Researches On Calibration Algorithm For The Master-slave Camera Composed By Panoramic Camera And PTZ Camera

In recent years, with the continuous development of the panoramic imaging technology, panoramic camera has wider application in fields of security monitoring, driver assistance, on-site monitoring, vehicle inspection, etc. The combination of panoramic camera and other monitoring equipment can improve monitor effects greatly and reach satisfactory results. The Master-Slave camera surveillance system is composed by panoramic camera and PTZ camera. The fish-eye camera is capable of monitoring a large field of view, but the partial scene is unclear due to lack of resolution. The PTZ camera has high mobility and zoom ability, but the narrow angle of view restricts it to monitor several different scenes simultaneously. The Master-Slave camera surveillance system combines advantages of both cameras to monitor total and partial scene clearly at the same time in order to achieve the best effect.With respect to general calibration methods’ limitation, this paper proposes several calibration methods which are suitable for the panoramic and PTZ Master-Slave camera surveillance system. Our contributions can be mainly concluded as following:(1) Classical methods used for calibrating the Master-Slave camera surveillance system are listed, and these methods are analyzed what the advantages and disadvantages are when used for the fish-eye and PTZ Master-Slave camera.(2) Manual calibration method based on data fitting is proposed. The method divides fish-eye image into M ?N squares, and PTZ camera is controlled manually to match the data points in every squares. The sampling points’ data is recorded by the look-up table, and the other pixel points are calculated by the bilinear interpolation. The mapping relationship between the panoramic image and the motion parameter of PTZ camera can be obtained.(3) An automatic calibration method based on camera model is proposed. The method satisfies the assumptions which the two cameras’ optical centers overlap. It establishes the panoramic camera coordinate system and the PTZ camera coordinate system and calculate transformation matrix model by matching the feature points in the scene. The mapping relationship between the panoramic image and the motion parameter of PTZ camera can be obtained.(4) Another automatic calibration method based on PTZ image mosaic is proposed. The method firstly composes a mosaic image by several PTZ images and then, gets the mapping relationships recorded in the look-up table by template matching. So the mapping relationship between the panoramic image and the motion parameter of PTZ camera can be obtained.The three methods proposed in this paper used for the Master-Slave camera surveillance system have their advantages and disadvantages. So the different calibration approach is needed according to the different circumstance.