A method for determining calibration data for at least two cameras (camera 1 , camera 2 ) in a multi view position, includes a step of determining respective parameters ((h 1 00 , . . . , h 1 22 ), (h 2 00 , . . . , h 2 22 )) for identifying at least one respective homographic transformation on respective images (image 1 ,image 2 ) taken by said cameras of a same scene, by performing respective geometry analyses on said respective images (image 1 , image 2 ), a step of performing at least one respective combined homographic transformation/feature detection step on said respective images thereby obtaining respective sets (feature set 1 , feature set 2 ) of features on respective transformed images, such that said calibration data are obtained from matches (m l , . . . , m k ) determined between said respective sets of features.

Method and arrangement for multi-camera calibration

Camera calibration in sport event scenarios

A method for determining calibration data for at least two cameras (camera 1, camera2) in a multi view position, includes a step of determining respective parameters ((h1 00,...,h1 22), (h2 00,...,h2 22)) for identifying at least one respective homographic transformation on respective images (image1, image2) taken by said cameras of a same scene, by performing respective geometry analyses on said respective images (image1, image2), a step of performing at least one respective combined homographic transformation/feature detection step on said respective images thereby obtaining respective sets (feature set1, feature set2) of features on respective transformed images, such that said calibration data are obtained from matches (m1,..,mk) determined between said respective sets of features.

In many sports and surveillance scenarios the action is dynamic and takes place on a planar surface, while being recorded by two or more zoom-pan-tilt cameras. Although their position is fixed, these cameras can typically rotate and zoom independently from each other. When rotation and zoom of each camera are known, one could reconstruct the dynamic event in 3D and generate different views of the action. Sensors exist which report zoom and orientation changes of pan-tilt units. In absence of such sensors, however, we prove that the varying zoom and rotation of two pan-tilt units can be extracted solely from the planar homography which exists between both cameras.

Maarten Aerts

Papers

Online calibration of two zoom-pan-tilt units for planar dynamic events