自动化装船场景下船体实时跟踪系统设计与应用

Design and Application of Real-Time Ship Hull Tracking System for Unmanned Loading Scenarios

摘要: 在无人装船作业中，高精度船体位姿跟踪是保障溜筒精准对位与作业安全的关键。针对传统PLC数据点云对齐方法精度有限的问题，本文提出一种基于预建船体地图的两阶段实时船体跟踪方法。该方法首先构建全局地图，装船时利用PLC数据为新点云提供初始位姿进行粗定位，再驱动ICP算法与地图进行精配准，实时解算船体精确位姿。实验表明，该方法在整个装船测试流程中可以实现稳定、可靠的跟踪效果，为无人装船系统的智能控制提供了稳定可靠的空间基准。

Abstract: In unmanned ship loading, accurate real-time tracking of the ship’s 6-DOF pose is vital for safe, collision-free, and precise chute positioning. To overcome the limited accuracy of PLC-only kinematic methods, we propose a robust two-stage approach: first, PLC encoder data provides a coarse initial alignment of each new LiDAR scan to a pre-built 3D ship hull map; then, the ICP algorithm refines this alignment to achieve high-precision pose estimation. Field tests at a working port show the method delivers stable and reliable tracking throughout the entire loading cycle while the ship remains docked—and intentionally fails when the ship departs, confirming its responsiveness to real conditions. This provides a dependable, real-time spatial reference essential for intelligent control of unmanned loading systems.