Dynamic Multi-information Measurement Method For Aircraft Physics Simulation Based On Stereo Vision System

Aircraft physical simulation test is an essential part of the aircraft development process.With the continuous development and progress of the aerospace technology,the requirements for the functions of aircraft simulation system are becoming increasingly diversified.However,in a large number of complex simulation experiments,it can be observed that although the simulation model,simulation technology,and simulation environment have made a qualitative leap,the source of the simulation data still remains at the ideal assumption or statistical level with large deviations.At present,the vision measurement technique based on the combination of photography and computer technology has become an extremely important non-contact measurement method.Therefore,on the basis of the application of vision measurement technology in aircraft ground simulation tasks,this dissertation aims at meeting the requirements of high-precision external parameter testing of motion parameters and relative motion relationships in aircraft physics simulation tests,as well as studying on the dynamic multi-information measurement technology of aircraft physics simulation.Subsequently,the information such as aircraft motion trajectory,macroshape and pose parameters can be obtained precisely in the full field of view,which provides credible data support for the dynamic performance analysis and evaluation of the aircraft ground test.The main research work is as follows:(1)The high accuracy calibration of the camera is crucial for accurate measurement in vision system.Aiming at the problem of camera calibration with mu lti-parameters in space target pose measurement ambient,a novel calibration method based on dual update strategy weighted differential evolution particle swarm optimization is proposed.A judgment factor is constructed to control the usage proportion of weighted differential evolution(WDE)algorithm and particle swarm optimization(PSO)algorithm in each iteration process,which allows the PSO or WDE algorithm to update individuals according to the probability laws.Moreover,through the information exchange mechanism,the individual obtained by the WDE operation is utilized to guide the individual evolution process operated by PSO algorithm.The proposed WDE-PSO(Weighted Differential Evolution and Particle Swarm Optimization)algorithm can ensure the diversity and effectiveness of the individual evolution of the population,and it is coupled with the camera nonlinear calibration model and parameters.Consequently,the proposed method can simultaneously realize the combined nonlinear and global continuous optimization of camera parameters.(2)The problem of full-field three-dimensional deformation measurement of slender or biggish specimens is investigated,upon which a novel stereo digital image correlation measurement method based on multi-camera network joint constraint optimization is proposed.By establishing the joint constraint relationship of the multi-camera network,any two CCD cameras can be bound as a whole to complete the partial measurement of the specimen,and lastly the measurement results are mapped to the unified reference coordinate system.As a consequence,the influence of large nonlinear imaging distortion caused by rough full-field measurement is diminished,as well as avoiding sacrificing the effective spatial resolution of the acquired images.In addition,a novel correlation function for speckle stereo registration is constructed based on the multi-camera network joint constraint relationship,which can accurately restrict the search region for stereo matching homologue points between image pairs to the epipolar line instead of the entire image.Thereby,the search region is reduced and the search efficiency is improved.At the sam e time,the digital image correlation method based on multi-camera network joint constraint optimization can realize the optimization of the position and attitude parameters between cameras,and further enhance the reconstruction accuracy of the three-dimensional coordinates of the speckle points,thereby improving the measurement accuracy of the deformation information.(3)The problem of the measurement for moving target pose parameters in a large field of view that is susceptible to factors,i.e.,cumulative error of the model,imaging distortion,as well as unknown feature information is fully addressed,upon which a novel measurement method is proposed.First,a multi-source feature data fusion model suitable for visual measurement process is established to overcome the simplicity and inefficiency of feature points.Then the closed measurement mode is constructed based on bidirectional transfer feature point cloud information,which changes the unidirectional process from image data to spatial feature information transmission in the traditional mode.The confirmed spatial data is returned as the control information to the measurement process that increases the redundant information.Consequently,the proposed method enhances the stability and reliability of the calculation of aircraft pose parameters.(4)Lastly,for the redundant time-consuming problem of speckle feature matching in the visual measurement process,it is difficult to improve the solutio n speed by orders of magnitude only by the optimization and improvement of the algorithm itself.Therefore,the GPU parallel operation mode of the CUDA source program compiled by NVCC is adopted to boost the sub-pixel search efficiency for speckle matching.The GPU parallel programming method with CUDA source programs compiled by NVCC has strong flexibility,and it solves the obstacles when Mex script interacts with other language programs.At the same time,it avoids being restricted by overloaded functions,so that the overall computing performance of the deformation measurement program can reach a better level and achieve a higher speedup ratio.In addition,the CUDA heterogeneous parallel algorithm is employed to carry out three-dimensional translation measurement and dynamic three-dimensional deformation measurement experiments on the large-scale simulation wing of the ground vehicle,which obtains the preferable experimental results and lays the theoretical and technical foundations for real-time measurement and output.