Stereo Camera Calibration For Autonomous Car Applications

Several stages will be carried out in developing an intelligent transportation system, including building stereo vision by utilizing two cameras with stereo performance. Before vehicle object detection, vehicle distance and speed detection are implemented in the user’s vehicle by using cameras. Camera calibration is a process that should be noticed in computer vision. The calibration process is usually carried out if the camera is to be used to estimate the camera’s distance to a particular object. The camera stereo calibration system has been carried out and successfully implemented to detect vehicle distance and relative speed of vehicles on autonomous cars. OpenCV with c920 stRealm pro and c922 stRealm pro cameras is used to implement stereo vision on the camera. Optimal calibration conditions were obtained on a chessboard pattern measuring 3 cm with a distance between the lenses of 9.8 cm and shooting ranges from 0.5 cm. The optimal conditions were obtained based on the RMSE value of 0.4932.