An Improved Randomized Hough Transform Algorithm for Circle Detection

摘要

Circle detection plays an important role in the vision measurement system. Randomized Hough transform (RHT) ,as an important way in circle detection,has the disadvantages of massive storage space and calculation in dealing with complex images. According to the arrangement regulations of pixels on circle boundary ,this paper presents an improved algorithm on circle detection based on RHT. In order to obtain the valid edges,corrosion and expansion operations of mathematical morphology are introduced. The search for the edge pixels of image by the FourNeighborhoods counterclockwise search within the semicircle range can greatly reduce the search time and storage space. The 1/4-length of the minimum continuous curve is named as the step to select the dots by Equal-Interval Hough transform,and It increases the probability of the necessary dots. Experiments demonstrate that the algorithm achieves rapid circle detection.