| abstract |
The invention discloses a method for identifying and evaluating the fault state of a transmission tower based on Mask RCNN. The steps include: step 1, establishing an image database; step 2, building a MaskRCNN transmission tower fault state identification and evaluation model, and modifying the last layer FC of the model The number of layers is classified, and the model is migrated and fine-tuned by using the image database to realize multi-target recognition and instance segmentation of birds, bird nests, insulators and their body faults; combined with image processing technology, the results of the identified bird droppings stains and bird pecking marks Correction and fault status assessment; Step 3, analyze the characteristics of bird-related faults, including the analysis of bird-related fault types and the characteristics of affected bird species; Step 4, use the analysis results of Step 3 to establish a fusion of the identification of bird species involved. The bird fault discrimination method realizes the identification and evaluation of all categories of bird-related faults. The method of the invention can more precisely escort the safe operation of the transmission line. |