Fully Automatic Matching of Circular Markers for Camera Calibration
Camera calibration is one of critical steps incomputer vision,also an exhaustive process becausesubstantial human computer interactions arefrequently required to deal with the matching problem.In this paper,an automatic matching method ofmarkers for camera calibration is presented based onthe local architecture characteristics of a new planarcircle pattern,which can be applied to a wide range ofspoiled images including those of distortion,noise,andblur.Firstly,a robust ellipse detector is developed toextract ellipses from the image.After the centroids ofthe ellipses are triangulated into triangle network byDelaunay method to find k annular neighborhood ofeach ellipse,a cost function is defined to locateorientation ellipses for the homography matrixbetween the image and the calibration pattern.Thenthe correspondences between the marker ellipses andthe marker circles are established by a homographymethod followed by a point sets registration strategy.Finally,the method is tested with real and transformedimages to show its accuracy and robustness,which isentirely unsupervised and easily embedded into theexiting algorithms
Renbo Xia Jibin Zhao Weijun Liu Jinting Xu
Key Laboratory of Industrial Informatic,Shenyang Institute of A utomation,Chinese Academy of Sciences,Shenyang 110016,China
国际会议
厦门
英文
1065-1070
2008-11-17（万方平台首次上网日期，不代表论文的发表时间）