Abstract
The reliability of the insulator has directly affected the stable operation of electric power system. The detection of defective insulators has always been an important issue in smart grid systems. However, the traditional transmission line detection method has low accuracy and poor real-time performance. We present an insulator defect detection method based on CenterNet. In order to improve detection efficiency, we simplified the backbone network. In addition, an attention mechanism is utilized to suppress useless information and improve the accuracy of network detection. In image preprocessing, the blurring of some detected images results in the samples being discarded, so we use super-resolution reconstruction algorithm to reconstruct the blurred images to enhance the dataset. The results show that the AP of the proposed method reaches 96.16% and the reasoning speed reaches 30FPS under the test condition of NVIDIA GTX 1080 test conditions. Compared with Faster R-CNN, YOLOV3, RetinaNet and FSAF, the detection accuracy of proposed method is greatly improved, which fully proves the effectiveness of the proposed method.
Highlights
The demand for electricity is increasing day by day, which poses a huge challenge to the inspection and maintenance of power grid
As an indispensable device in the power system, the self-destruction of the insulator will seriously endanger the safe operation of the power grid system
On the basis of the above research, we propose a defect insulator detection algorithm based on WDSR and CenterNet, which uses ResNet50 as the backbone network
Summary
The demand for electricity is increasing day by day, which poses a huge challenge to the inspection and maintenance of power grid. Compared with the two-stage, the one-stage can achieve end-to-end detection and has a faster detection speed, but its accuracy is reduced, mainly including: YOLO [9], SSD [10], YOLOv2 [11], YOLOv3 [12], CenterNet [13], etc Whether it is a two-stage detection model or a one-stage detection model, the information assistance of a priori box is usually needed to regress to the ground truth. Since Law and Deng proposed the Cornernet model without anchor boxes [14], some corresponding anchorless frame models have attracted widespread attention from scholars [13, 15,16,17,18] Most of these detectors take key points, such as corners or centers, as positive samples to regress to the objects.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
