Sub-pixel localization algorithm of micro-vision based on Hough transform

According to geometric characteristics of micro-image, sub-pixel localization algorithm of crisscross image based on Hough transform (SLACHT) and sub-pixel localization algorithm of round image based on Hough transform (SLACHT) are presented. Sub-pixel localization of linear and round image of micro-image is studied in detail. In contrast with classic Hough transform (HT) algorithm detecting straight line and classic random Hough transform (RHT) algorithm detecting circle, CSLHT and RSLHT show more advantages. Reliability of the algorithms is proved by comparing with algorithm written by Matlab software. Many experiments show that localization accuracy of the algorithms is less than 0.01 pixel, and that they are accurate, running fast and strong robust.