【摘要】 计算机视觉技术在产业升级中发挥了重要的作用，是建设下一代智慧图书馆的关键技术，而位姿计算技术是计算机视觉技术重要的研究应用方向。论文首次探索了基于视觉的位姿计算技术在图书馆中的应用，探讨了该技术对图书馆服务的重要意义，并且基于开源框架成功实现了基于视觉的位姿计算技术，将位姿计算系统实际部署在西北工业大学图书馆中。代表性场景的位姿计算结果表明，论文基于视觉的位姿计算技术方法，其位移精度在10厘米以内，角度精度在2度以内，定位平均耗时约为50毫秒，位姿精度和时间性能满足图书馆场景下定位的需求。论文是将计算机视觉技术用于构建智慧图书馆的一次有益尝试，也为其他相关技术在图书馆的落地实践提供了较好的借鉴作用。更多还原

【Abstract】 Computer vision plays an important role and value in industrial upgrading, and are key technologies for constructing next-generation intelligent library. Vision-based 6 dof location method is an important research and application direction of this filed. This work is the first time to explore applying vision-based location and pose calculation method in library scenario, and discuss the important value of the related applications. We implemented the vison-based location and pose calculation method based on some open source frameworks, and successfully deploy the location system in NWPU library. Representative scenarios show that transition and rotation accuracy of our vision-based location and pose calculation method are within 10 cm and 2 degree respectively, execution time of one calculation is about 50 ms, which can satisfy library requirements. This work is a valuable attempt of applying computer vision technology in intelligent library field, and has good reference function of practicing related technologies.更多还原