基于机器学习的航天装备金属表面缺陷检测方法综述

打开文本图片集

中图分类号：TP301.6 文献标志码：A

Abstract：Quality control of metal surfaces helps ensure the reliabilityand maintainability of aerospace industry products.Machine learning methods are widely used in automatic detection of metal surface defects and have achieved good results.The current status of metal surface defect detection technology is summarized firstly. Then，metal surface defect detection methods are discussed and compared in terms of traditional algorithms and deep learning algorithms.Among them，traditional defect detection algorithms mainly include clustering，support vector machines，etc；defect detection algorithms based on deep learning mainly include CNN and YOLO series deep convolutional neural networks，etc.Finall，the existing problems in metal surface defect detection are analyzed，and the development of metal surface defect detection methods for aerospace equipment is prospected.

Key words：machine learning;deep learning；aerospace；metal surface defect detection

0 引言

金属材料被广泛应用于航天装备与航天器制造，是构成航天器各个系统与支撑构件的重要部分。（剩余12442字）