Methodoligy Of Atitude Measurements For Aircrafts In Shooting Range Based On Computer Vision Technique

It is a very important value that tracking and video recording attitude flight of aircraft to obtain three-dimensional attitude of aircraft flight,not only is an important basis for the evaluation of aircraft flight stability,and analysis of aircraft fault aerodynamic shape modeling,and improve the aircraft launch and tracking system in the range of experiment.In this paper,based on the background of the experimental task of measuring the attitude parameters of the space flying object with large aperture electro-optical theodolite,the method of high accuracy attitude measurement of spatial object is studied on the machine vision algorithm.This paper establishes the long baseline distance of theodolite imaging model,deduces the conversion relationship between photography coordinate system and theodolite coordinate system,introduces the definition of the attitude parameters,proposed a simplified model for solving attitude parameters.In view of the image quality directly to the success of feature detection,the paper first carriees out the study of image preprocessing techniques.It analyzes the reasons of theodolite image distortion,uses radial distortion correction and linear optimization of the basic matrix to complete camera calibration experiment,and better to eliminate the geometric distortion of the image for electro-optical theodolite;In view of the interference stripe,the rectangular low-pass filtering algorithm is proposed,which greatly improves the image signal to noise ratio.The above methods can improve the image quality effectively,and make a good foundation for the attitude vision measurement.The key technology of aircraft attitude measurement is the feature extraction of the target.It not only directly affects the accuracy of the subsequent attitude measurement,but also the attitude measurement system is effective.In order to improve the accuracy and save resources,the traditional speed up robust features(Speed Up Robust Features,SURF)operator to make improvements,improve the robustness of the geometric deformation and the position error and matching speed.Based on the feature points,a high precision attitude solving model of constructing optimization function is proposed,and the accuracy of attitude measurement is analyzed in detail.In order to further enhance the robustness of the algorithm,in view of the range target,put forward the line feature extraction algorithm based on kernel density estimation,using the straight line segment features as a constraint optimization to obtain target attitude parameters.The aircraft attitude accuracy of vision measurement in the range is a very important technical index of system.To evaluate the influence of the algorithm error on the whole system is a difficult problem,the difficulty comes from the uncertainty of error sources,the current the research methods are not many,carry out Monte Carlo methods for estimation of the measurement error estimation algorithm based on visual attitude.Theodolite shooting,aircraft image recording,image preprocessing by removing the distortion,noise,using multi-scale order two descriptors for target feature extraction,the method of constructing optimal attitude characteristic function and line constraint.The aircraft images shot and recorded by theodolite wihich are preprocessed to preprocess the distortion,noise,are extracted target feature by using multi-scale second-order descriptors.This paper estimates target attitude parameters by constructing optimal attitude characteristic function and line constraint.Shoot test results show that in the target imaging pixel number is more than 500 pixels,measurement accuracy is less than 0.1°.