Research On The Situation Surveillance Of Reposition And The Technology Of Collision Avoidance

Airport traffic dispatch and monitoring system,which plays an important role in maintaining the normal operation of the airport,is mainly responsible for surveillance of traffic dispatch situation and real-time computation of the minimum separation between aircrafts to prevent collision during the traffic dispatching process.In this paper,due to the massiveness and complication of dispatch system,we chiefly focus on its four parts as followed: tracking algorithm for airplane target in the airport,the error correction of the monitoring camera generated from aircraft movement,the initial pose estimation of airplane target based on CAD model,and the traffic collision avoidance algorithm.1.In this paper,concerning the key issues in airport tracking,a high-speed scale adaptive kernelled correlation filtering(KCF)algorithm is put forward,which can realize the real-time and robust target tracking in difficult conditions with illumination variation,weak light and the obstruction of similar targets.2.In the airport,the internal and external parameters of the monitoring camera vary with the existence of vibration caused by aircraft motion.To tackle this problem,the physical simulation and computer simulation experiments are carried out,in which a variety of iterative algorithms are employed to correct camera pose errors.Meanwhile,the influences of factors including the number of control points,the extracting precision and the focal length of the camera on the different algorithms are discussed as well.3.As the simulation library with a true value of pose is obtained from 360 degree field of view projection of the CAD model,the pose estimation is successfully transformed into the matter of 2D-3D alignment via the target pose estimation algorithm based on the CAD model,the wrinkle/ major concern of which lies in how to align the simulated image with the actual image.In this thesis,the method proposed by Aubry Mathieu is refined.To begin with,the model base is classified into several categories,then optimized by the “coarse to fine” alignment strategy.Finally,the effectiveness of this method is verified by researching on various complex scenarios with airplanes included.4.Considering the main function of the airport collision avoidance system is providing real-time early warning,this paper presents a simplified algorithm for the shortest distance between the convex hulls of objects based on the saliency detection,which brings about the calculation of real-time minimum distance between the target airplane and other aircrafts or its surrounding obstacles in image.