A LiDAR–Camera Spatiotemporal Synchronization Method for Unmanned Aerial Vehicle-Based Ground Target Perception

Unmanned aerial vehicles (UAVs) have been widely adopted in military reconnaissance, urban management, and agricultural monitoring. However, the increasing complexity of real-world environments, characterized by dynamic targets and variable ambient conditions, presents significant challenges for multisensor data fusion. While integrating heterogeneous sensors such as light detection and ranging (LiDAR) and cameras enhances ground target perception, differences in sampling rates and sensor characteristics often lead to spatial and temporal misalignment. Traditional synchronization methods, including static extrinsic calibration and hardware-based triggers such as inertial measurement units (IMUs) or global navigation satellite system (GNSS), struggle to adapt to rapidly changing scenarios and complex backgrounds. To address these challenges, we propose a novel and cost-effective spatiotemporal synchronization method that leverages prominent ground features—such as terrain contours and building outlines—as reference benchmarks, combined with a timestamp-driven pulse width modulation (TSD-PWM) sampling technique. This approach eliminates the need for additional synchronization hardware while ensuring adaptability across diverse environments, ranging from dense urban landscapes with intricate architectural structures to rural areas with natural terrain variations. Extensive real-world experiments demonstrate that our method improves time synchronization accuracy by 82.65% over interpolation-based techniques and enhances spatial synchronization accuracy by 74.23% compared to reprojection methods based on static extrinsic parameters. These results highlight the effectiveness of our approach in mitigating spatiotemporal asynchrony in UAV-based ground target perception using heterogeneous sensors.