| Least squares analysis is often used to evaluate the circularity of a set of discrete data obtained from a cylindrical surface. It is quite possible that such an analysis can lead to an erroneous determination of the center and radius. The significance of the research is that for the first time, we can provide criteria for the arrangement of data points obtained from surface of a cylindrical part with Coordinate Measurement Machine such that it will render a unique measure of the departure from circularity. We have shown that for {dollar}4 le n le{dollar} 200 and n is a multiple of 4, there exists a unique circle which best fits the data points when the data points are scattered around an annulus with equal spacing angle-wise and randomly located radius-wise. The annulus is constructed in such a way that the radius of the inner circle is 1 and the width of annulus, {dollar}kappa{dollar}, is no greater than 0.05. |
