Two-stage insulator self-explosion defect detection method based on Mask R-CNN
Miao Hu, Xiaoming Ju · 2021
With the continuous development of power systems, the application of UAVs in the inspection of power systems has become more and more extensive. Aiming at the problem of few defect insulator samples and complex background, a detection method of Two-stage insulator self-explosion defect detection method based on Mask R-CNN is proposed. This paper uses Mask R-CNN to obtain the mask of the insulator and the corresponding original image part, and send them into the classification model with SPP network. The final insulator detection accuracy reached 87.5%.