Affiliations: [a] NRC Aerospace, Ottawa, ON, Canada | Université Laval, Canada | Tohoku University, Japan
Correspondence:
[*]
Corresponding author: Marc Genest, NRC Aerospace, 1200 Montreal
Rd, Building M14, K1A 0R6, Ottawa, ON, Canada. Tel.: +1 613 949 1326; Fax: +1
613 952 7136; E-mail: [email protected]
Abstract: In its current state, the wide acceptance of the Magneto-Optical
Imaging (MOI) technique is hindered due to noise, lack of recordable results,
and impossibility of data post-processing. This paper presents some add-ons
made to a commercial MOI system to ease the image interpretation, archiving and
reporting of the results. In addition, a few image processing techniques are
also employed in an attempt to perform automatic flaw detection. The recording
capability of the MOI instrument output images was addressed by digitizing the
video signal in video or image files. To help with the identification of the
damage location and distance between images, a rotary quadrature encoder was
mounted onto the MOI scan head. The use of the encoder allowed the
identification of the inspection location with respect to a reference position,
such as the beginning of the scan. Moreover, it allowed saving images at fixed
intervals, which were then stitched into a single image, thus simplifying the
post inspection analysis process. Both live and post-inspection image
processing capabilities were made available. Implemented image processing
included background subtraction, de-noising, contrast adjustment and
morphological operation, among others. Contrast stretching transform and
background subtractions were found to be among the most powerful techniques
that could be used in simplifying the image interpretation.
