Design Of Four-rotor UAV Control And Target Detection

Quadcopter UAVs rely on their flexible and free operability and strong scalability,making their application scenarios more and more abundant.At present,the application of quadrotor UAVs in target detection,real-time tracking and other aspects can greatly reduce labor costs and improve work efficiency.In order to improve the target detection ability and control stability of quadrotor UAV,this paper has completed the following main work:Firstly,according to the actual needs of the project,the software and hardware aspects are designed,the processing module,sensor,power component and image acquisition module of the UAV are selected,and the target detection system based on the quadrotor UAV flight platform is built according to the framework requirements of the target detection system,which lays a foundation for the research of subsequent control methods.Then,aiming at the problem of low detection accuracy caused by large difference in target scale,complex detection scene,small and dense target from the perspective of UAV,a real-time target detection algorithm for YOLOv5 n for aerial photography small targets is proposed.By introducing the lightweight channel attention(ECA)module in the basic model,the ability of the convolutional neural network to extract the effective information in the feature map is improved.After the output of the feature pyramid network,the adaptive feature fusion module(ASFF)is added to improve the recognition accuracy of feature maps of different sizes.The EIOU loss function is used to calculate the difference between the prediction box and the target frame,which accelerates the convergence speed and improves the detection accuracy.The detection head of YOLOv5 n is improved to optimize the detection performance of the model for small targets.After the improved object detection algorithm is verified on PC,it is deployed on a high-performance processor(Jetson nano)to verify the real-time and effectiveness of the object detection algorithm.Finally,in order to provide a stable flight target detection and control platform and reduce the external interference of the quadcopter during flight,this thesis combines the kinematic principle of quadcopter UAV to build a dynamic model in the simulation environment in MATLAB software,and determines the cascade PID controller of angle ring and angular velocity ring by analyzing the model characteristics of quadcopter UAV,on the basis of which a cascade PID controller based on extended state observer and measurement error estimator(ESO-MEE)is designed.The model simulation experiment was carried out to achieve stable attitude control,and then the hardware was debugged and the actual test of physical flight test and target detection was carried out,which verified the practicality of the quadrotor UAV target detection and control system.