Real-time implementation of motion-based filtering in Magneto-Optic Imager

The Magnetio-Optic Imager (MOI) is a relatively new Nondestructive Inspection Technique, that produces analog images of the magnetic field associated with the eddy current on the surface and subsurface structures. However, the magnetic domain wall structures of the sensor add serpentine noise patterns to the MOI images that degrade the inspection capability of the MOL A Motion-based Filtering (MBF) algorithm is shown to represent an effective technique for removing the domain-related noise patterns. The MBF utilizes the dynamic nature of sample features relative to the static noise introduced by the sensor. The filtering algorithm is based on a multiple inter-frame subtraction of the images obtained during scanning the sample. This paper presents the results of MBF along with a systematic attempt to make the implementation fast enough to meet real-time processing constraints.