Enhanced on-line and post-production defect detection

Important results from the project

The project has reached its goals: - The detection of cracks in hot bars and imprints in cold plates, were evaluated and successfully demonstrated in lab environments. - Various IR imaging solutions were evaluated including Middle-Wave InfraRed (MWIR) and LongWave InfraRed (LWIR) cameras. - Image processing techniques based on edge detection and Hough transform have been implemented and evaluated, and provided a clear enhancement of crack defects. - A project proposal for further development and evaluation of these techniques in industrial environments has been submitted,

Expected long term effects

The pre-study has: - confirmed the clear benefits of IR imaging for the detection of cracks in hot products through a set of successful lab experiments, - emphasized the opportunities provided by image analysis tools to automate defect detection for process monitoring and quality control, - underlined the need to develop camera housings compatible with production environments. Next project phase will enable the implementation of quality control on warm and cold steel products by detecting and identifying surface defects that occur at different manufacturing stages.

Approach and implementation

- Hot products (bars from Ovako, slabs from SSAB): The detection of defect (cracks) turned out to be relatively easy with conventional IR intensity imaging solutions. Good contrasts were obtained both with MWIR and LWIR cameras. Image processing tools can be used to further enhance the detectability of cracks, and make automation possible. - Cold products (sheets from SSAB with imprints or oxides) Defects can be detected by measuring differences in reflectance between defect and non-defect areas, using MWIR or LWIR cameras.