Ship Target Recognition and Positioning Based on Aerial Photos

In order to obtain the ship information in the domestic canals and territorial sea more conveniently, a method of acquiring ship type information and geographical position information by aerial photography is proposed. In this paper, the YOLOv4 target detection algorithm is used to identify ships. The method of data enhancement and adversarial training is used to improve the robustness of ship target detection model. According to the data of the inertial navigation system and the attitude angle data of the high-definition camera, the pixel coordinates of the ship target are converted to the longitude and latitude coordinates of WGS-84. The positioning experiment of ship target on water surface is designed. And then the positioning experimental data were calculated and analyzed. The positioning accuracy of ship target is improved by analyzing the error factors of experiment. Experiments show that the proposed method can meet the requirements of practical work and lay a foundation for fast and accurate acquisition of ship information.