Research On UAV Target Detection Tracking And Remote Monitoring System

In recent years,small fixed-wing unmanned aerial vehicles(UAVs)have been widely used in security patrol tasks.Small fixed-wing UAVs equipped with intelligent visual tracking algorithms have become the preferred choice for autonomously and efficiently completing covert reconnaissance,emergency tracking,personnel control,and other urgent and dangerous tasks.However,existing small fixed-wing UAVs based on visual target capture and monitoring have problems such as low autonomy,high equipment cost,fixed application scenarios,and slow algorithm inference speed.Therefore,this paper proposes an efficient target detection and tracking algorithm by building a fixed-wing UAV experimental platform,and based on this method,a UAV target remote monitoring system is designed and implemented using remote MQTT communication,achieving remote automatic target detection,tracking,and control of UAVs.Through real-world UAV flight experiments to verify the feasible remote inspection solution proposed in this paper,the research content and research results of this paper are as follows:(1)In view of the shortcomings of existing target detection and tracking methods,the construction,improvement,deployment,and application of target detection and tracking methods were studied.Based on deep learning methods,a high frame rate and high-precision tightly coupled target detection and tracking method for small fixedwing UAVs,namely the YOLO-KCF algorithm,is proposed.Firstly,the joint region calculation is optimized to improve the detection success rate and accuracy.Secondly,a dual-thread tracking mechanism is proposed to optimize the use of onboard computer performance.Finally,the use of frame difference verification improves the tracking response speed and detection and tracking effect.The experimental results show that the autonomous detection accuracy of ground stationary or moving targets in complex backgrounds can reach 85.7% with an average FPS of 23.6,and the reaction speed can reach the millisecond level,realizing tracking and control with a collision accuracy of0.95.The proposed method has certain theoretical and practical significance in the field of target detection and tracking,and provides powerful technical support for the practical application of fixed-wing UAVs.(2)To realize the application of the tightly coupled target detection and tracking algorithm YOLO-KCF algorithm under remote distance conditions,a communication method for UAV remote monitoring is designed and studied.In terms of communication methods,the MQTT communication protocol is selected,and the message server framework is designed.By using the message proxy method,the publish-subscribe mode of information exchange between UAVs and the monitoring system is realized.The design of the UAV’s automatic communication mechanism realizes the coordinated operation of the UAV’s messages and the remote monitoring system’s internal and external fields,thereby ensuring the real-time communication capability required by this paper’s method.The information structure design of physical objects is implemented for efficient data storage and concise data management.(3)Combining the target detection and tracking algorithm and remote communication means,a UAV remote monitoring system software is designed in a targeted manner.After the requirement analysis,specific design is carried out for each module of the system,including the system main interface module,status communication module,image display module,map waypoint planning module,data management system module,real-time target detection and tracking module,and other functional modules.The functions of each module of the remote monitoring system proposed in this paper and the remote communication control of UAVs are fully implemented.