Development of a cloud-based road surface quality assessment system
This research aims to develop a cloud-based system utilizing the You Only Look Once version 8 (YOLOv8) model for assessing road surface quality. The system is designed to address critical road maintenance challenges and the need for high accuracy and fast response road surface quality monitoring. Da...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Qeios Ltd.
2024
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/43459/ http://umpir.ump.edu.my/id/eprint/43459/1/04NBI2.2.pdf |
| Summary: | This research aims to develop a cloud-based system utilizing the You Only Look Once version 8 (YOLOv8) model for assessing road surface quality. The system is designed to address critical road maintenance challenges and the need for high accuracy and fast response road surface quality monitoring. Data acquisition involved images from the Internet, dashcams, and smartphones, with subsequent processing through advanced image techniques. The YOLOv8 model demonstrated efficacy in detecting various road surface defects, achieving a precision of 0.457 and a recall of 0.486. While exhibiting potential in identifying patches and potholes, further refinement is required for crack detection. The model’s processing speed, with 9.7 milliseconds per image, indicates its capability for near real-time analysis. Finally, the model is deployed on cloud infrastructure hosted by Digital Ocean to provide scalability and accessibility. The cloud-based system enables users to upload videos for automated defect detection and offers downloadable results, fostering collaborative initiatives in road surface monitoring. While the model shows promise, particularly in detecting patches and potholes, crack detection has room for improvement. Future work could focus on enhancing the model’s performance for this challenging defect class. |
|---|