LTA-OM: Long-Term Association LiDAR-IMU Odometry and Mapping
Vložit
- čas přidán 10. 09. 2024
- LTA-OM [JFR2024]: an efficient, robust, and accurate SLAM system. We integrate FAST-LIO2, STD-LCD (Loop Closure Detection), loop optimization and false positive loop closures rejection into LTA-OM. Long-term association (LTA) mapping is proposed to employ the corrected history map to provide direct global constraints to the LIO mapping process, resulting in nice global map consistency. A multi-session mode is also implemented and this mode provides consistent map stitching performance. LTA-OM shows superior accuracy, map consistency and efficiency in the benchmark experiment against existing SLAM systems. It is applicable to both multi-line spinning (360 degrees FOV) and solid-state LiDARs (small FOV). Code is available on github.com/hku....