manufacturing feature recognition based on neural networks are analyzed. Meanwhile, the similarities and differences of different neural network methods are analyzed. In addition, the development direction of current feature recognition approach using neural networks is summarized. This overview can provide some theoretical guidance and technical support for the research in related fields.