A Fast Circle Detection Algorithm Based on Circular Arc Feature Screening

Circle detection is a crucial problem in computer vision and pattern recognition. In this paper, we propose fast circle algorithm based on circular arc feature screening. order to solve the invalid sampling time consumption of traditional algorithms, improve fuzzy inference edge by adding main contour screening, refinement, arc-like determination enhance positioning accuracy remove unnecessary edges. Then, strengthen features with step-wise two matrices set auxiliary points for defective circles. Finally, built square verification support region further find true complete constraints. Extensive experiments were conducted complex images, including defective, blurred-edge, interfering images from four diverse datasets (three publicly available one built). The experimental results show that our method can up 89.03% filtering superior RHT, RCD, Jiang, Wang, CACD terms speed, accuracy, robustness.