Abstract: With the widespread application of low-slow-small (LSS) targets such as drones in many lowaltitude airspace, LSS targets have gradually become a significant threat to low-altitude security. In response to the requirements for real-time and accurate target identification in the low-altitude airspace, a trajectory extraction algorithm for LSS targets based on single-photon lidar is proposed in this work. Single-photon lidar, with its photon-level detection sensitivity, has been widely used in fields such as target ranging. This algorithm utilizes the target distance information measured by lidar to transform the trajectory extraction problem into a line fitting problem based on the random sample consensus (RANSAC) algorithm, and then adaptively optimizes the RANSAC algorithm to extract target trajectories from a large number of echo point clouds. Experimental results show that when the target's movement speed varies between 45 m·s −1 and 75 m·s −1 , and the background noise intensity varies between 500 counts·s −1 and 3500 counts·s −1 , the algorithm demonstrates good adaptability, with errors always being controlled within 7 m and runtime within 75 ms. This method provides a feasible technical solution for real-time and accurate extraction of LSS target trajectories based on single-photon lidar.

Key words: laser techniques, single-photon lidar, low-slow-small target, RANSAC algorithm, trajectory extraction