Design And Implementation Of Laser Anti-counterfeiting Code Character Segmentation Algorithm

In recent years, with the development of computer science and informationtechnology, character segmentation technology is becoming mature. In manyindustries, such as financial, revenue, transportantion and book publishing industry,Character segmentation technology has been widely used. However, the algorithm inhand still many problems, these problems highlighted in the following two points:Firstly, at present,most of these algorithms are designed for some specific fields,notfor all. Secondly, when character segmentation is applied in the real scenario,problems arises. In terms of the existing of various interference factors, charactercomplex background and mutual adhesion between characters always appear thatmakes the segmentation of characters rather difficult, which differs from thetraditional character segmentation technique. Therefore, recently adhesion charactersegmentation has gradually become a hot issue in research field. In this paper, aimingat the actual demand of a corporation, I do some research on the pretreatment of Laseranti-counterfeiting code image, zone location and segmentation technique.Main works is as follows:(1)Research on the algorithm for the pretreatment of Anti-counterfeiting codeimage. Before zone located the Anti-counterfeiting code characters,I conduct aserious of pretreatment on the Anti-counterfeiting code image, including strenthen theAnti-counterfeiting code image, filtering and noise reduction and Binarization.Experimental results show that if the pretreatment works well, it can improve thequality of the pictures and makes a good preparation for the Character orientation andsegmentation.(2)Research on the algorithm for zone location of Anti-counterfeiting codecharacter. There are two approaches for locating Anti-counterfeiting code: the oneapproach is Based on the vertical edge detection and projection, and the other one isthe approach of Straight line detection based on Hough transform. Both approachescan accomplish the detection and location for the character zone accurately, and knowthe position of character’s connected body in Anti-counterfeiting code image. (3)propose the Improved water algorithm for the segmentation ofAnti-counterfeiting code character. This algorithm spread the width of the zone of thecharacters correspondingly, and eventually realise the accurate segmentationAnti-counterfeiting code characters successfully. Experimental results show that thecharacter segmentation method introduced in this paper can adapt to theAnti-counterfeiting code character segmentation in complex environment, and it has ahigher segmentation accuracy, a shorter segmentation time, and a better instantaneityand robustness.According to the actual test, the algorithm in this paper can completely meets therequirement of the project, it can be used for a reference for the application of relatedcharacters.