Crack detection in magnetic tile images using nonsubsampled shearlet transform and envelope gray level gradient

A novel algorithm based on nonsubsampled shearlet transform (NSST) and envelope gray level gradient (EGLG) is proposed for detecting low-contrast cracks in variably illuminated magnetic tile images. The algorithm first adopted NSST to decompose the original images into multiple subbands at different directions and scales. Then a novel column filtering based on EGLG was employed to remove the uneven background in the approximation subband, and a shearlet coefficient variance discriminator (SCVD) was used to eliminate interferences of noise and textures in the detail subbands. Finally the cracks were extracted from the reconstructed images. To verify the validity of the proposed algorithm, extensive experiments were conducted in a novel machine vision based system and its comparison with traditional algorithms was given. Experimental results show that this method achieves an accuracy rate of 95.5% in detecting cracks longer than 0.9 mm with an average runtime of 0.576 s, and outperforms traditional methods in terms of accuracy and robustness.