基于视觉的集装箱海上并靠吊装位姿识别

舰船在远海进行并靠补给时受波浪影响两船之间会产生无规律的相对运动，给补给集装箱吊装工作带来了很大挑战，急需一种能够在较高海况下辅助人工的自动化吊装系统。针对这一问题，本文提出了一种基于单目视觉和ArUco标志点相结合的集装箱位姿检测方法。首先，介绍了带有视觉系统的并联吊装系统主要组成部分，简要分析了控制流程及方法；然后，根据实际集装箱码放流程，设计了适用的ArUco标志点布放和检测方案，介绍了相机标定、ArUco标志点检测及集装箱位姿估计方法；最后，搭建模拟试验平台测试视觉位姿识别的精度、速度和不同光照条件下的鲁棒性。试验结果表明，在普通工控PC平台上，最低识别速度在15Hz以上，姿态角平均误差为 ，位移量平均误差为 ，可以满足并靠吊装要求。

Vision-based Container Posture Recognition of Side-by-side Replenishment Ships at Sea

When ships are operating in the open sea and are affected by waves when they are replenished, there will be irregular relative movement between the two ships, which brings great challenges to the lifting of supply containers. It is urgently needed to be able to assist humans in higher sea conditions Automatic lifting system. Aiming at this problem, this paper proposes a container pose detection method based on a combination of monocular vision and ArUco landmarks. First, the main components of a parallel hoisting system with a vision system are introduced, and the control process and method are briefly analyzed. Then, according to the actual container stacking process, an applicable ArUco landmark placement and detection scheme is designed. ArUco landmark detection and container pose estimation method. Finally, a simulation test platform is built to test the accuracy, speed and robustness of visual pose recognition under different lighting conditions. The test results show that on a common industrial PC platform, the minimum recognition speed is above 15Hz, the average error of the attitude angle is , and the average error of the displacement is , which can meet and rely on the lifting requirements.