All Issue

2025 Vol.35, Issue 4 Preview Page

Research Article

Proposing an Object Recognition Model for Analyzing Traffic Characteristics in Urban Pothole-Prone Areas

도심지 포트홀 발생 구간의 교통 특성 분석을 위한 객체 인식 모델 제안

Jeongdu Noh1
,
Seokjin Oh2
,
Hyeok Seo2*
노 정두1
,
오 석진2
,
서 혁2*
1Researcher, Engineering Technology Institute, Chosun University
2Assistant Professor, Department of Civil & Environmental Engineering, Honam University
1조선대학교 공학기술연구원 연구원
2호남대학교 토목환경공학과 조교수
*Corresponding Author
31 December 2025. pp. 787-797
PDF XML Citation
Abstract

Gwangju Metropolitan City has recently seen an increase in pothole occurrence, and these occurrences are concentrated in areas with heavy traffic volumes. Previous research has primarily focused on road pavement material characteristics or pothole detection using object recognition technology, but traffic characteristics such as traffic volume and vehicle type have not been adequately reflected. This study proposes a prediction model that utilizes YOLO11-based object recognition technology to identify vehicle types and traffic volumes from road image data and analyze traffic characteristics in areas where potholes occur. To train the prediction model, approximately 250,000 images of data, including AI Hub, public data, and our own video footage, were collected. The model was then trained to evaluate object detection performance for heavy vehicles such as trucks, buses, and special-purpose vehicles. As a result, the proposed model achieved high performance, with a precision of approximately 0.95, a recall of approximately 0.92, and an mAP50 of approximately 0.96. This study is expected to be used as basic data to support efficient road maintenance decision-making by identifying the correlation between heavy vehicle traffic and pothole occurrence through image-based traffic analysis.

Keywords
YOLO
object detection
traffic analysis
heavy vehicle detection
pothole-prone areas
image-based analysis

광주광역시는 최근 포트홀 발생이 증가하는 추세이며, 이는 중대형 차량의 통행이 많은 곳에 밀집되어 나타나고 있다. 기존 연구는 주로 도로 포장 재료 특성이나, 객체 인식 기술을 이용한 도로 내 포트홀 탐지 등에 집중되어 수행되었으나, 교통량, 차량 종류 등과 같은 교통 특성은 충분히 반영되지 못했다. 본 연구에서는 YOLO11 기반의 객체 인식 기술을 활용하여 도로 영상 자료에서 차량의 종류와 통행량을 식별하고, 포트홀 발생 지역의 교통 특성을 분석할 수 있는 기초 모델을 제안하였다. 예측 모델 학습을 위해 AI Hub, 공공데이터, 자체 영상 촬영 등 약 25만장의 데이터를 구축하였고, 이를 학습하여 트럭, 버스, 특수차 등 중차량의 객체 탐지 성능을 평가하였다. 그 결과 제안된 모델은 정밀도 약 0.95, 재현율 약 0.92, mAP50 약 0.96의 높은 성능을 나타냈다. 본 연구는 영상 기반의 교통 분석을 통해 중차량 통행과 포트홀 발생의 연관성을 규명하고 효율적인 도로 유지관리 의사결정을 지원하는 기초자료로 활용될 수 있을 것으로 기대된다.

키워드
YOLO
객체 인식
교통량 분석
중차량 탐지
포트홀 발생 지역
영상 기반 분석
References
1

Choi, S.H., Kim, W.J., Le, V.P., Lee, H.J., Hwang, S.D., 2016, Moisture damage evaluation of asphalt mixtures depending on the types of anti-stripping agent, International Journal of Highway Engineering, 16(4), 45-50 (in Korean with English abstract).

10.7855/IJHE.2014.16.4.045
2

Gwangju City, 2024, 2023 Gwangju metropolitan city transportation-related basic survey summary report, 1-380.

3

Han, D., 2020, Road patrol strategy based on pothole occurrence characteristics considering rainfall effects, Journal of the Korea Academia-Industrial cooperation Society, 21(12), 603-611 (in Korean with English abstract).

