Research On Helicopter Blade Flapping Distance Measurement Technology Based On Image Super-Resolution Reconstruction

Helicopter blade motion parameters measurement is an important inspection item in the development,production and maintenance process of helicopters,flapping distance is one of blade motion parameters,therefore,it has great significance to research on a high-precision flapping distance technology for ensuring the safe operation of helicopters.In this paper,the image super-resolution reconstruction method is used to develop a research of the helicopter blade flapping distance measurement technology.The main research contents and results are as follows:(1)In this paper,the motion parameters measurement system of helicopter rotor based on four-view stereo vision developed by our research group is improved,the system operability is enhanced,and the system function is perfected.In the hardware system,according to the requirements of the experimental site conditions and technical specifications,the construction of high-speed industrial cameras,computers,frequency multipliers,non-flicker photography lights,remote control computer and other hardware modules is completed.In the software system,the original system designed by our research group is improved,a software subsystem of blade motion parameter measurement is added,this subsystem includes coordinate of global coordinate system transformation,hub coordinate system establishment,coordinate of hub coordinate system transformation,blade motion parameter calculation and other functional modules,and these modules are implemented.The test results under actual measurement conditions indicate that the system performance well and satisfies the actual measurement requirement.(2)A super-resolution reconstruction method for blade image based on image internal learning is researched.The number of pixels occupied by the marker points in the blade image is not enough,resulting in low positioning accuracy,in order to solve the problem above,a super-resolution reconstruction method of the blade image based on image internal learning is studied and implemented.Firstly,a fully convolutional network model for super-resolution reconstruction is designed,and the model structure is determined experimentally.Secondly,generating a training dataset by data enhancement using a blade image to be super-resolution reconstructed.Thirdly,after training the network model with the training data set,the blade image to be reconstructed is used as a network input to generate a super-resolution image.Finally,the contrast experiment with other methods based on deep learning is carried out byusing the real blade image as the experimental data,the experimental results verify the effectiveness of the method proposed.The 3D coordinate measurement accuracy verification experiment is carried out,comared with the original blade image,in the blade image super-resolution reconstructed by this method,the average absolute error of the lateral distance of the marked points,was reduced by 0.014 mm,and the average absolute error of the longitudinal distance was reduced by 0.321 mm.The experimental results show that the proposed method satisfies the requirements of blade motion parameter measurement.(3)Helicopter blade flapping measurement technology based on image super resolution is studied.Firstly,the 3D coordinates of the marked points detected in the reconstructed blade image are transformed to a global coordinate system.Secondly,the hub coordinate system is established by global coordinates.Thirdly,the 3D coordinates of the marked points in the global coordinate system are transformed to the hub coordinate system.Finally,the blade motion parameters are measured by using the 3D coordinates of the marker points in the hub coordinate system.The large number of rotor measurement experiments carried out in the rotor test tower of a helicopter research institute are completed,the experimental results verified The feasibility and effectiveness of the method proposed.the blade flapping distance measurement method proposed is effective and has high measurement accuracy.compared with the method in literature [3],it is verified that the method can improve the accuracy of blade roll measurement.