Research On Key Technologies Of UAV Monitoring Software System

UAV is a remote control by radio for an aircraft,having a flexible,low cost and so on.Currently,UAV plays an important role in all areas of military and civilian use.In order to sense and collect external information during the flight,UAV is often equipped with equipment such as sonar,laser,and visual sensors.Compared to other,the information collected by the vision sensor is richer and more intuitive,enabling realtime monitoring of targets in the scene.The research on the UAV ground moving target monitoring system has always been a multi-disciplinary application research topic.This paper focuses on the development of computer vision technology,and selects relevant algorithms to solve the problem of moving target detection and target tracking.The actual algorithm of UAV monitoring is improved.At the end of this paper,the UAV monitoring system is designed and implemented.The specific research includes:(1)The background compensation difference method is used to solve the moving target detection problem of the UAV monitoring system.Because of the relative motion between the airborne camera and the monitoring picture,this paper uses the affine motion parameter model to model the background motion of the background of the monitoring picture.In order to accurately estimate the background motion parameters,the Harris corner detection algorithm and the diamond matching rule are used to perform feature extraction and feature point matching on adjacent frame images in the monitoring image.For the error points in the matching point pairs,the distance statistical histogram is used.The form is culled,and then the background motion parameters of the monitor picture are estimated by the least squares method.Finally,the background motion parameter is used to perform background compensation on the reference frame and then differentiate from the current frame to achieve the purpose of detecting the moving target in the monitoring screen.Experimental results show that the algorithm can effectively detect moving targets in the video stream.(2)In order to adapt to the target scale change in the target tracking process of the UAV,this paper selects the DSST algorithm for target tracking.Aiming at the problem that the target may be occluded during the tracking process,firstly,the mapping relationship between image plane coordinates and space coordinate system coordinates is established,and then the motion state of the target in the space coordinate system is predicted by Kalman filter.In the target tracking of image plane,this paper uses APCE parameters to measure the target tracking result and adopts a high confidence update strategy.If the target does not occlude,update the DSST tracking model and convert the tracking result to the spatial coordinate system to update Kalman.Filter;if the target is occluded,the current tracking model is retained,and the predicted value of the Kalman filter is converted to the image plane as the tracking result of the current frame and the input parameter of the next frame of the DSST algorithm.Experiments prove the effectiveness of the algorithm.(3)The UAV monitoring system was modularized and implemented using technologies such as QT and OpenCV.