The Application of Independent Component Analysis Method on the Mura Defect Inspection of LCD Process

In the Mura defect inspection for TFT-LCD, the uneven brightness of image have directly influence to the inspection results. In order to adjust the brightness unevenness of LCD image, this paper proposed a new method which combining the homomorphic transform and the independent component analysis method. The homomorphic transform method transformed the multiplicative uneveness into additive one and then the independent component analysis method estimated and separated the mixed source signals and noise signals. The inverse homomorphic transform method estimated signals without noise, and the target image after brightness adjustment was gotten finally. The experiment results show that this method can restrain the brightness unevenness and the moire fringe of image and strengthen the defects.