Fast Visual Localization Algorithm Based on Nearest Neighbor Search

The primary means of positioning for unmanned aerial vehicles (UAVs) relies heavily on the Global Navigation Satellite System (GNSS). However, in certain extreme scenarios, GNSS signals can become unreliable and unstable. To ensure uninterrupted positioning functionality, a supplementary method is indispensable. Visual positioning, with its robust resilience to interference, offers a viable solution by enabling positioning under such challenging conditions. By correlating the UAV's onboard downward-facing images with satellite maps, we can accurately pinpoint the coordinate position of the center point within the UAV image on the satellite map. Subsequently, utilizing the latitude and longitude data extracted from the satellite map, we can calculate the precise location coordinates of the UAV. Nevertheless, prevailing visual positioning algorithms typically suffer from slow processing speeds, which compromises real-time performance and significantly impacts the overall positioning accuracy. In response to this challenge, this research paper introduces an innovative approach that employs the nearest neighbor search method to refine the process of filtering and matching satellite maps, thereby accelerating the positioning speed.