关键词: 智能行车系统, 机器人操作系统, ORB-SLAM3算法, 深度相机

Abstract: In view of the technical bottlenecks of high deployment complexity and poor portability faced by existing intelligent crane systems in industrial scenarios, this study proposes an intelligent crane system based on the Robot Operating System (ROS) and the ORB-SLAM3 algorithm. Through functional analysis, the overall architecture of the intelligent crane system is constructed. Specifically, a depth camera is first utilized to capture real-time environmental information of material piles. Subsequently, the ORB-SLAM3 algorithm is optimized by incorporating a dense mapping thread and a safe material-picking thread, thereby enabling real-time crane positioning, three-dimensional reconstruction of the material storage area, and planning of target material-picking points. Finally, by leveraging the standardized motion control interface of ROS, the stable and reliable operation of the crane is ensured. Simulation experiments demonstrate that the proposed intelligent crane system achieves high accuracy in positioning and three-dimensional reconstruction, and is capable of generating safe and efficient material-picking strategies in real time. Furthermore, the distributed architecture of ROS and the rapid deployment of the depth camera contribute to reducing deployment complexity and enhancing the portability of the intelligent crane system. The research findings are anticipated to provide a feasible technical approach for addressing the long-standing issues of high deployment complexity and poor portability within the industry.

Key words: intelligent driving system, Robot Operating System (ROS), ORB-SLAM3 algorithm, depth camera