Defect Inspection in Low-Contrast LCD Images Using Hough Transform-Based Nonstationary Line Detection

Abstract

In this paper, we propose a Hough transform-based method to identify low-contrast defects in unevenly illuminated images, and especially focus on the inspection of mura defects in liquid crystal display (LCD) panels. The proposed method works on 1-D gray-level profiles in the horizontal and vertical directions of the surface image. A point distinctly deviated from the ideal line of a profile can be identified as a defect one. A 1-D gray-level profile in the unevenly illuminated image results in a nonstationary line signal. The most commonly used technique for straight line detection in a noisy image is Hough transform (HT). The standard HT requires a sufficient number of points lie exactly on the same straight line at a given parameter resolution so that the accumulator will show a distinct peak in the parameter space. It fails to detect a line in a nonstationary signal. In the proposed HT scheme, the points that contribute to the vote do not have to lie on a line. Instead, a distance tolerance to the line sought is first given. Any point with the distance to the line falls within the tolerance will be accumulated by taking the distance as the voting weight. A fast search procedure to tighten the possible ranges of line parameters is also proposed for mura detection in LCD images.Experimental results have shown that the proposed method can effectively detect various mura defects including spot-, line-, and region-mura. It performs well for the test images containing nontextured and structurally textured patterns in the unevenly illuminated surfaces.

A simple computational framework for defect detection system with orthogonal polynomials transcoded coefficients

“…Use of multi-level threshold is reported in [41] to identify the spot-type defects in Thin Film Transistor -Liquid Crystal Display (TFT-LCD). With the help of Independent Component Analysis (ICA) and minimum spatial correlation for spatial redundancy, design of defect detection scheme of uneven brightness in low contrast images, in LCD is reported in [42,43].…”

Bottle cap scratches detection with computer vision techniques

“…After we get the optimal θ, we transform the test bottle cap edge image into the same orientation with the template bottle cap edge image according to (5).…”

“…In [4], singular value decomposition (SVD) is combined with wavelet transform to detect defects during TFT-LCD manufacturing processes. In [5], authors proposed a modified Hough transform to cope with the line detection problem in a nonstationary signal. In steel industry, gradient texture information is combined with support vector machines to detect scratches [6].…”

Automatic Mura Detection for Display Film Using Mask Filtering in Wavelet Transform

“…There are various existing methods for detecting mura, such as removing the background image [3]- [7] and using mura features [8], [9]. To detect mura by removing the background image, Lu and Tsai [3], [4] and Wang et al [5] proposed a method for estimating the background image using singular value decomposition (SVD).…”

“…This method requires many calculations, making detection slow, and changes parameters according to image contrast. Li and Tsai [9] proposed a method for detecting mura by applying threshold values to a distance between a direct line decided by modified hough transform and pixel value of the input image. However, this method has a weak point that the limited range of Hough transform parameter has to be redecided according to images.…”