A Fast Method for Extracting Parameters of Circular Objects.

A region-based method is proposed for extracting parameters of circular objects in images. The Hough space is defined by a 2-dimensional accumulator with less columns. By analysing the voting values distribution in each column of the accumulator, a quadratic function, which denotes the relationship between the voting value and the voting distance, is deduced. Regrading all columns, a linear function formula is deduced by analysing the relationship between mean voting distances and the voting angles. After all region pixels have voted for 2D accumulator, a quadratic function is fitted in each column. The circle radius is calculated based on the fitted coefficients. Then a linear function is fitted to mean voting distances corresponding to every voting angles. The circle center coordinates are computed based on the fitted coefficients. Synthetic images and real-world images are used to test the proposed region-based method. Experimental results show the proposed method is fast and accurate even in the presence of contour defects and outliers.