.png){kind=icon}
.png){kind=icon}
Return
Lightweight Insulator Defect Detection Algorithm Based on YOLOv8
.png){kind=icon}
OA
AI
DOI：10.3788/LOP250801.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A lightweight insulator defect detection algorithm based on YOLOv8 is proposed to address the problems of excessive size and high computational complexity in current transmission line insulator defect detection models. First, a reversible column network is used as the backbone network to ensure the effective retention of underlying information by repeatedly passing input to each column, thus maintaining high accuracy of the network. Second, a context guided network is used in the neck network to improve joint features and enhance focus on key features by utilizing global contextual information, thereby significantly reducing the parameters and floating point operations (FLOPs). Finally, an improved depthwise separable convolution that preserves the original information is introduced in the prediction layer to make the model lightweight while retaining as much feature layer information as possible, in order to improve the detection performance of the model. Experimental results show that, the parameters and FLOPs of the proposed algorithm are 1.32x10(6) and 3.4x10(9), respectively, with a mean average precision of 90.0% for insulator defect detection. Its comprehensive performance meets the practical needs of detecting insulator defects in transmission lines for application on edge devices.
Keywords:
insulator
defect detection
lightweight
feature fusion
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
L
IF:
1
Papers:
91
Citations:
3.9K