Attitude Measurement Of Antenna Based On The UAV Aerial Images

The pitch angle and azimuth angle of the mobile base station antenna are the most important physical engineering parameters,which have a decisive influence on the electromagnetic coverage of the base station antenna.Therefore,operators need to measure these two parameters of the base station antenna from time to time to ensure the correct electromagnetic coverage of the antenna.However,conventional measurement methods and current research have obvious defects.Conventional measurement methods generally use dinometer,mechanical compasses,antenna alignment tool or other measuring tools to perform antenna attitude measurement.The use of these tools inevitably requires manual climbing and contact with the antenna,thus presenting a huge safety hazard.The dinometer and mechanical compasses are complicated to operate,and the inaccurate reading is easy to cause a large error.The antenna alignment tool needs to be fixed on the antenna during operation,so the operation is complicated as well,and it is expensive.In order to avoid safety hazards,researchers have begun to pay close attention to non-contact attitude measurement methods based on computer vision.At present,the measurement methods based on computer vision mainly include two kinds of measurement methods: monocular vision and binocular vision,and the calculation accuracy of monocular vision and the robustness to error are better than binocular vision,but whether it is monocular or binocular vision,it is necessary to perform camera calibration in advance,that is,to solve the internal and external parameters of the camera.Therefore,these two methods are not universal.The 3D reconstruction algorithm can estimate the internal and external parameters of the camera based on the sequence image,and then obtain the 3D information of the target through the matching of the feature points to solve the pose.However,the running time of the 3D reconstruction algorithm is too long,and the failure rate of outdoor reconstruction is large.Since this method estimates the internal and external parameters of the camera,the method produces a large error sometimes and thus cannot be put into practical use.Aiming at various problems in the field of antenna attitude measurement,this thesis studies and proposes a non-contact and real-time antenna attitude measurement scheme based on aerial image of UAV.The scheme first obtains the antenna image during the aerial photography of the UAV and the attitude data of the UAV at the corresponding moment,at the same time the measurement personnel operates the UAV based on the aerial image and determines the state of the antenna in the current image.If the state of the antenna in the image is determined to be measurable at a certain moment,the characteristic line representing the current antenna is extracted based on the appearance characteristics of the antenna itself,and finally the pitch angle and the azimuth angle of the antenna are calculated according to the characteristic line combined with the posture of the UAV.In addition,this thesis adopts the short-term target tracking algorithm to track the target antenna which selected by the surveyor initially during the single-operation process,so that even if the UAV is affected by the wind,fluttering in the air and causing the picture to shake,the smallest region where the target antenna is located can be automatically determined.In this way,the influence of the wind on the operation is avoided.Therefore,compared with the existing antenna attitude measurement method,the main innovation of this scheme lies in the noncontact,real-time and high efficiency of the measurement scheme.In this thesis,we use a self-built antenna to design and carry out multiple experiments from multiple perspectives,and make comparative analyses based on the experimental data.The experimental results verify the feasibility,accuracy and stability of the antenna attitude measurement scheme.