Research On Mobile Target Tracking System Of UAV Based On Machine Vision

The identification and tracking technology for specific targets in the mobile state is of great significance for industrial scene monitoring,intelligent transportation system management,power maintenance and even military applications.The drone's compact size,remote control and low cost make it possible to perform complex tasks in areas that are inaccessible to humans.The rapid development of machine vision technology has promoted the development of tracking technology with drones as the core.Combining drones and machine vision technology to realize the identification and autonomous tracking of UAVs for specific targets will have broad application prospects.In this paper,the tracking of the ground moving target by the drone is taken as the research content.The drone can capture the target in the field of view and extract the information in real time,and solve the relative positional relationship between the drone and the target,and then combine the inertia of the body.The attitude information obtained by the navigation module is sent out,and an adjustment command is issued to control the attitude and flight path of the aircraft to achieve the effect of autonomously tracking the target.The main work of this paper is the communication between the visual unit of the machine and the hardware unit of the drone,the software development of the drone unit,the software development of the machine vision unit,the machine vision unit and the flight control unit.The machine vision unit software design adopts the AprilTag algorithm to optimize the algorithm according to the actual application scenario,which meets the functional requirements of the system while reducing the performance requirements of the control chip.The design of the flight control software adopts the double closed loop PID control algorithm,combined with the attitude calculation based on inertial navigation to ensure that the drone can fly stably and reliably.This paper independently developed a four-rotor UAV flight control unit and machine vision unit based on embedded microprocessor,which realized the tracking of moving targets and carried out a lot of experimental tests.The experimental results show that the UAV tracking system designed in this paper has high stability and can accurately identify moving targets and track them autonomously under different scene environments and target occlusions.