Limits...
Unsupervised classification of surface defects in wire rod production obtained by eddy current sensors.

Saludes-Rodil S, Baeyens E, Rodríguez-Juan CP - Sensors (Basel) (2015)

Bottom Line: An unsupervised approach to classify surface defects in wire rod manufacturing is developed in this paper.The defects are extracted from an eddy current signal and classified using a clustering technique that uses the dynamic time warping distance as the dissimilarity measure.It is shown that it outperforms other classification alternatives, such as the modified Fourier descriptors.

View Article: PubMed Central - PubMed

Affiliation: Centro Tecnológico CARTIF, Parque Tecnológico de Boecillo 205, 47151 Boecillo, Valladolid, Spain. sersal@cartif.es.

ABSTRACT
An unsupervised approach to classify surface defects in wire rod manufacturing is developed in this paper. The defects are extracted from an eddy current signal and classified using a clustering technique that uses the dynamic time warping distance as the dissimilarity measure. The new approach has been successfully tested using industrial data. It is shown that it outperforms other classification alternatives, such as the modified Fourier descriptors.

No MeSH data available.


Macro photography of individual defects belonging to each class.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4482004&req=5

f3-sensors-15-10100: Macro photography of individual defects belonging to each class.

Mentions: The operators of a manufacturing plant of wire rods identified and labeled the surface defects obtained for several production shifts. This has been a very tedious and time-consuming task, because it required unwinding long wire rod coils, searching the surface defects by visual inspection and classifying and putting them in correspondence with the signal recorded by the eddy current inspection system. After this manual process, a collection of labeled defects is available for validation of the developed unsupervised classification method. The surface defects have been classified by the experts into four different classes. The corresponding eddy current signals associated with them have been represented in the complex impedance plane and labeled as defects belonging to Classes A, B, C and D, respectively. An individual sequence representing each of these groups is depicted in Figure 2. The length of the available labeled sequences ranges between 101 and 996 samples. Samples of the defect classes are shown in Figure 3.


Unsupervised classification of surface defects in wire rod production obtained by eddy current sensors.

Saludes-Rodil S, Baeyens E, Rodríguez-Juan CP - Sensors (Basel) (2015)

Macro photography of individual defects belonging to each class.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4482004&req=5

f3-sensors-15-10100: Macro photography of individual defects belonging to each class.
Mentions: The operators of a manufacturing plant of wire rods identified and labeled the surface defects obtained for several production shifts. This has been a very tedious and time-consuming task, because it required unwinding long wire rod coils, searching the surface defects by visual inspection and classifying and putting them in correspondence with the signal recorded by the eddy current inspection system. After this manual process, a collection of labeled defects is available for validation of the developed unsupervised classification method. The surface defects have been classified by the experts into four different classes. The corresponding eddy current signals associated with them have been represented in the complex impedance plane and labeled as defects belonging to Classes A, B, C and D, respectively. An individual sequence representing each of these groups is depicted in Figure 2. The length of the available labeled sequences ranges between 101 and 996 samples. Samples of the defect classes are shown in Figure 3.

Bottom Line: An unsupervised approach to classify surface defects in wire rod manufacturing is developed in this paper.The defects are extracted from an eddy current signal and classified using a clustering technique that uses the dynamic time warping distance as the dissimilarity measure.It is shown that it outperforms other classification alternatives, such as the modified Fourier descriptors.

View Article: PubMed Central - PubMed

Affiliation: Centro Tecnológico CARTIF, Parque Tecnológico de Boecillo 205, 47151 Boecillo, Valladolid, Spain. sersal@cartif.es.

ABSTRACT
An unsupervised approach to classify surface defects in wire rod manufacturing is developed in this paper. The defects are extracted from an eddy current signal and classified using a clustering technique that uses the dynamic time warping distance as the dissimilarity measure. The new approach has been successfully tested using industrial data. It is shown that it outperforms other classification alternatives, such as the modified Fourier descriptors.

No MeSH data available.