摘要
以埋弧焊管焊缝的X射线检测图像为对象,通过图像处理、特征提取和模糊识别实现了对缺陷的识别。为提高识别精度与实时性,采用主成分分析法对采集图像的像素矩阵进行了主元分析,结合模糊识别中的模糊C均值聚类算法对圆形缺陷和线形缺陷进行识别。相比于传统的通过提取缺陷的若干几何特征分类识别的方法,此方法具有算法简单、占用内存空间小、识别准确率高、实时性强等特点。最终平均识别率可达到90.93%,能够较准确地对焊缝缺陷进行分类识别。
The recognition of the defects in the welding seam of submerged-arc welded pipe is finished through image processing, feature extraction and fuzzy recognition. In order to improve the accuracy and real-time performance of defect recognition, the principal component analysis （PCA） of the pixel matrix of the acquired image is carried out, and the circular defects and linear defects in the submerged-arc welding seam are recognized using fuzzy C-means clustering （FCM） algorithm. Compared with the traditional classifica- tion identification method based on the extracted geometric features of the defects, this algorithm is simpler, its occupied memory space is smaller, its recognition accuracy is higher and its real-time performance is stronger. The average defect recognition ratio of the algo- rithm can reach to 90.93% ,and the accurate classification identification of the defects in the submerged-arc welding seam can be finished using this method.
出处
《西安石油大学学报（自然科学版）》
CAS
北大核心
2016年第4期115-121,共7页
Journal of Xi’an Shiyou University（Natural Science Edition）
基金
陕西省自然科学资助项目(编号:2013JQ8049)
陕西省教育厅重点实验室科研计划项目(编号:14JS079)
陕西省教育厅自然科学专项(编号:2013JK1077)
陕西省自然科学基础研究计划青年人才项目(编号:2015JQ5129)
关键词
缺陷识别
焊缝缺陷
X射线检测
主成分分析法
模糊C均值聚类
像素矩阵
defect recognition
welding seam defect
X-ray detection
principal component analysis
fuzzy C-means clustering
pixel matrix
