基于人工智能的光交箱资源管理方法

Resource management method of optical delivery box based on artificial intelligence

随着全光网时代的到来,光交箱作为光缆接入网的最后一公里触点,是电信业务建设和维护中最频繁操作的平台。针对光交箱资源信息人工维护方式存在的低效问题,基于WS-DAN细粒度分类算法和YoloV5目标检测算法,在手机APP端开发资源现场核查功能,对光交箱的工艺和端子信息进行智能识别,提升光交箱资源现场核查效率和数据准确性,可在电信运营商推广应用。

With the advent of the all-optical network era, as the last mile contact of the optical cable access network, the optical delivery box is the most frequently operated platform in the construction and maintenance of telecommunication services. Aiming at the inefficiency of the manual maintenance method of the optical delivery box resource information, a fine-grained classification algorithm based on WS-DAN and YoloV5 target detection algorithm are proposed. The on-site verification function of resources is developed on the mobile phone app to intelligently identify the craftsmanship identification and terminal information of the optical delivery box, so as to improve the efficiency and data accuracy of on-site verification of the optical delivery box resources, and it can be promoted by telecom operators.