.png){kind=icon}
.png){kind=icon}
Return
Lightweight rust detection network model for lightning rod descaling robots
.png){kind=icon}
PRE
AI
DOI：10.1007/s12206-025-1145-x.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper presents an improved lightweight rust detection network model, YOLO-LR, specifically designed for use in descaling robots to address the difficulties encountered in detecting rust on transmission tower lightning rods. The model is based on YOLO11 and introduces several key improvements. First, the ADown module replaces standard convolution operations, allowing the model to capture high-level image features effectively. Second, the C3K2-MS module is incorporated to mitigate the detrimental effect of edge noise. In addition, the SDEC-head module is designed to effectively reduce the model's size and parameter count. When evaluated on the transmission tower lightning rod rust dataset, YOLO-LR achieves reductions of 0.809 M parameters, 1.1 G FLOPs, and 1 MB in model size compared to YOLO11. Overall, YOLO-LR serves as a valuable reference for the development of lightning rod descaling robots and lightweight YOLO-based models.
Keywords:
Descaling robots
Lightning rods
Lightweight YOLO
Rust detection
Journal
IF:
1.7
Papers:
601
Citations:
1.2W