Automatic mura inspection using the principal component analysis for the TFT-LCD panel

In this paper, we propose a principal component analysis (PCA)-based mura detection algorithm. Recent conventional algorithms divide the panel image with the detection window and detect the mura in each detection window. However, these algorithms have several problems due to the limitation of the detection window size. To overcome these problems, we first estimate the background image of the entire panel image by applying the PCA method. And then, the difference image between the input panel image and the estimated background image is used for the mura region decision. The experimental results show that the proposed algorithm outperforms the conventional mura detection algorithm.