In an automatic evaluation strategy for 3D X-ray computed tomography data which does not use CAD information it is useful to start with an object labelling stage. One problem of simple segmentation methods is that parts of big voids within the sample or porosity at the surface of the specimen are often assigned to the background due to similar density values. The subsequent evaluation method which acts on the labelled object regions is therefore not always able to recognize defects of the above mentioned kind. This paper presents new techniques for assigning such defects to the object to improve the result of the automatic inspection method. The performance of the method is presented by the means of an aluminum sample from the automobile industry.
Object segmentation techniques for automatic evaluation of industrial 3D CT data
Objektsegmentierungsverfahren zur automatischen Bewertung dreidimensionaler industrieller Computer-Tomographiedaten
2002
8 Seiten, 3 Bilder, 4 Quellen
(nicht paginiert)
Aufsatz (Konferenz)
Datenträger
Englisch
Techniques for Automatic Image Segmentation
Online Contents | 1999
|Techniques for Automatic Image Segmentation
Online Contents | 1999
|Automatic Moving Object Segmentation with Accurate Boundaries
Springer Verlag | 2006
|Automatic Moving Object Segmentation with Accurate Boundaries
British Library Conference Proceedings | 2006
|Moving object segmentation in video sequences by user interaction and automatic object tracking
British Library Online Contents | 2001
|