Least-square method is usually adopted for fitting of geometric elements in coordinate measurement. However

in the process of short arc measurement

it is difficult to keep precision due to lack of feature points and influence of noise jamming. Aimed at measurement of short arc based on machine vision system

centre constraint least-square method is proposed in the study. Influence of central angle on measurement error of circle centre and radius adopting least-square method is analyzed. It was proved that the measurement error is less than 0.01mm by using our method to measure the radius of a short arc with 20-degree central angle. Results of comparative experiment show that centre constraint least-square method can improve the accuracy of circle fitting efficiently in short arc measurement.