Tracking and Monitoring Method for Multi-Target Moving Ships Based on High-Resolution Remote Sensing Images

To address the problem of ship target probing and detection in vast sea areas, a new method is proposed to realize rapid, positioning identification, and motion tracking of multitarget moving ships by comprehensively utilizing the characteristics and advantages of satellites in orbits of different altitudes. According to the characteristics of ship targets and the prechecking of geostationary orbit satellites, low-earth orbit satellites are deployed to coordinate investigation and capture high-resolution remote sensing images, which can be used to quickly designate target candidate areas, effectively improve the search efficiency, and achieve accurate positioning of ship targets and fine-grained recognition and classification of “Ship-Category-Model”. The DeepSORT algorithm was improved to use a single small neural network structure by adding the functions of speed and course estimation, track plotting, and dynamic management, and it was used to extract the apparent features of ship targets and integrate them with common position information to participate in the inter-frame data association. Through satellite video inspection and simulation experiments, the feasibility and robustness of accurate and stable tracking in remote sensing scenarios and multi-target moving ship tracking and detection tasks were verified by mining and utilizing the advantages of high-resolution images.