Investigación del método de localización y seguimiento de objetivos con lidar de recuento de fotones únicos en...

Single-photon counting lidar exhibits distinct advantages in detection sensitivity and spatiotemporal resolution, making it crucial for precision target detection. However, laser point clouds are susceptible to noise interference due to the high sensitivity of single-photon detectors (SPDs), and achieving stable target tracking still requires further research. This paper proposes a dual-branch tracking method (DBTM) for single-photon lidar target detection. First, a multi-layer denoising approach is performed in a coarse-to-fine manner to filter out noise point-clouds, separating the echo data of the moving target from the background interference. Subsequently, a dual-branch bidirectional guidance technique is designed, with a parameter estimation method integrated into the adaptive kernel Kalman filtering algorithm to estimate the target position information. This strategy fully utilizes both target intensity and depth information under typical water mist interference conditions. Simulation results show that the DBTM can still track the target effectively and accurately even under a photon noise rate of 1 MHz. Field experiments demonstrate the real-time ranging and tracking of a ship model at a distance of 84 m under water mist interference conditions, with the positioning error of 0.161 m, which verifies the effectiveness of DBTM. Performance analyses confirm that the DBTM outperforms typical conventional approaches in trajectory fitting, error control, and positioning accuracy. Moreover, the dual-branch structure supports adaptation to diverse feature scenarios and holds significant practical value for the precise positioning of small moving objects at a long range in complex environments.