Position and attitude calibration method of the CCD camera using the genetic algorithm

In unmanned vehicles, it is important to know the accurate position and attitude of a CCD camera so as to analysis the target region using the 3D modeling data. In general, a tracker like GPS and gyroscope is attached to the CCD camera in order to track the position and attitude of it. However, it is hard to know the accurate position and attitude of the CCD camera, because of the various error factors. Especially, mechanical misalignments can cause changes in the position and attitude of the projected virtual images that are difficult to compensate. In this paper, we describe the position and attitude calibration method of the CCD camera using the genetic algorithm. The real image obtained from the CCD camera is combined with the virtual image rendered by a computer using the position and attitude of the tracker and the 3D modeling data in the static environment. And the combined image is displayed the monitor screen. The genetic algorithm can search the optimal position and attitude of the CCD camera to minimize the registration error using several feature points that a user chooses. This method can search the accurate position and attitude of the CCD camera in the world coordinate system. And it can also find the 3D transformation matrix to adjust the misalignment between the CCD camera and tracker, even though the position and attitude of the CCD camera in the world coordinate system are not known wholly. Therefore when the system works in dynamic environment, the error caused by the distance and attitude with respect to the CCD camera and tracker can be adjusted.