It is significant to detect circles

obtain the parameters，which can be applied to industrial computed tomography（CT） image measurements as well as the reverse engineering. A method to precisely measure circles in industrial CT image has been developed based on Facet sub-pixel edge detection and improved algorithm for calculation of probability of circles exitence. To overcome the defects of long-time calculation and memory consuming in the current method of existence probability map for detecting circles

the present method first makes sure the potential range of the center of a circle

and store the message which meet the required numerical values in the chain. By applying the method to the measurement of 10 circles in an image obtained from an industrail CT system with its spatial resolution as 2.0lp/mm

the result shows that this method has greatly improved the calculation rate

and the circles measurement accuracy is better than 0.5%.