基于单目视觉的集装箱角件自动对位

打开文本图片集

关键词：单目视觉；集装箱；自动对位；目标识别；位姿估计中图分类号：TH247;TP29 文献标识码：Adoi：10.37188/OPE.20253319.3150 CSTR：32169.14.OPE.20253319.3150

Abstract： To enhance the eficiency of container loading and unloading and realize automated operations， this study investigates monocular-vision-based automatic alignment between spreader and container，integrating deep learning and image-processing techniques. Monocular images of container hoisting conditions were analyzed with emphasis on the regional characteristics of corner components.To addressthe low pixel proportion of corner regions in high-resolution images，a two-stage "coarse positioning-fine segmentation" strategy Was proposed.Based on the segmentation results，key feature points were detected，2D-3D point correspondences were established，and pose estimation was performed using the Levenberg-Marquardt algorithm. Validation was conducted on an AUBO-ilO manipulator alignment platform in laboratory setings. Experimental results demonstrate that the mean average precision （mAP） for detection of container corner components exceeds 95% in both laboratory and real-scene environments. Mean intersectionover-union （mIoU） for corner segmentation reached 98.15% and 93.89% ，respectively-improvements of （20 1.24% and 1.64% over the baseline SegFormer-BO-while model computational cost was reduced by approximately 23.2% . At a camera-to-corner distance of about 2m ，the aiming error of the alignment positionwas below 1.0mm . Absolute translation errors on the X ， Y ，and Z axes were all below 5.0mm ， and absolute rotation error was below 0.5∘ . These results indicate that the proposed method achieves reliable accuracy and satisfies the requirements for automatic alignment of single-angle components.

Key words： monocular vision; containers; automatic alignment; target recognition; pose estimation

1引言

近年来，随着全球经济的快速增长，集装箱运输业得到了快速发展，集装箱运输通过标准化的装载和运输方式，大大提高了货物的运输效率和安全性，降低了物流成本，促进了国际贸易的繁荣[1]。（剩余14356字）