Benchmarking YOLOv8 Variants with Transfer Learning for Real-Time Detection and Classification of Road Cracks and Potholes

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Performance evaluation of YOLOv8 variants shows high precision in detecting road damage, highlighting the efficacy of transfer learning methods.

YOLOv8m achieved the highest precision of 99.00% in road damage detection, indicating superior performance among variants.
In tests for potholes, precision reached 98.70% and recall 99.30%, demonstrating effective classification capability.
Benchmarking was performed using a dataset of 3,846 categorized images from Roboflow, enhancing the training process.
Results suggest YOLOv8 methods can transform infrastructure monitoring, reducing reliance on manual inspections and human error.

Kurniadi et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

Performance evaluation of YOLOv8 variants shows high precision in detecting road damage, highlighting the efficacy of transfer learning methods.