智能感知融合技术在干散货堆场动态环境中的应用

Application of intelligent perception fusion technology in dynamic environment of dry bulk yard

以干散货码头堆场为代表,针对其大型设备自动化作业在堆场复杂动态环境的感知问题,提出多传感器智能感知融合技术框架,基于激光雷达、惯性传感器(IMU)和实时动态(RTK)等多种传感器数据融合,利用原始的RTK数据信息对前端里程计输出的位姿进行优化,应用动态物体移除技术实现稳健、精准的动态要素感知,结合八叉树立方体查询实现高效、实时的全局地图更新,最终建立一个高精度且实时更新的散货堆场地图,为后续的堆场作业提供支持。将智能感知融合技术应用于大型干散货码头,结果证明该技术可实现堆场动态环境的高精度和实时性能,有效提高干散货码头自动化作业的感知能力。

This article takes the dry bulk terminal yard as a representative and proposes a multi-sensor intelligent perception fusion technology framework for the perception problem of large-scale equipment automated operations in the complex dynamic environment of the yard.Based on the fusion of multiple sensors such as LiDAR,inertial sensor(IMU),and real time kinematic(RTK),the article uses the original RTK data information to optimize the bit pose of front-end odometer output.It also uses dynamic object removal technology to achieve robust and accurate dynamic element sensing and utilizes the octree cube query to achieve efficient and real-time global map updates to finally establish a highly accurate and up-to-date bulk yard map and support the subsequent yard operations.The intelligent perception fusion technology framework is applied to a real-world large dry bulk terminal.The results prove that the technology can achieve its high accuracy and real-time performance effect for the yard s dynamic environment,effectively improving the perception ability of automated operations in dry bulk terminals.