计算机工程与应用2019，Vol.55Issue(24)：196-201,6.DOI:10.3778/j.issn.1002-8331.1810-0297
迁移学习用于牵引变电所视频多目标识别研究
Research on Multi-Target Recognition of Traction Substation Video Based on Transfer Learning
摘要
Abstract
In order to realize video image recognition of traction substation and provide technical support for remote intel-ligent inspection of traction substation. Based on the theory of transfer learning, the SSD(Single Shot Multibox Detector) and YOLOv2(You Only Look Once v2)models are used to realize automatic recognition of meters, split indicator lamp status of High Voltage Switchgear and disconnector status in the video image of traction substation. The multi-target rec-ognition method implemented by TensorFlow platform is fast and robust.关键词
牵引变电所/迁移学习/目标检测/YOLOv2模型/SSD模型Key words
traction substation/transfer learning/object detection/You Only Look Once v2(YOLOv2)model/Single Shot Multibox Detecto(r SSD)model分类
信息技术与安全科学引用本文复制引用
完颜幸幸,傅钦翠,吉鑫..迁移学习用于牵引变电所视频多目标识别研究[J].计算机工程与应用,2019,55(24):196-201,6.基金项目
2017年度江西省教育厅科学技术研究项目(No.GJJ170375) （No.GJJ170375）
江西省2018年度科技计划项目(No.20181BBE58011). （No.20181BBE58011）
