Chip Character Area Segmentation and Localization Algorithm Based on Machine Vision
The characters on the chip surface are important for sorting,and the positioning of the characters is a key step in the sorting work.In order to improve the localization efficiency and localization accuracy of the sorting work,a method is proposed to segment and localize the character area on the chip surface based on the improved region growing algorithm and convex packet detection algorithm.Firstly,we perform pre-processing operations on the acquired images;use the improved Canny algorithm to obtain the interference-free image edges,use the image edges as the seed points of the region growth method and segment the ima-ges with the image adaptive threshold as the growth criterion;use the maximum inline rectangle algorithm to coarsely locate the character regions;then use the Harris corner detection algorithm to obtain the charac-ter corner distribution locations;finally,filter the corner points and extract the key points Finally,a modified convex packet detection algorithm is used to locate the character region.After experimental verification,the designed algorithm can completely segment and locate the character area on the chip surface,and the posi-tioning accuracy and efficiency are improved by 5.3% and 15.4% compared with the original method,meeting the requirements of actual industrial production.
NETL
NSTL
万方数据
维普
