R-FCN disconnecting link detection method combining local features and global features

• Main Authors: TAO ZHENGYONG, ZONG QIZHEN, XIAO ZHENYUAN, CHU HONGJIAN, LI YOUWEN, ZENG QINGXUAN
• Format: Patent
• Language: Chinese; English
• Published: 29.07.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; MEASURING; MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES; PHYSICS; TESTING

The invention provides an R-FCN disconnecting link detection method combining local features and global features. According to the method, state detection is carried out on a disconnecting link of a transformer substation based on a deep learning network model R-FCN. Network camera centralized control software is embedded into a transformer substation auxiliary monitoring system to achieve camera multi-position and multi-angle disconnecting link image acquisition in different outdoor weather environments and backgrounds, and diverse disconnecting link data are constructed; the R-FCN output prediction network is connected in parallel and fused into the global feature prediction module, the defect that the receptive field is insufficient due to the fact that an original network predicts the disconnecting link only through local features is overcome, the detection accuracy of the partially-shielded disconnecting link is improved, and the omission ratio and the false detection rate of the disconnecting link in th