SummaryThe main objective of the MARMOR project is to develop methods and tools which can detect and classify surface defects also on marbled surfaces and surfaces with textures. These tools will be demonstrated at production lines for steel and copper stripes.ObjectivesParsytec's targets of this project are twofold: firstly, to increase the percentage of defects which can be detected and classified correctly on weakly marbled surfaces using the on-line characteristics of weakly marbled surfaces to control the parameters of standard defect detection tools already in use. The second target are significantly marbled or textured surfaces, where the current approaches are not suitable for defect detection. Here the project partners will analyze different methods how two-dimensional textures can be characterized by certain parameters. The capabilities of the different approaches will be tested by applying them to a large volume of acquired images coming from real world environments of both HTS-2 and HTS-2W. After this phase we expect that we will be able to specify those production lines where these tools will overcome the actual shortage of approaches which currently makes impossible any industrially applicable defect detection.TechnologyThis project analyses different software approaches which come from the wide manifold of non-statistical and statistical tools both for the quantification of structures at surfaces (such as analytic decompositions and entropy-based characteristics) and for stochastic processes. |
![]() |
|
| FuE Gemeinschaftsvorhaben-OST (FUEGO) gefördert vom BMBF/AiF Projekt Nummer: 9502C8 | ||
ProjektpartnerGFaIParsytec GmbH |
