基于模糊模式识别的铝护套切割装置缺陷自动检测方法研究  

Research on Automatic Defect Detection Method of Aluminum Sheathed Cutting Device Based on Fuzzy Pattern Recognition

作　　者：刘毅川 陈明坚 欧敏飞 LIU Yichuan;CHEN Mingjian;OU Minfei(Guangzhou Southern Investment Group Co. ,Ltd. ,Guangzhou 510260,China;Guangzhou Nanfang Electric Power Technology Engineering Co. ,Ltd. ,Guangzhou 510600,China)

机构地区：[1]广州南方投资集团有限公司,广东广州510260 [2]广州南方电力技术工程有限公司,广东广州510600

出　　处：《自动化仪表》2021年第11期44-48,52,共6页Process Automation Instrumentation

摘　　要：针对当前铝护套切割装置缺陷检测方法存在铝护套切割装置缺陷检测效果较差、检测时间较长的问题,提出基于模糊模式识别的铝护套切割装置缺陷自动检测方法。采用模糊聚类方法,对铝护套切割装置缺陷特征进行聚类;根据模糊相似关系,获取铝护套切割装置缺陷特征数据作为检测数据。利用模糊神经网络,检测铝护套切割装置缺陷类型,对缺陷特征数据进行模糊标准化,提高检测效率。根据最大隶属度原则,对缺陷特征类型进行判断,精准识别缺陷特征类型,实现铝护套切割装置缺陷自动检测。试验结果表明,所提方法能够有效提高不同铝护套切割装置缺陷类型的查全率、正确率和F值,并能准确检测缺陷类型、缩短缺陷检测时间。Aiming at the problems of poor detection effect and long detection time in the current defect detection methods of aluminum sheathed cutting device,an automatic defect detection method of aluminum sheathed cutting device based on fuzzy pattern recognition is proposed.The fuzzy clustering method is used to cluster the defect characteristics of aluminum sheathed cutting device.According to the fuzzy similarity relation,the defect characteristic data of aluminum sheathed cutting device is obtained as the test data.The defect types of aluminum sheathed cutting device are detected by using fuzzy neural network,and the defect characteristic data are fuzzy standardized to improve the detection efficiency.According to the principle of maximum membership degree,the defect feature types are judged to accurately identify the defect feature types,and automatic detection of the aluminum sheathed cutting device is realized.The experimental results show that the proposed method can effectively improve the recall rate,accuracy rate and F value of the defect types of different aluminum sheathed cutting devices,accurately detect the defect types and shorten the defect detection time.

关 键 词：模糊模式识别 铝护套 切割装置 缺陷检测 模糊聚类 模糊神经网络 最大隶属度原则 模糊相似关系 

分 类 号：TH136[机械工程—机械制造及自动化]