| 叶宏武.机械零件图像表面瑕疵的检测算法[J].轻工机械,2015,33(2): |
| 机械零件图像表面瑕疵的检测算法 |
| Detection Algorithm of Parts Surface Image Defect |
| |
| DOI：IO. 3969/j. issn. 1005 -2895. 2015. 02. 011 |
| 中文关键词: 机械零件 零件表面瑕疵检测 机器视觉 自动检测 缺陷识别 |
| 英文关键词:mechanical parts parts surface image defect detection machine vision automatic detection defect recognition |
| 基金项目:国家自然基金( 51303157)；浙江教育厅科研项目(Y201431078)；宁波市自然基金项目(2013A610044) |
|
| 摘要点击次数: 1859 |
| 全文下载次数: 1334 |
| 中文摘要: |
| 针对机械零件表面瑕疵检测问题，将机器视觉技术用于零件表面图像瑕疵的提取和分析，提出一种基于粒子群
优化算法加权模糊C均值聚类的零件缺陷图像智能分割算法，精确定位了机械零件表面的瑕疵区域。缺陷的形状特征
是判断其类型的重要依据，提取缺陷的形状特征，设计支持向量机分类器来检测划痕、裂纹、砂眼等表面瑕疵。研究结果
表明，该方法具有较强的实用性，在实验数据库上达到90%以上的正确识别率。 |
| 英文摘要: |
| According to the mechanical parts' surface defects problem, the machine vision technology was used to extract
and analyze the surface defect of the mechanical parts image. A fuzzy C means clustering weighted defect edge detection
algorithm was proposed based on particle swarm optimization. It could locate mechanical parts surface defect precise
region accurately. Shape characteristics of defects were important criterion for its type. The defect shape feature was
extracted and support vector machine classifier was designed to detect surface defects such as scratches. cracks and
trachoma. The research results show that this method has strong practicality and the correct recognition rate is above
90% in the experimental database. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
