| Issue |
E3S Web Conf.
Volume 483, 2024
The 3rd International Seminar of Science and Technology (ISST 2023)
|
|
|---|---|---|
| Article Number | 03016 | |
| Number of page(s) | 8 | |
| Section | Trends in Mathematics and Computer Science for Sustainable Living | |
| DOI | https://doi.org/10.1051/e3sconf/202448303016 | |
| Published online | 31 January 2024 | |
Applying Digital Images to Identify Pavement Damage in Support of The Road Infrastructure Development Program
1 Politeknik Negeri Banyuwangi, Civil Engineering Department, 68461 Banyuwangi, Jawa Timur, Indonesia
2 Politeknik Negeri Banyuwangi, Informatics Engineering Department, 68461 Banyuwangi, Jawa Timur, Indonesia
3 Politeknik Negeri Banyuwangi, Tourism Business Management Department, 68461 Banyuwangi, Jawa Timur, Indonesia
4 Universitas Jember, Mathematics Education Department, 68121 Jember, Jawa Timur, Indonesia
* Corresponding author: siska_aprilia3@poliwangi.ac.id
Road damage can cause discomfort while driving and even lead to accidents. According to the National Road Network Condition Map Data in 2017, the level of severe and minor road damage in the East Java region had reached 288 kilometers. Based on this data, periodic road condition assessments and maintenance are essential to minimize damage. Road maintenance efforts are crucial to support road infrastructure development programs. The initial step in road maintenance is to identify road damage, determining the necessary actions to be taken. In this research, road pavement damage identification is carried out using the Yolov5, Yolov6, and Yolov7 methods. Test results indicate that the Yolov5 method performed the best with a validation mAP (mean Average Precision) score of 42%, a Precision value of 0.544, and a Recall value of 0.453. These scores indicate that the accuracy of road pavement damage detection using the YOLO algorithm for depression, corrugation, potholes, and alligator cracking is at its maximum level.
© The Authors, published by EDP Sciences, 2024
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.