IP protection

1 Please provide a short description of the state-of-the-art and/or current trends in the field? How does the result fit into it?
           The reliable detection and evaluation of welding defects is  important part of the nondestructive welding control techniques. A significant number of algorithms have already been developed and used (Kehoe and Parker, 1990; Cherfa et al., 1998; Liao and Li, 1999; Lashkia, 2001; Herold et al., 2002; Wang and Liao, 2002; Kumar and Pang, 2002; Alghalandis and Alamdari, 2006; Eisele, 2002). The most renown ones are described in the brief below:

  • A method for automatic recognition of welding defects, which is performed in two steps: in the first one, the defects are searched for in the high-frequency areas of the image. For this purpose, the processed image is divided in sub-blocks of 32 x 32 pixels; then, for each sub-block the discrete Fourier spectrum is calculated. In the second step, the likely defects are compared consecutively with a set of typical defects, stored in a pattern library;
  • A method for intelligent segmentation of industrial radiographic images, in which processing starts with image segmentation based on adaptive threshold and the obtained binary image is then processed with neural networks back-propagating any error;
  • A background subtraction method for weld inspection, in accordance with which every column of the digital image is approximated with a polynomial function, which is subtracted from the brightness of the original one. The resulting image is processed with fuzzy k-nearest neighbor, multi-layer perceptron neural networks and the classification of the welding defect type is performed by a fuzzy expert system. The main features used for the classification are: distance from the center, circularities, compactness, major axis, width and length, elongation, diameter, average intensity and standard intensity deviation, etc.;
  • A method for radiographic recognition of welding defects with linear classifiers, which comprises the following steps: in the first one, the quality of the X-ray image is improved using a median filter followed by contrast enhancement; in the second step (the defects detection) the following is performed: 1) Segmentation of the likely defects; 2) Extraction of their geometric and gray value features (contrast, position, aspect ratio, width-area ratio, length-area ratio, and roundness). The correlation between the so extracted features and each of the defect classes (slag inclusion, porosity, lack of penetration, and undercutting) is evaluated analyzing the linear correlation coefficient; 3) The most relevant features are used as input data for a hierarchical linear classifier;
  • A method for defects recognition, using shape features. The detection comprises the following steps: 1) Segmentation: regions of pixels are fixed and isolated from the rest of the X-ray image using a watershed algorithm and morphological operations (erosion and dilation); 2) Feature extraction: the regions are measured and shape characteristics (diameter variation and main direction of inertia based on invariant moments) are quantified; 3) Classification: the extracted features of each region are analyzed and classified using a k-nearest neighbor classifier;
  • A method for defects recognition, using texture features based on co-occurrence matrix and 2D Gabor functions, i.e., Gauss-shaped band pass filters with dyadic treatment of the radial spatial frequency range and multiple orientations. The most relevant features are used as input data on a statistical classifier;
  • A method for defects detection using image data fusion, based on edge extraction, wave profile analyses, segmentation with dynamic threshold and weld district extraction.

The methods for welding defects detection are getting more and more sophisticated and intelligent in correspondence with new requirements, which force the fast development of new technologies and systems for efficient automatic control.

   Expert systems used in practice are usually not presented to wide public. The system developed by Padua et al., is using preprocessing and detects various kinds of discontinuities, for which was used the Savitzky-Golay filter and fast Fourier transform. Empirically are set the optimal parameters of the filter. Another system is presented by Ho, White and Lukas, is based on the use of the Sobel operator. For detection of cracks (Kabir and Rivard) is used grey level co-occurrence matrix and discrete wavelet transform.

 

References:

Kehoe, A. and Parker, G. (1990) ‘Image processing for industrial radiograph IC inspection: Image enhancement’. British Journal for NDT, Vol. 32 (4), pp. 183-190.

Cherfa, Y. Kabir, Y. and Drai, R. (1998) ‘X-Rays Image Segmentation for NDT of Welding defects’. 7th ECNDT’98, Copenhagen, Denmark.

Liao, W. and Li, Y. (1998) ‘An Automated Radiographic NDT System for Weld Inspection: Part II – Flaw detection’. NDT&E International, Vol. 31 (3), pp. 183-192.

Lashkia, V. (2001) ‘Defect detection in X-ray images using fuzzy reasoning’. Image and Vision Computing, Vol. 19, pp. 261- 269.

