A novel circle detector based on sub-image point pairs

In this paper a novel circle detector that can be used in real image processing is proposed. Since the drawbacks of standard Hough transform (SHT) and other circle detection methods such as least square method require large computation and the large storage, the new method is proposed to reduce computation cost and storage requirement. First, edge image is divided into several sub-images by the properties of circle, and then point pairs are chosen from the sub-images to calculate the parameters of circles. Second a data merging method is used to process the parameters. At last, both synthetic and real images are processed with the proposed method, the experimental results show that the proposed detector is capable of detecting circle features with superior accuracy under various image conditions.