Font Size: a A A

Welding Defects Recognition Algorithm Reasearching Based On Principal Component Analysis

Posted on:2018-01-25Degree:MasterType:Thesis
Country:ChinaCandidate:J ZhangFull Text:PDF
GTID:2348330515988717Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Weld defect recognition is a key technology in the process of welding quality detection.Detecting the weld defect effectively and guaranteeing the quality of weld are of great significance for oilfield safely production.In the aspect of defect detection,manual assessment has a big subjectivity,reatly influenced by testing personnel's professional quality,low level of automation.With the rapid development of computer technology and electronic technology,the technology of computer aid assessment based on image processing has become possible.And more misjudgments are easy to be produced.The computer Assistant Assess System has greatly reduced the workload of the worker,and improver the working efficiency is some extent,thus,the process of the film assessment is more scientific and more standard.But when we put the welding line image into computer,it has much noise,the edge of defect fuzzy and lack contrast and so on,so these weak points make drawing the defect information,segmentation and recognition much difficulty.In this research,the X-ray images of submerged-arc welded pipe are taken as the object.To identify the welding defects,such as: air hole and crack,through image processing,feature extraction and neural network recognition is used to identify defect successfully.First,image filter,image enhancement,image segmentation and edge detection arithmetic is used to find the welding seam boundary in the entire image.Second,by analyzing the distribution of gray curve,the positions of the weld defects were located at first.Third,build up the feature vector by calculating 7 classes geometric and shape feature of welding defects,use the Principal Component Analysis and Kernel Principal Component Analysis method preprocess this vector.Last,the recognition between air hole and crack is studied based on the Neural Network,of which the Original data and processed vector are taken as the input parameter.The experiment result indicated the dataset had been carried out feature selection could be mapped to a lower feature space by using KPCA and PCA for feature extraction.As a result,the performance of the classifier was improved.Simulation result,we could find KPCA less than PCA 1 in feature extraction,KPCA was more effective than PCA;we could find RBF neural network BP neural network more than 0.72% in defects recognition,RBF neural network was more effective than BP neural network.
Keywords/Search Tags:Welding defects, Image process, Principal Component Analysis, Kernel Principal Component Analysis, Neural Networks
PDF Full Text Request
Related items