Abstract: A detection method of the crosslines overprinting accuracy is presented to compensate the shortage of human eyes resolution. The crosslines are segmented in CLE L*a*b* color space by FCM algorithm and denoised by morphology dilation operation. A simple and reliable approach to determine the crossline center by pixels projection is proposed. The deflection angle of a crossline is computed through coordinate transformation.

Key words: printing crossline, overprinting accuracy detection, CLE L*a*b* color space, FCM algorithm

摘要： 为弥补人眼分辨十字线的不足，建立了一种检测十字线套准的方法。运用FCM聚类算法在CIE L*a*b*色彩空间完成了十字线的分割，用形态学膨胀运算处理分割后的噪声点。提出了一种用像素投影确定十字线中心的简单可靠方法，并通过Hough变换确定了十字线的偏转角度。实验表明，该方法精度高，抗干扰性强，可满足工业需求。

关键词: 印刷十字线, 套准检测, CIE L*a*b*色空间, FCM算法