Abstract

This study explores the deployment of YOLOv8s for detecting anomalies in fiber optic cables mounted 
on poles, with a focus on climbing activities and environmental impediments. To address the lack of 
climbing-related annotations in current datasets, a custom dataset was generated, covering a variety 
of scenarios to enhance model adaptability. During training, various augmentation approaches were 
used, which greatly enhanced model performance and reduced overfitting. The proposed model was 
trained and tested over many epochs, with detection performance progressively improving: [email protected] 
increased from 78.9% at 20 epochs to 87.5% at 50 epochs and 97.3% at 100 epochs, after which further 
increases plateaued. In comparison, the trained YOLOv8s-modified model outperformed the other 
models on all key metrics. It achieved a mAP@50 of 97.3% and a mAP@50:95 of 71.5%, outperforming 
YOLOv8-original at 89.6% and 59.0%, respectively. Additionally, it achieved higher precision (96.9%) 
and recall (86.6%), demonstrating superior detection accuracy, dependability, and robustness in 
detecting complex anomalies. These improvements due to model backbone optimizations and the 
usage of a well-balanced, scenario-rich dataset. This study shows that YOLOv8s is a highly accurate and 
efficient method for detecting anomalies in fiber optic infrastructure, making it appropriate for real
time deployment in operational field environments.