Research Of Insulator Fault Recognition In Power Inspection Image

With the deepening of China ’s urban network transformation and smart grid construction,the number of insulator devices in overhead lines has increased.Most of the overhead lines are erected in remote areas,and the difficulty of detecting insulator devices has also increased.Insulators play an important role as key components in overhead cables.They are exposed to the environment during work and are vulnerable.For the detection of insulators,the traditional manual detection is time-consuming and laborious.With the development of UAV technology,the use of aerial photography technology to assist power inspection has gradually become the mainstream.In this paper,the insulator in the overhead line of the power system is taken as the research object.In the process of insulator inspection,due to the background interference,small fault area and low resolution of the collected image,the detection is difficult,the accuracy is low,and the problems of missed detection and false detection are easy to occur.In this paper,an improved target detection algorithm based on YOLOv5 is proposed in combination with super-resolution network,and an insulator fault detection system based on UAV aerial image is built to detect the insulator aerial image.The experimental results show that the method improves the accuracy and speed of insulator fault detection.The main research work of this paper is as follows :1.Aiming at the problems of limited image resources,interference of environmental factors and limitation of shooting angle in the actual collection process of insulator data set,the data set is amplified,and the labeling tool is used to label the category and location,and the insulator data set is constructed for subsequent experiments.2.Aiming at the problem of low image resolution in the process of aerial image acquisition,this paper analyzes the difference between high-resolution images and lowresolution image feature maps in the high-frequency domain,and proposes a DLMAN algorithm.The residual network is used to add dynamic attention,optimize the network structure,and improve the learning efficiency of the network.Finally,experiments are designed to compare with other algorithms.The experimental results show that the improved algorithm improves the quality of the image and improves the accuracy of the subsequent detection process for the low-resolution insulator image in the acquisition process.3.Aiming at the limitations of manual detection and the low accuracy of current insulator detection algorithms,an improved LMYOLOv5 detection algorithm based on YOLOv5 is proposed.The backbone network is constructed by a lightweight module,a small-scale network detection layer is added,and a receptive field module is designed so that the network can better extract feature information.In order to verify the effectiveness of the improved algorithm proposed in this paper,experiments were carried out by setting up comparative experiments and ablation experiments.The experimental results show that the algorithm performs better in detection accuracy,speed and accuracy.4.An insulator fault detection system based on UAV aerial image is built.By designing the UAV hardware,the host computer graphical interface is constructed,and the insulator fault detection algorithm is run to realize the intelligent detection of insulators.Fig.[58] table [6] reference [81]...