An accurate inspection of PDP-mesh cloth using Gabor filter

This paper proposes an accurate inspection technique of fabric defects in an electromagnetic wave shield mesh of the plasma display panel (PDP-mesh) for automatic detection. The proposed method utilizes two-dimensional discrete Fourier transform (DFT) and a Gabor filter for accurate detection. The DFT of the whole image enables the user to detect global defects, such as yarn density and yarn orientation, of the manufactured PDP-mesh. The determined global features can be utilized to design the optimum Gabor filter to detect the local defects such as clogging, yarn lack, and yarn slippage. Defective areas can be extracted from the output of the Gabor filter by using a variable threshold method. Experimental results show that the proposed method can accurately detect the defects from scanned PDP-mesh images.