Joint Mask R-CNN and Least Square Method for Antenna Parameters Measurement

Abstract

Traditional antenna measurement methods conducted with professional worker to climb towers tend to raise safety and inefficient concerns. This paper firstly proposes a novel antenna parameters measurement method using unmanned aerial vehicle (UAV) to solve the problem. Firstly, an antenna dataset, named UAV-Antenna, containing 10,715 images, is constructed by image capturing with the help of UAVs fly around various base stations. Secondly, UAV-Antenna is adopted to train an optimal instance segmentation model. Finally, least squares method is utilized for measuring antenna parameters. Experimental results show that the proposed method has a better performance than methods of measuring antenna parameters with practical hardware.