You are here: Home :: Area NEM :: Engineering :: Machine construction

Binjian Xin, Michael Heizmann, Sören Kammel, Christoph Stiller

Analysis of Image Sequences for the Inspection of Grinded Surfaces

This contribution addresses the automated visual inspection of defects on ground surfaces. To this aim, novel strategies based on image sequences are introduced that are acquired with varied azimuth of a directional illumination. Meaningful features are defined in the parameter space of the illumination that allow classification of defective areas in the grinding texture. As compared to standard methods, which usually analyze the texture in question by means of local operators such as the structural tensor, the presented strategies require only a considerably reduced image resolution. Moreover, they have proven reliable and robust even for grinding textures which are difficult to illuminate.

tm – Technisches Messen, Oldenbourg Wissenschaftsverlag

Print ISSN: 0171-8096
Volume: 71, 04/2004
Pages: 218 - 226

Journal homepage (external site)

Show all available items of this journal

 

Area NEM