10.5762/KAIS.2020.21.12.603
4

Hong, S.P., Cho, Y.J., 2022, A study on road porthole detection based on deep learning, Journal of Information Technology and Applied Engineering, 12(2), 35-41 (in Korean with English abstract).

10.22733/JITAE.2022.12.02.005
5

Jo, M.H., Kim, N., Seo, Y., 2008, Asphalt concrete pavement response to moving load and viscoelastic property, KSCE Journal of Civil Environmental Engineering Research, 28(4D), 485-492 (in Korean with English abstract).

10.12652/Ksce.2008.28.4D.485
6

Kim, D., Jeon, J., Kim, D., 2019, Predictive modeling of pavement damage using machine learning and big data processing, Journal of the Korean Society of Hazard Mitigation, 19(1), 95-107 (in Korean with English abstract).

10.9798/KOSHAM.2019.19.1.95
7

Kim, J.Y., Ryou, J.S., Woo, B.H., 2025, Selective area emphasis preprocessing of images for pothole detection in asphalt pavement, Journal of the Korean Asphalt Institute, 15(1), 149-156 (in Korean with English abstract).

10.22702/JKAI.2025.15.1.14
8

Kim, S., Lee, D., Lee, S., Kim, Y., Kim, K.W., 2015, Evaluation of aging level of asphalt mixture at pothole site, Korean Asphalt Society, 5(1), 18-26 (in Korean with English abstract).

9

Lee, Y., Sun, J., 2020, Predicting highway concrete pavement damage using XGBoost, Korean Journal of Construction Engineering and Management, 21(6), 46-55 (in Korean with English abstract).

10.6106/KJCEM.2020.21.6.046
10

Ma, W., 2016, Simulate initiation and formation of cracks and potholes, MSc Thesis, Northeastern University, 1-93.

11

Sah, H., Ahn, D., 2024, Super resolution based mobile long-distance pothole detection system, Journal of the Korea Institute of Information and Communication Engineering, 28(6), 696-703 (in Korean with English abstract).

10.6109/jkiice.2024.28.6.696
12

Ultralytics, 2025, Performance metrics deep dive, Retrieved from https://docs.ultralytics.com/ko/guides/yolo-performance-metrics/.

13

Woo, B.H., Yoo, D.H., Choi, J.S., 2024a, Asphalt pavement pothole detection using a 3 channel based bayesian-convolutional neural network, Journal of the Korean Asphalt Institute, 14(2), 192-201 (in Korean with English abstract).

10.22702/JKAI.2024.14.2.18
14

Woo, B.H., Yoo, K.S., Choi, J.S., 2024b, Improving the accuracy of pothole tracking and prediction using Spiking Neural Network, Journal of the Korean Asphalt Institute, 14(1), 28-36 (in Korean with English abstract).

10.22702/JKAI.2024.14.1.4
15

Yang, S.L., Hwang, S.D., Kim, Y.M., Jeong, K.D., 2008, Evaluation of moisture susceptibility on asphalt mixtures mixing anti-stripping agent, International Journal of Highway Engineering, 10(4), 41-52 (in Korean with English abstract).

16

Yudakul, M., Tasdemir, S., 2025, An enhanced YOLOv8 model for real-time and accurate pothole detection and measurement, arXiv preprint arXiv:2505.04207.

17

Zhang, S., Wei, H., Liu, J., Tong, H., 2025, WRI-YOLO11: A deep learning approach for detecting surface defects in metal shafts, Engineering Research Express, 7, 025260.

10.1088/2631-8695/add792
Information
  • Publisher :Korean Society of Engineering Geology
  • Publisher(Ko) :대한지질공학회
  • Journal Title :The Journal of Engineering Geology
  • Journal Title(Ko) :지질공학
  • Volume : 35
  • No :4
  • Pages :787-797
  • Received Date : 2025-12-12
  • Revised Date : 2025-12-24
  • Accepted Date : 2025-12-24
  • DOI :https://doi.org/10.9720/kseg.2025.4.787
Journal Informaiton The Journal of Engineering Geology The Journal of Engineering Geology
Online Submission
  • NRF
  • KOFST
  • KISTI Current Status
  • KISTI Cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close