Research On Key Problems Of Vision Based Motion Measuring System For Vehicle In Ground Test

Ground test is essential in the development of vehicle, which can improve the reliability, reduce the costs, and shorten the cycle. In the ground test, the position and attitudes of vehicle are important, which are the basic parameters in kinematics and kinetics, and can be used to evaluate the results of test, determine the causes of possible accidents, and design the system of vehicle. Vision measurement is widely used in the measuring of motion parameters. For the high precision and wide range measurement of motion parameters for vehicle in ground test, a measuring method based on the vectors of light beams is proposed referring to the traditional vision measuring methods. Meanwhile, the key issues in the development of the measuring system, including optimization of the vision measurement system, calibration of camera with large FOV(Field-of-View), and motion correspondence of projective spots, are also studied.First, the issue of optimization for vision based motion measuring system is studied. Optimization of vision measurement system can provide a theoretical basis for the design of the system. After the measuring principle based on the vectors of light beams and the corresponding system scheme are introduced, the interferences in the measurement are fully considered, and a multi-objective optimization model is established with the objectives of both measuring precision and measuring range. On the basis of this, a multi-objective particle swarm optimization algorithm based on the interaction of multi-level information is proposed. This algorithm provides the Pareto Optimal Set for the optimization of the measuring system, which avoids the interferences of measurement, while satisfies the different design requirements for measuring precision and range so as to enhance the scalability of the system.Second, the issue of onsite calibration for camera with large FOV is studied. In order to achieve the high precision, fast and onsite calibration for camera with large FOV in the vision based motion measuring system for vehicle in ground test, theory foundations of camera calibration are analyzed, and a method based on planar homography is proposed, which can be divided into flexible calibration based on planar homography and onsite adjustment based on prior information two parts. This method achieve the onsite calibration of camera with large FOV by means of several feature points on the planar target, which improves the efficiency of calibration, while the precision is also guaranteed, which is equal to that of traditional calibration methods with large scale target.In the following, the issue of motion correspondence for projective spots of light beams in cooperation targets is studied. In the measurement, in order to acquire the vectors of light beams, projective spots belong to the same light beam should be identified, which can be solved with the motion correspondence of projective spots. After the problem of motion correspondence and the related methods, including statistical methods and heuristic methods, are stated, a motion correspondence method for projective spots of light beams based on multi-source information is proposed. This method makes full use of multi-source information, such as the predicted position of projective spot, the motion constraints, and the relevant information of measuring system, so as to guarantee the accuracy of motion correspondence, in which the situation that the projective spot is out of FOV is efficiently handled.Finally, the issue of wide range measurement for motion parameters of vehicle in ground test is studied. To achieve high precision and wide range measurement of motion parameters for vehicle in ground test, cooperation target and light beam in each cooperation target are increased, and a wide range measuring method based on multiple cooperation targets is proposed on the basis of measuring method based on the vectors of light beams, which effectively solves the problems of calculation of vectors of light beams in missile coordinate system, the selection of cooperation target that participates in the calculation, and the singularity of transformation from rotation matrix to Euler angle. Meanwhile, the coordinate systems involved in the measurement are analyzed, and the corresponding calibration method and data transfer method are presented. Finally, the performance of the vision based motion measuring system for vehicle in ground test is tested. The results show that the measuring system in this paper can achieve high precision and wide range measurement for the motion parameters of vehicle in ground test.