Research on Sensor Data Spatiotemporal Synchronization Technology for Onboard Visual Navigation Systems

Against the backdrop of the rapid development of unmanned aerial vehicle (UAV) technology, onboard visual navigation systems have become the core support technology for achieving precise positioning and efficient map construction of UAVs. This paper deeply explores the key technologies of sensor data spatiotemporal synchronization in onboard visual navigation systems, aiming to solve the synchronization issues in multi-sensor data fusion in variable environments, and enhance the navigation accuracy and reliability of UAVs. The research covers both spatial and temporal synchronization dimensions, with a focus on the technical implementation of Visual-Inertial Odometry (VIO). For spatial synchronization, a target-based method is proposed to construct an error equation for solving the extrinsic parameters between the camera and the Inertial Measurement Unit (IMU), achieving precise alignment of sensor coordinate systems through the identification and matching of visual target feature points. In terms of temporal synchronization, considering the data frequency differences between the camera and IMU, this paper designs a time alignment strategy that realizes efficient temporal alignment of data through data buffering and timestamp matching. Additionally, the initialization and local sliding window optimization process of the Visual-Inertial Odometry are detailed, achieving high-precision fusion of visual and inertial data through feature extraction, optical flow tracking, and motion structure recovery. Experimental results show that the methods proposed in this paper can technically provide strong support for the autonomous flight of UAVs, effectively improving the navigation performance of UAVs in complex environments, and enhancing the robustness of the system.