Regular Article
Recognition of Partial Circular Shapes from Segmented Contours

https://doi.org/10.1006/cviu.1996.0023Get rights and content

Abstract

This paper presents a circle-finding method to deal with partially occluded circles using segmented contours. Contours are segmented using a criterion based on the derivative of the curvature. Constant curvature segments are primitives inputted into a clustering algorithm which brings out the relationships among contour segments which are likely to represent the same circles. A minimization criterion is used to find the circle parameters, providing an accurate parameter estimation. The method allows recovery of the contour segments used to estimate circle parameters, providing useful information in some practical applications.

References (0)

Cited by (53)

  • Image processing for the experimental investigation of dense dispersed flows: Application to bubbly flows

    2019, International Journal of Multiphase Flow
    Citation Excerpt :

    Although the Hough transformation is very interesting for the detection of spherical objects, it is known to be expensive in time and memory and to be unsuitable for the detection of partially occluded circular shape objects. To overcome this difficulty, Pla (1996) proposed a method based on the calculation of the curvature of objects contours to detect connecting points that correspond to the intersection of two bubble contours: segments having a constant curvature are identified to be object contours, then, the overlapping contours are separated thanks to the analysis of the derivative of the curvature. The application of this method on real complex images enables the detection of full and partially occluded spherical objects.

  • An efficiency improved recognition algorithm for highly overlapping ellipses: Application to dense bubbly flows

    2018, Pattern Recognition Letters
    Citation Excerpt :

    The curvature of a 2D set boundary exists if the arc boundary of the regarded set is twice differentiable with continous second derivative. Methods to compute the curvature of a boundary set on a 2D image can be found in [20]. These connecting points can therefore be interpreted as concavity points, which will be used by the detection algorithm.

  • An integrative image measurement technique for dense bubbly flows with a wide size distribution

    2015, Chemical Engineering Science
    Citation Excerpt :

    Thus, ignoring these clusters in the measurement would bias the estimations. To consider the size of such bubble clusters, some authors have proposed to approximate the overlapping bubbles through an object recognition approach which fits an ellipsoidal shape to the object areas (Pla, 1996; Honkanen et al., 2005). Some other reports have focused upon segmenting these clusters by implementing watershed algorithm (Bonifazi et al., 1999; Lin et al., 2008; Zhou et al., 2010; Zhang et al., 2011; Lau et al., 2013).

  • On covering a digital disc with concentric circles in ℤ<sup>2</sup>

    2013, Theoretical Computer Science
    Citation Excerpt :

    In later years, several approaches have been suggested for improving the method of circle construction or circle approximation [2,4,7,25,32,35,43,46–48]. Added to this, is the more challenging problem on recognizing circular arcs/objects in digital images, which have several solutions based on characterization and parameterization of circular arcs [10–12,14,16,23,26,30,36,37,39,40]. This paper is focused on identifying and characterizing the absentee pixels (2D points with integer coordinates) in the cover of a digital disc with concentric digital circles.

View all citing articles on Scopus
View full text