#### DMCA

## A.: A novel approach to ball detection for humanoid robot soccer (2012)

Venue: | In: Advances in Artificial Intelligence (LNAI 7691 |

Citations: | 8 - 7 self |

### Citations

1926 |
An algorithm for least-squares estimation of nonlinear parameters
- Marquardt
- 1963
(Show Context)
Citation Context ...the accuracy and efficiency of the algorithm is compared to previous ball detection approaches, including the previous NUbots system [1], implementing Levenberg-Marquardt least squares circle fitting =-=[10]-=-; and a circular Hough transform based method [16], similar to those implemented by many RoboCup teams [11, 15]. 2 Ball Detection in Context In computer vision, a mapping from an arbitrary 3-component... |

299 |
Algorithm as 136: A k-means clustering algorithm
- Hartigan, Wong
- 1979
(Show Context)
Citation Context ...e Sect. 2) by a generalised module capable of determining multiple candidate points. In an environment where the maximum number of balls is known a priori, this is accomplished via k-means clustering =-=[8, 14]-=-. Concretely, given a set of m data points P = {x(1), . . . , x(N)} (x(i) ∈ Rm), k-means clustering attempts to partition P into K sets (known as clusters) S = {S1, . . . , SK} such that the following... |

252 | Computer Vision: Algorithms and Applications
- Szeliski
- 2010
(Show Context)
Citation Context ...e Sect. 2) by a generalised module capable of determining multiple candidate points. In an environment where the maximum number of balls is known a priori, this is accomplished via k-means clustering =-=[8, 14]-=-. Concretely, given a set of m data points P = {x(1), . . . , x(N)} (x(i) ∈ Rm), k-means clustering attempts to partition P into K sets (known as clusters) S = {S1, . . . , SK} such that the following... |

198 |
Well separated clusters and optimal fuzzy partitions
- Dunn
- 1974
(Show Context)
Citation Context ...In general, for an environment where the number of balls is known but with no guarantee every ball is present in a given image, an internal cluster validation criteria (such as the Dunn’s based index =-=[5, 6]-=-) is applied to the each cluster for K = {1, . . . , b}, with the K value yielding the best results indicating the number of balls in the current image. 4 Determining Location and Size As outlined in ... |

55 |
Another efficient algorithm for convex hulls in two dimensions
- Andrew
- 1979
(Show Context)
Citation Context ...een pixel coordinates are added to a list of points. The green horizon then becomes the upper convex hull of these points, determined by a modified implementation of Andrew’s monotone chain algorithm =-=[3]-=-. 3. Generate colour transitions: Processing of the image to locate potential field object candidates is a colour transition level operation. To generate colour transitions, each pixel along each scan... |

49 | Princen, 'Comparative study of Hough transform methods for circle finding
- Yuen, J
- 1990
(Show Context)
Citation Context ...mpared to previous ball detection approaches, including the previous NUbots system [1], implementing Levenberg-Marquardt least squares circle fitting [10]; and a circular Hough transform based method =-=[16]-=-, similar to those implemented by many RoboCup teams [11, 15]. 2 Ball Detection in Context In computer vision, a mapping from an arbitrary 3-component colour space C to a set of coloursM assigns a cla... |

32 |
The RoboCup humanoid challenge as the millennium challenge for advanced robotics,
- Kitano, Asada
- 2000
(Show Context)
Citation Context ... extraction, object recognition, clustering 1 Introduction The problem of developing a team of humanoid robots capable of defeating the FIFA World Cup champion team, coined “The Millennium Challenge” =-=[9]-=-, has been a milestone that has driven research in the fields of artificial intelligence, robotics and computer vision for over a decade. One crucial skill of soccer, the accurate, robust and efficien... |

18 |
Development of open humanoid platformDARwin-OP
- Ha, Tamura, et al.
(Show Context)
Citation Context ... subsequent advances in processor performance over the last decade, from the 384 MHz RISCbased processors of the Sony AIBO ERS-210 (2002) to the 1.6 GHz Intel Atom processors of the Robotis DARwIn-OP =-=[7]-=- platform (2012), the temporal and spatial complexity of feature extraction algorithms to solve this task has grown accordingly. With past research suggesting that colour-based algorithms are suboptim... |

16 |
Extensions to Object Recognition in the Four-Legged League
- Seysener, Murch, et al.
- 2004
(Show Context)
Citation Context ...ty of feature extraction algorithms to solve this task has grown accordingly. With past research suggesting that colour-based algorithms are suboptimal for object recognition in a RoboCup environment =-=[12, 13]-=-, particularly in the presence of varying lighting conditions, a paradigm shift from colour-based to shape-based feature extraction has been evident amongst RoboCup teams [12]. This shift has been amp... |

9 | Evaluation of colour models for computer vision using cluster validation techniques
- Budden, Fenn, et al.
- 2013
(Show Context)
Citation Context ...y RoboCup teams [11, 15]. 2 Ball Detection in Context In computer vision, a mapping from an arbitrary 3-component colour space C to a set of coloursM assigns a class label mi ∈M to every point cj ∈ C =-=[5]-=-. If each channel is represented by an n-bit value and k = |M | represents the number of defined class labels, then C →M, where C = {0, 1, . . . , 2n − 1}3 and M = {m0,m1, . . . ,mk−1} . Where computa... |

7 |
Real-time generic ball recognition in RoboCup domain.
- Martins, Neves, et al.
- 2008
(Show Context)
Citation Context ...he previous NUbots system [1], implementing Levenberg-Marquardt least squares circle fitting [10]; and a circular Hough transform based method [16], similar to those implemented by many RoboCup teams =-=[11, 15]-=-. 2 Ball Detection in Context In computer vision, a mapping from an arbitrary 3-component colour space C to a set of coloursM assigns a class label mi ∈M to every point cj ∈ C [5]. If each channel is ... |

2 |
en ligne), S.S.: Pattern recognition and machine learning
- Bishop
- 2006
(Show Context)
Citation Context ...ining candidates are twofold. Firstly, compared to other common clustering techniques such as mean shift and expectation maximisation, k-means is computationally efficient, with time complexity O(Km) =-=[4]-=-. In addition, as clustering only takes place over the set of colour transitions and very few iterations are required, this method is able to be executed in real time on the DARwIn-OP platform [7]. Se... |

2 | S.: Combining edge detection and colour segmentation in the four-legged league
- Murch, Chalup
- 2004
(Show Context)
Citation Context ...ty of feature extraction algorithms to solve this task has grown accordingly. With past research suggesting that colour-based algorithms are suboptimal for object recognition in a RoboCup environment =-=[12, 13]-=-, particularly in the presence of varying lighting conditions, a paradigm shift from colour-based to shape-based feature extraction has been evident amongst RoboCup teams [12]. This shift has been amp... |

1 |
A.: Dutch nao team, team description paper for robocup
- Velthuis, Verschoor, et al.
- 2012
(Show Context)
Citation Context ...he previous NUbots system [1], implementing Levenberg-Marquardt least squares circle fitting [10]; and a circular Hough transform based method [16], similar to those implemented by many RoboCup teams =-=[11, 15]-=-. 2 Ball Detection in Context In computer vision, a mapping from an arbitrary 3-component colour space C to a set of coloursM assigns a class label mi ∈M to every point cj ∈ C [5]. If each channel is ... |