Algorithm Performance Evaluation Of Aircraft Position And Pose Photoelectric Measurement

In recent years,with the increasing attention to computer vision and pose measurement algorithms,they have been widely used in navigation technology such as shipboard helicopter landing,UAV take-off and landing,aircraft condition monitoring and so on.With more and more pose measurement algorithms proposed and improved,it is particularly important to analyze the performance of the measurement algorithm and select the correct and effective algorithm.Good algorithms must meet the requirements of functional and efficient performance that can be performed correctly.In this thesis,a general framework is designed and implemented for performance evaluation of pose measurement algorithms,the purpose of which is to enable the staff to analyze the performance and the practicability of the algorithm quickly and effectively without knowing the implementation details of the algorithm,and to select the appropriate measurement algorithm.Firstly,the research background and significance of the algorithm evaluation software for aircraft pose measurement are introduced,and the application status of the performance evaluation index of the measurement algorithm in recent years is introduced in detail.The principle of vision measurement and the design basis of system requirements are expounded.According to the theoretical basis of precision,uncertainty and complexity,the performance evaluation index of the algorithm is analyzed.Secondly,based on the visual GUI interface platform provided by MATLAB,a set of evaluation system with human-machine interface is designed.Through the modules of model loading,environment parameter setting,analysis report setting,precision analysis,uncertainty analysis and complexity analysis,the detailed modular design of the system is carried out.Based on the theoretical basis of precision,uncertainty and complexity,and the functions needed to be realized,the design flow of sub-modules are described in detail.Thirdly,the simulation model and measurement model are designed by evaluating the requirements of the software for the example algorithm.This paper introduces the design principle of monocular vision measurement algorithm and the functions needed to realize the measurement algorithm based on OpenCV vision library,which provides a clear idea for the programming of the sample program of position and posemeasurement algorithm.At the same time,the example algorithm program is encapsulated in the form of loading dynamic library DLL.The interface of the software call measurement algorithm is designed to make the software system more standardized and reasonable.At the end of this thesis,the function of the evaluation system is designed.By setting the input parameters and the system parameters,the simulation model and measurement algorithms are called respectively and tested from three sub-modules:precision,uncertainty and complexity.According to the test results of each function module,the performance of the measurement algorithm is analyzed to verify the function of the evaluation software.