Motion-based filtering of magneto-optic imagers Unsang Park a, * , Lalita Udpa b , George C. Stockman a a Department of Computer Science and Engineering, Michigan State University, 3208 Engineering Building, East Lansing, MI 48824, USA b Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI 48824, USA Received 9 May 2003; received in revised form 26 September 2003; accepted 2 October 2003 Abstract The magneto-optic imager (MOI) is a powerful device for the nondestructive inspection (NDI) of aging aircraft. MOI produces analog images of magnetic flux leakage associated with eddy current distribution around surface and subsurface structures. The main advantages of using MOI are its fast inspection speed and easy interpretation compared with conventional Eddy Current NDI instruments. However, due to the magnetic domain wall structures of the sensor, the MOI images are corrupted by noise, which lowers the MOI inspection capabilities. The domain walls produce serpentine pattern noise, which can be reduced by improving the sensor or by use of image processing methods. This paper introduces a motion-based image processing method to reduce the background noise. Initial results of implementing the algorithm on real field data are presented. q 2003 Elsevier B.V. All rights reserved. Keywords: Magneto-optic imager; Serpentine pattern noise; Motion-based filtering 1. Introduction The magneto-optic imager (MOI) designed by Physical Research Instrumentation (PRI) [1] is a relatively new variation of eddy current technology, which can inspect large areas at high speed compared with classical Eddy Current methods. MOI mainly uses eddy current excitation and polarized light in conjunction with a MO sensor to form visual images of the magnetic flux density associated with induced eddy currents. The main advantages of MOI are its fast and easy inspection capabilities in comparison with other conventional nondestructive inspection instruments. MOI inspection is up to 10 times faster than conventional eddy current methods because of its large inspection area and the ease of inspection and interpretation [2,3]. However, MOI images are severely corrupted by a characteristic serpentine pattern associated with domain walls in the sensor. The noise hinders detecting small cracks and corrosion located in second and third layers, limiting the capability of images. This leads to the need for an image processing algorithm for reducing this background noise. Conventional image processing methods are focused on processing each frame of MOI in a scan sequence [4,5]. In this paper, we introduce a new image filtering technique based on the characteristics of MOI revealed in the sequence of images generated during scanning of the sample surface. In MOI images, noise associated with the domain structures in the sensor is overall stationary as the sensor moves while images associated with structures (rivet) or corrosion move from frame to frame due to relative motion of sensor and sample. The algorithm presented in this paper is based on separating the moving parts from the stationary parts in the sequence of images. Noise is reduced by retaining and boosting moving parts and suppressing the stationary part of images. In the remainder of this paper, we first analyze the characteristics of MOI images and describe the motion-based filtering technique. Finally, experimental results of motion-based filtering with a real MOI field data are presented. 2. Analysis of MOI images The MOI inspection is generally conducted by a human operator by scanning the surface of sample with the sensor. 0262-8856/$ - see front matter q 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.imavis.2003.10.001 Image and Vision Computing 22 (2004) 243–249 www.elsevier.com/locate/imavis * Corresponding author. Tel.: þ 1-517-3559319; fax: þ 1-517-4321061. E-mail addresses: parkunsa@egr.msu.edu (U. Park), udpal@egr.msu. edu (L. Udpa), stockman@cse.msu.edu (G.C. Stockman).