Research On UAV Monitoring Algorithm Based On Image Transmission Signal

In recent years,with the development of civilian UAVs,the illegal flying UAVs have caused a series of safety problems.How to effectively implement UAV monitoring is an urgent problem to be solved.Passive monitoring has the advantages of not emitting electromagnetic waves,not affected by factors such as the shape and color of the monitoring target,etc.It has been widely used in the field of UAVs monitoring in recent years.At present,the passive positioning method of the array antenna has high equipment complexity and limited application.The image positioning technology based on the Perspective-n-Point(Pn P)problem is weak against outliers,and needs to know the camera focal length in advance,which cannot be well adapted to the application scenarios of non-partner UAV monitoring.This thesis aims at the problems of the existing Pn P positioning algorithm,starting from improving the anti-outlier capability of the known focal length algorithm and improving the stability and solution speed of the unknown focal length algorithm,a deep research on Pn P-based image location technology was conducted,the main research as follows:1.According to the positioning requirements of UAV monitoring,an overall monitoring scheme based on Pn P problems was designed.The data for Pn P was obtained by capturing and analyzing the image signal from UVA.According to the characteristics of the data type and transmission protocol of the UAV image transmission signal,under the Linux system,a scheme for capturing and analyzing the UAV image transmission signal using Wireshark + wireless network card is proposed.Based on the camera imaging model,the relationship between the camera world system and the world coordinate system is described,which lays a theoretical foundation for the subsequent research on Pn P related algorithms.2.Under the condition that the focal length is known,a binary weighted algorithm based on orthogonal iteration is proposed to solve the problem that the existing algorithm is weak against outliers.Considering the matching situation between the UAV image and the indeed 3D information,on the base of the orthogonal iterative algorithm,a binary weighted system is introduced to determine the reliability of the reference point,and a binary weighted Pn P algorithm based on orthogonal iterative is proposed.The algorithm selects the maximum value of the quartile of the reprojection error as one of the thresholds,and at the same time introduces a threshold that is proportional to the maximum reprojection error and inversely proportional to the focal length to accelerate the convergence rate of the algorithm when the outlier rate is low.Taking the larger of the two thresholds as the final threshold to weight all reference points,which reduces the effect of outliers on the objective function and improves the stability of the algorithm.The simulation verifies the feasibility of the algorithm.3.In the case of unknown focal length,aiming at the problems of low stability and slow running speed of existing algorithms,a perpendicularity projection model is established,and a PNP algorithm based on line and plane characteristics is proposed.On the base of the original imaging model,the 3D reference point was mapped vertically to the imaging surface,and the relationship between the 3D reference point and the 2D image point was established through the proportional relationship model.Further,the solution of the rotation matrix was constructed as a polynomial system without the unknown focal length using the characteristics of the line and the flat surface.In this algorithm,the polynomial problem was transformed into the sparse matrix decomposition problem by using the eigenvalue method,which improves the efficiency of pose estimation algorithm.Through vertical projection of 3D reference point,the influence of unknown focal length is eliminated and high stability was obtained.The feasibility of the algorithm was verified by experiments on both synthetic data and real image.