Insulator fault detection method based on collaborative deep learning

The invention discloses an insulator fault detection method based on collaborative deep learning, belongs to the technical field of insulator fault detection, and aims to solve the problem that the positioning and detection precision of an insulator fault area is affected due to the complex background of an insulator in an aerial image. The method is characterized in that an FCN-8s model is constructed through a jump structure, insulator image segmentation is completed based on an FCN algorithm, and the purpose of background filtering is effectively achieved; an insulator data set is clusteredand analyzed by using a K-means++ clustering algorithm with small randomness, the initial anchor point frame parameter of a YOLOv3 algorithm is optimized, and the positioning and detection precisionof the target detection model is further improved; and an insulator fault detection model cooperating with the FCN algorithm and the YOLOv3 algorithm is constructed, through experimental comparison, compared with an original Y