Detección Eficiente de Elipses en Imágenes. Aplicación al Posicionamiento 3D de un Robot Industrial

Resumen

En este artículo se presenta un algoritmo para la detección de elipses en imágenes, cuyo objetivo es el cálculo d e la posición 3D de una característica circular en una aplicación robótica. El algoritmo emplea un procedimiento estocástico RANSAC cuya eficiencia ha sido mejorada. El muestreo aleatorio ha sido sustituido por un muestreo guiado sobre las cadenas de contorno de la imagen, que son ordenadas de acuerdo a un criterio de probabilidad de formar parte de la elipse buscada. Esta estrategia disminuye notablemente la cantidad de muestras necesarias, permitiendo que el algoritmo sea adecuado para tiempo real.

Abstract

In this paper, we present a ellipse detection algorithm developed to measure the 3-D position of a circular feature in a robotic application. The algorithm uses a RANSAC stochastic procedure whose efficiency has been significantly improved, substituting the random sampling with a guided sampling on the curve segments in the image. The contours of the image are first split analyzing their curvature. Then the curve segments are sorted according to their likelihood to be part of the ellipse. We have used the length as a prior indicator of this likelihood. The RANSAC algorithm starts considering only the longer curve segments whilst shorter curve segments are progressively incorporated. This strategy notably diminishes the amount of samples needed and makes the algorithm suitable for real time.

Referencias no citadas

Ballard, 1981, Bouguet, 2010, Chaumette and Hutchinson, 2006, Cheng and Liu, 2004, Chum and Matas, 2005, DeSouza and Kak, 2002, Fischler and Bolles, 1981, Fitzgibbon et al., 1999, Fuente and de la, 2011, Gracia and Pérez, 2010, Halir and Flusser, 1998, Hahn et al., 2008, Hartley and Zisserman, 2004, Hermann and Klette, 2006, Hough, 1962, Illingworth and Kittler, 1998, Kanatani and Liu, 1993, Kanatani and Ohta, 2004, Liu and Qiao, 2009, Mai et al., 2008, Marji, 2003, McLaughlin, 1998, McLaughlin and Alder, 1998, Miguel Trespaderne and Fuente, 2009, Myatt et al., 2002, Mundy and Zisserman, 1992, OpenCV, 2011, Rosenhahn et al., 2004, Rosin and West, 1995, Rousseeuw and Leroy, 1987, Soria et al., 2008, Thanh et al., 2009, Tordoff and Murray, 2002, Xu et al., 1990, Zhang, 1997 and Zhang, 1998.