| In order to ensure the safe landing of the aircraft,airports at home and abroad often use multiple cameras to monitor the landing process of the aircraft.However,the current aircraft landing surveillance is limited to the pixel level,which means cameras are independent and unrelated,and they need to be manually switched in multiple pictures to judge the landing status of the aircraft,as well as the potential danger to the aircraft cannot be alarmed automatically.Aiming at this problem,this thesis proposes a multi-camera collaboration method for aircraft monitoring,which realizes the motion state measurement and attitude estimation of the aircraft during landing,and provides a guarantee for the safe landing of the aircraft.The main work carried out in this thesis is as follows.Firstly,a moving target detection and tracking algorithm for a single camera static background is studied.In view of aircraft detection,this thesis studies detection principles of inter-frame difference method,three-frame difference method and background difference method,and then proposes an improved three-frame difference method as the aircraft target detection algorithm of this study.Compared with the previous detection algorithm,this improved three-frame difference method can achieve a more complete aircraft profile and more accurate aircraft positioning.Meanwhile,this thesis deeply studies the principle of KCF target tracking algorithm,and verifies that the algorithm can simultaneously meet real-time and accuracy requirements in aircraft target tracking.Secondly,a method based on video image for aircraft motion state measurement and attitude estimation is studied.In this thesis,the spatial position of the aircraft is obtained by solving the perspective projection equation under additional constraints,and the motion velocity of the aircraft is further estimated.The theoretical error of the algorithm is analyzed,as well as the theoretical error of the algorithm is analyzed.For the attitude estimation of the aircraft,this thesis proposes a method for solving the attitude of the aircraft based on the calibration image after studying the SFM attitude solving algorithm.The simulation results show that the proposed aircraft motion state measurement method and aircraft attitude estimation method in this thesis both meet the monitoring accuracy requirements.Finally,aiming at solving the continuous tracking problem of multi-camera aircraft,this thesis proposes a multi-camera collaboration tracking strategy on the basis of the visual field boundary line,which can achieve continuous tracking of the aircraft by multicamera.In addition,in order to obtain the global position of the aircraft,this thesis studies the image mosaic fusion method based on the visual field boundary line and establishes the global coordinate system of the airport runway.Then,for the case of single camera position solution error,this thesis studies the positioning error correction method on the basis of binocular ranging principle.The simulation results show that the error correction effect of the method is positive. |
