Infrared image identification method of substation equipment fault under weak supervision
This study presents a weakly supervised method for identifying faults in infrared images of substation equipment. It utilizes the Faster RCNN model for equipment identification, enhancing detection accuracy through modifications to the model's network structure and parameters. The method is exemplified through the analysis of infrared images captured by inspection robots at substations. Performance is validated against manually marked results, demonstrating that the proposed algorithm significantly enhances the accuracy of fault identification across various equipment types.
updated: Sun Nov 19 2023 03:33:46 GMT+0000 (UTC)
published: Sun Nov 19 2023 03:33:46 GMT+0000 (UTC)