Font Size: a A A

The Application Of Wavelet Analysis For The Welding Line Flaw Detection

Posted on:2005-02-07Degree:MasterType:Thesis
Country:ChinaCandidate:L LuoFull Text:PDF
GTID:2168360125951050Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
In this paper, the X-ray digital image of welding lines in natural gas pipelines is studied. According to the current developing status of technology to detect flaws in welding line, the wavelet is chosen as the leading analysis tools, and three methods based on wavelets' advantages of multi-scale and singularity are presented.In the first method, the flaws in gas welding lines distribute in a round line. The two dimension image is first translated into one dimension signal on the precondition of not losing flaws' information. Then it is analyzed respectively by three wavelets commonly utilized in image processing, the optimal wavelet which is best available for detecting this kind of flaw is found out, and at last the method of modulus maxima is used to finish flaw's detection. The results show that this method is effective and with fine real-time property. It is fit for occasions that the shape and type of flaws are unknown.In the second method, firstly, the image is de-noised by the two dimension wavelet and also these methods of de-noising are compared, then the two dimension image is decomposed and analyzed by wavelet and its decomposing coefficients and energy percent are gotten. The conclusion is that the coefficients distribute more in vertical and horizontal direction than it does in diagonal direction. At last the Laplacian operator is utilized to detect flaw's edge. By comparing with edge detection in vertical, horizontal, 45 and 135 direction, the results show this method is effective in recognizing flaws and takes short calculation time.In the third method, firstly the multi-scale edge detection is done by two-dimension wavelet, and then compare with the results of utilizing traditional edge detection operators. The simulation results show that this method is effective for exiguous flaw's detection. Its disadvantages are complicated calculation and bad real-time property.At last, a simple human-machine interface is developed by a dynamic link between VB and MATLAB. Its reliability and practicability are testified by simulation experiments. The above-mentioned three algorithms can be directly used in the current pipeline welding lines' fully-automatic film interpretation system through this HMI without transplanting algorithms' programs.
Keywords/Search Tags:welding line flaw, edge detection, wavelet transform, one-dimension wavelet, two dimension wavelet, mother wavelet, simulation, VB, MATLAB
PDF Full Text Request
Related items