Image detection method for text defects in printed matter based on block template matching
Printing defect detection has gradually become one of the important links to improve printing quality with the rapid development of automated production.In the detection of printed text defects,due to the influence factors such as the flexibility of the raw paper,there is a problem of low local template matching accuracy,resulting in high false detection rate.In order to further improve its detection accuracy,this paper proposes a printed text defect image detection method based on the block template matching.This method takes the template image as the detection standard,uses the HoughLines to detect the crop line of the printed matter,completes the overall posture and size correction of the scanned image,and adopts the block template matching method to complete the local high-precision matching between the template image and the scanned image,and then obtains the defect image such as missing print,ink spot and ink color uneven by positive and negative difference and fusion processing of the image,and finally performs the connection domain analysis to remove the pseudo-defect,realizing the image detection of the text defect of the printed matter.Experiments have verified that this method can meet the detection requirements well and has certain application value in the printed matter industry.
NETL
NSTL
万方数据
