| Currently, the domestic main technical means adopted by surface surveillance are the surface movement radars, ADS-B, and multi-lateration, etc.. Binocular stereo vision measuring, as a low cost of emerging technology, has the potential to be an alternative to non-cooperative surveillance radar for small and medium airport surface. Therefore, the study of surveillance applications based on computer vision technology for the scene is necessary. Image matching is an identification process of the same points by a certain matching algorithm between two or multiple images.The topic is through binocular stereo vision measuring principle, to establish and validate a two-dimensional mathematical model of camera visual measurement of the airport scene for airport surface surveillance technology, and to provide a theoretical basis of the technology and the reliability analysis of the implementation in airport. Additionally an image coordinates matching method is proposed based on the mathematical model.The binocular measurement principle is based on the camera coordinate system, the image coordinate system and the world coordinate system. By using the pinhole imaging, a two-dimensional surface surveillance measurement model is established with a pitch angle and rotation angel of the camera. The model can recover the target location coordinates in a 3D scene from two-dimensional image. Using experimental data this model is verified with the required accuracy and a good performance.The error analysis of the model is studied for each of the model parameters, including the internal camera parameters and camera location parameters, such as the pitch angle, the rotation angle, the focal length, the pixel size, the relationship between the two camera positions etc. By making the simulation analysis, the impact to the accuracy of the target model by each parameter error is simulated, as well as the total effect of various parameters error synthesized. This accuracy is measured by distance error. Then using experimental data, the validity of the focus error analysis of the model is verified, which proves that this error analysis method is reliable.According to the pinhole imaging principle, a new method for image matching is proposed for the model. Based on the formula of the distance between two different spatial straight lines, and that the distance between two intersecting straight lines is zero, a new matching mathematical formula is revealed. The formula is focus on the distance. By calculating the distance between two straight lines that connect any two image coordinates to the optical center, whether it is a pair of matching points or not is determined. Using experimental data this formula is verified, which is proved to be correct.The results showed that, with the pitch angle and the rotation angel, the two-dimensional target solution model is correct and reliable. Its precision accuracy meet the requirements of the airport surface surveillance, and it is fit for airport surface surveillance target solution. The impact of the focal length error for the model is significant. The mathematical matching formula is reliable and easy to operate. As the theory is proved correct, the model can be used for image matching. |
PDF Full Text Request