A quantile-quantile plot based pattern matching for defect detection

摘要: Pattern matching has been used extensively for many machine vision applications such as optical character recognition, face detection, object and defect detection. The normalized cross correlation (NCC) is the most commonly technique in pattern matching. However, it computationally intensive, sensitive to environmental changes lighting shifting, suffers from false alarms a complicated image that contains partial uniform regions. In this paper, pattern-matching scheme based on quantile-quantile plot (Q-Q plot) proposed detection applications. Q-Q plot, quantiles of an inspection are plotted against corresponding template image. p-value Chi-square test resulting then quantitative measure similarity between two compared images. quantile representation transforms 2D gray-level information into 1D one. It can therefore efficiently reduce dimensionality data, accelerate computation. Experimental results have shown fast tolerable minor displacement process variation. excellent discrimination capability detect subtle defects, with traditional NCC. With proper normalization be moderate light changes. assembled PCB (printed circuit board) samples, IC wafers, liquid crystal display (LCD) panels efficacy