Detection of Surface Defects on Red Hot Wire
During the production of wires in a rolling mill the red hot steel is transported at very high speed up to 110 m/s. Several fast line scan cameras are scanning the complete surface continuously to detect surface defects in an early state of production. In a first step, only sections are scanned and stored to disk to develop those algorithms, which should provide an in line inspection at a later project state.
The topic of this thesis is the detection of surface defects on images of red hot steel cylinders, scanned with line scan cameras at very high speed. The images will show high dynamics, resulting from special illumination conditions, and typical patterns, which come from normal production process. The correction of the illumination and motion induced artefacts and a segmentation of possible defect regions will be a first step. This will be followed by a feature calculation and classification procedure, where a separation between real defects and false positive hits from the segmentation process will be performed and optimized.
The main tasks of the thesis are the
- Illumination correction
- Segmentation of possible defect regions
- Feature generation and classification – all with the background of the necessity to process very high data rates at a final production system
- Performance tests of several developed solutions on a set of provided images with manually segmented defects
The thesis will be conducted in the scope of the COMET-project Vision+ - Integrating visual information with independent knowledge (http://www.comet-visionplus.at).
C/C++, Texture Analysis, Pattern Recognition
this mastertheses at JOANNEUM RESEARCH DIGITAL is scheduled for approx. 6 month.
JOANNEUM RESEARCH offers interested students cooperation with all Austria but also European or non-European educational institutions and is willing to assist you looking for a supervisor.