摘要
X光透射法是检验铸造零部件内部的缩孔和缩松的重要方法。目前,主要采用的人工视检手段存在检测效率低和缺乏客观量化统计标准两个问题。鉴于此,文章首先基于YOLOv4模型设计了缺陷检测系统,使系统具备对铸造件X光照片进行缺陷检测的能力;然后,通过提出最小组间间距的概念,基于k-mean算法设计了自适应类别数量的缺陷区域聚类方法,使无序分布的缺陷有序地自组织并成为具备统计意义的缺陷区域,为评估工艺参数提供客观量化指标。数值算例和实验验证了文章方法的有效性。
X-ray transmission method is an important method to inspect shrinkage cavity and porosity in casting parts.At present,the main manual visual inspection methods have two problems:low detection efficiency and lack of objective quantitative statistical standards.In view of this,the article first designed a defect detection system based on YOLOv4 model,so that the system has the ability to detect defects in the X-ray photos of castings.Then,by proposing the concept of minimum group spacing,a defect area clustering method with adaptive number of categories is designed based on k-mean algorithm,so that the randomly distributed defects can be orderly self organized into defect areas with statistical significance,providing objective quantitative indicators for evaluating process parameters.Numerical examples and experiments verify the effectiveness of the proposed method.
作者
王湛昱
黄金泉
许善新
周宗震
沈聪
WANG Zhanyu;HUANG Jinquan;XU Shanxin;ZHOU Zongzhen;SHEN Cong(Suzhou Sanji Casting Equipment Co.,Ltd.,Suzhou,Jiangsu 215106,China;College of Energy and Power Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China;School of Computer Science and Engineering,Changshu Institute of Technology,Changshu,Jiangsu 215006,China)
出处
《计算机应用文摘》
2023年第1期117-119,共3页
Chinese Journal of Computer Application
