Construction and Optimization of Multimodal Perception Model for Defect Identification of Power Grid Equipment
In the maintenance and management of power grid equipment,defect identification is a key link in preventing equip-ment failure.However,traditional defect identification methods mainly rely on the processing of visible light and infrared data,and will face problems such as low recognition accuracy and poor generalization.In response to these challenges,this study pro-poses a multi-modal defect identification method based on Transformer.By integrating multiple data modalities such as visible light and infrared,the limitations of single modal data are overcome and more abundant information is provided for defect iden-tification;the U-Net network structure is used to effectively extract feature information from power grid equipment images,which provides a solid foundation for subsequent defect identification.The Transformer structure has been optimized to improve its performance in power grid equipment defect identification tasks,achieving precise positioning and defect identification of power grid equipment such as transformers,bushings,and circuit breakers.Experimental results show that this method has a-chieved significant improvement in the defect recognition task.It not only improves the recognition accuracy,but also enhances the robustness of the model,allowing the model to better adapt to the recognition tasks of different equipment and defect types.
defect identificationpower grid equipmentmultimodal perceptionTransformer modelmodel building
NETL
NSTL
万方数据