Herold, F. Bavendiek, K. and Grigat, R. (2002) ‘A new analysis and classification method for automatic defect recognition in X-ray images of castings’. In 8th European Conference on Non-Destructive Testing, Barcelona.

Wang, G. and Liao, W. (2002) ‘Automatic Identification of different types of welding defects in radiographic images’. NDT&E International, Vol. 35, pp. 519-528.

Kumar, A. and Pang, G. (2002) ‘Defect Detection in Textured Materials Using Gabor Filters’. IEEE Transactions on Industry Applications, Vol. 38 (2), pp. 425-440.

Alghalandis, S. and Alamdari, G. (2006) ‘Welding Defect Pattern Recognition in Radiographic Images of Gas Pipelines using Adaptive Feature Extraction Method and Neural Network Classifier’. 23rd World Gas Conference, 13 pages, Amsterdam

Eisele, H. (2002) ‘Automated defect detection and evaluation in X-ray CT images’. Ph. D. Thesis, Stuttgart, Germany, 62 p.

Ho, White and Lukas “A vision system for automated crack detection in welds”, Meas. Sci. Technol. 1, 287-294 (1990)

G. Padua, R. da Silva, D. Mery, M. Siqueira, J. Rebello, L. Caloba. “Detection and classification of weld discontinuities in radiographic images, Materials Evaluation, 1139-1145 (2007)

S. Kabir and P. Rivard. “Damage classification of concrete structures based on grey level co-occurrence matrix using Haar’s discrete wavelet transform.

2 What is the problem/need/knowledge gap that the research result is responding to?  How was it addressed before?
The methods, mentioned above have relatively high complexity. Most of them are based on intelligent algorithms, which require much time and computational power and need setting the values of large number of parameters.In this R&D result, a new approach is presented for automatic defects detection in digital X-ray or photo images of welded joints. The method is based on the following sequence of basic operations:

  • Preprocessing of the digital image, which comprises adaptive noise filtration and correction of the uneven background illumination;
  • Image segmentation used for the extraction of the likely welding defects;
  • Individual analysis of the detected likely defects, used to determine their geometrical and brightness parameters;
  • Classification of the detected defects on the basis of the extracted features.

The basic advantage of the new approach is the lower computational complexity.

Additional advantages are the high accuracy of the defects boundaries and the limited number of parameters, whose values should be changed depending on the sensor characteristics.

3 What is the potential for further research?
The main advantages of the new method are:- the relatively low computational complexity;- the limited number of parameters, whose values should be set depending on the sensor;

- the high accuracy of the boundaries of the detected defects.

4 What is the potential of the research result for synergy with other research areas either in the same or in a different discipline?
The same method could be used for efficient archiving of old handwritten documents. For this, text segmentation is first performed. After that, the image background is compressed with some kind of lossy compression, and the extracted text – with llossless. In result is obtained very high compression ratio, with retained visual quality of the restored document. 
5 What is the proposed method of IPR-protection? (patent, license, trademark etc.)
All kinds of IPR-protection are possible.
6 What are the steps that need to be taken in order to secure the IPR-protection? What is the cost of IPR-protection?
Regular costs and activities in EU.
7 What is the expected impact of the research result? (industry, society, administration etc. and target groups of beneficiaries)
The impact of the research result is very high for industry in the area of manufacturing of tubes and other similar parts, needed for gas transfer. Same method could be used for analysis of bridges or concrete (cracks).The additional application area (for archiving of old handwritten documents) has high social impact in the area of the cultural heritage protection and access.
8 What is you overall assessment of the scientific maturity of the research result?
The product is technically feasible. It offers very high efficiency when the method parameters are adapted to a pre-defined sensor and could be used as a basis for the creation of intelligent expert systems for nondestructive control.

KEYWORDS QUANTITATIVE ASSESSMENT (0-5).

  1 2 3 4 5
Scientific maturity         x
Synergies       x  
State-of-the-art/innovation       x  
IPR-potential         x

Bookmark the permalink. Follow any comments here with the RSS feed for this post. Post a comment or leave a trackback: Trackback URL.

Post a Comment

You must be logged in to post a comment.

Visit the other applications of the INTERVALUE Platform: R&D Repository | IP Agreements

© 2009-2010 INTERVALUE Project