Study On The Stereo Matching Method For Speckle Measurement Of Aero-engine Exhaust Flow Field

With the continuous development of civil aviation industry,the safety of aircraft navigation has received more and more attention.The data show that the fault caused by aero-engine is one of the important factors that affect the flight safety.The measurement of the exhaust flow field of the aero-engine can realize the monitoring of the engine running state,and then predict the safety hazards of the aircraft engine in advance.The subject “ Study on the stereo matching method for speckle measurement of aero-engine exhaust flow field” is to study a speckle measurement stereo matching method to meet the measurement requirements of the aerodynamic exhaust flow field environment based on the existing visual measurement research.Based on this demand,the main research contents of the paper include:1.Combining the measurement principle of the binocular vision measurement system and its limitations in the application of aero-engine exhaust flow field measurement,This paper studies the virtual stereo vision measurement system,which simplifies the acquisition process and solves the problem of shooting nonsynchronization in the binocular vision measurement system.2.In the calibration process of virtual stereo vision measurement system,The paper proposes a camera internal reference calibration method by using BP(Back Propagation)neural network.This method combines the nonlinear mapping ability of BP neural network,and the high-precision calibration of the camera's internal parameters was achieved by training the distortion at the calibration point,and experimental verification was performed.3.In the process of speckle measurement,the paper proposes a stereo matching method for speckle measurement by using the improved KLT(Kanade-Lucas-Tomasi)algorithm.This method is based on the KLT algorithm which mainly used in the video interframe processing,and uses the image pyramid to realize the improvement of the KLT algorithm.At the same time,in order to optimize the stereo matching effect,the MBE(Maximum Bidirectional Error)and external polar constraints are adopted to establishing a mismatched deletion mechanism.Thus,the high accuracy and high efficiency matching of speckle measurement is realized,and experimental verification is carried out.4.The virtual stereo vision measurement system is verified by experiment,experiment results show that the measuring system principle is correct,and the speckle measurement stereo matching method by using the improved KLT algorithm has higher accuracy,besides,the vector of the flow field plotted according to the matching results is in line with expectations,which further verifies the accuracy of the stereo matching method for speckle measurement.