Abstarct: The number of traffic control cameras is quickly increasing. Many of them lack the automated processing and are supervised by human operator. Extracting data by user needs manual configuration and maintenance in addition to highly qualified staff and in many cases the requires camera calibration. To remove manpower, reduce costs and increase the processing speed we need fully automated method for obtaining camera parameters. In this thesis we propose a method for fully automatic calculation of internal and external parameters of roadside cameras toward estimating real vehicle dimensions. This method allows the calibration of the camera (including scale) without any user input and can be conducted with just a few minutes of video surveillance. This method has practical purposes, including vehicle speed and dimensions measurement and vehicle classification. An important assumption for the vehicle movement is that at least a part of the vehicle motion is approximately straight that this assumption does not limit the usability of our approach experimentally. Due to the fact that obtaining vehicle dimensions are not dependent on video dimensions, at first reduce the video size to decrease the amount of calculations at the First step we compute three orthogonal vanishing points, through moving vehicles. The second step is make the three-dimensional box of each car and calculation of scene scale. In the third step we use the dimensions of the 3D bounding boxes for camera calibration. The results indicate that the proposed method have high precision and speed.