An Algorithm for Detection of Calibration Points for Photogrammetric Camera Calibrations Independent of Illumination Variations

Getting 3D information from 2D images depends on the appropriate camera calibration of the internal and external CCD (Charge Couple Device) camera parameters. The photogrammetric camera calibration technique needs a set of world coordinates and their corresponding calibration object image coordinates as an input. During this work we developed an algorithm to detect the image coordinates of a calibration object for different illumination conditions with subpixel accuracy. In other words, the proposed algorithm allows that any photogrammetric camera calibration technique that needs image coordinates of the calibration object will not be sensitive to illumination changes within a 200 to 6000 lux interval. The algorithm was tested in a vision system composed by a P.C., a robotic arm and a CCD camera. The system automatically controls the position of the camera within the work space and it captures and processes the images using Matlab®. Key-Words: Machine Vision, Image Processing and Applications.