Improved YOLOv5-Based Real-Time Road Pavement Damage Detection in Road Infrastructure Management
Deep learning has enabled a straightforward, convenient method of road pavement infrastructure management that facilitates a secure, cost-effective, and efficient transportation network. Manual road pavement inspection is time-consuming and dangerous, making timely road repair difficult. This resear...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-09-01
|
| Series: | Algorithms |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-4893/16/9/452 |
| _version_ | 1827727547328102400 |
|---|---|
| author | Abdullah As Sami Saadman Sakib Kaushik Deb Iqbal H. Sarker |
| author_facet | Abdullah As Sami Saadman Sakib Kaushik Deb Iqbal H. Sarker |
| author_sort | Abdullah As Sami |
| collection | DOAJ |
| description | Deep learning has enabled a straightforward, convenient method of road pavement infrastructure management that facilitates a secure, cost-effective, and efficient transportation network. Manual road pavement inspection is time-consuming and dangerous, making timely road repair difficult. This research showcases You Only Look Once version 5 (YOLOv5), the most commonly employed object detection model trained on the latest benchmark Road Damage Dataset, Road Damage Detection 2022 (RDD 2022). The RDD 2022 dataset includes four common types of road pavement damage, namely vertical cracks, horizontal cracks, alligator cracks, and potholes. This paper presents an improved deep neural network model based on YOLOv5 for real-time road pavement damage detection in photographic representations of outdoor road surfaces, making it an indispensable tool for efficient, real-time, and cost-effective road infrastructure management. The YOLOv5 model has been modified to incorporate several techniques that improve its accuracy and generalization performance. These techniques include the Efficient Channel Attention module (ECA-Net), label smoothing, the K-means<sup>++</sup> algorithm, Focal Loss, and an additional prediction layer. In addition, a 1.9% improvement in mean average precision (mAP) and a 1.29% increase in F1-Score were attained by the model in comparison to YOLOv5s, with an increment of 1.1 million parameters. Moreover, a 0.11% improvement in mAP and 0.05% improvement in F1 score was achieved by the proposed model compared to YOLOv8s while having 3 million fewer parameters and 12 gigabytes fewer Giga Floating Point Operation per Second (GFlops). |
| first_indexed | 2024-03-10T23:07:58Z |
| format | Article |
| id | doaj.art-dbd5e72356ce4ab4baa20063e7fa9572 |
| institution | Directory Open Access Journal |
| issn | 1999-4893 |
| language | English |
| last_indexed | 2024-03-10T23:07:58Z |
| publishDate | 2023-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Algorithms |
| spelling | doaj.art-dbd5e72356ce4ab4baa20063e7fa95722023-11-19T09:13:20ZengMDPI AGAlgorithms1999-48932023-09-0116945210.3390/a16090452Improved YOLOv5-Based Real-Time Road Pavement Damage Detection in Road Infrastructure ManagementAbdullah As Sami0Saadman Sakib1Kaushik Deb2Iqbal H. Sarker3Department of Computer Science & Engineering, Chittagong University of Engineering & Technology, Chottogram 4349, BangladeshDepartment of Computer Science & Engineering, Chittagong University of Engineering & Technology, Chottogram 4349, BangladeshDepartment of Computer Science & Engineering, Chittagong University of Engineering & Technology, Chottogram 4349, BangladeshDepartment of Computer Science & Engineering, Chittagong University of Engineering & Technology, Chottogram 4349, BangladeshDeep learning has enabled a straightforward, convenient method of road pavement infrastructure management that facilitates a secure, cost-effective, and efficient transportation network. Manual road pavement inspection is time-consuming and dangerous, making timely road repair difficult. This research showcases You Only Look Once version 5 (YOLOv5), the most commonly employed object detection model trained on the latest benchmark Road Damage Dataset, Road Damage Detection 2022 (RDD 2022). The RDD 2022 dataset includes four common types of road pavement damage, namely vertical cracks, horizontal cracks, alligator cracks, and potholes. This paper presents an improved deep neural network model based on YOLOv5 for real-time road pavement damage detection in photographic representations of outdoor road surfaces, making it an indispensable tool for efficient, real-time, and cost-effective road infrastructure management. The YOLOv5 model has been modified to incorporate several techniques that improve its accuracy and generalization performance. These techniques include the Efficient Channel Attention module (ECA-Net), label smoothing, the K-means<sup>++</sup> algorithm, Focal Loss, and an additional prediction layer. In addition, a 1.9% improvement in mean average precision (mAP) and a 1.29% increase in F1-Score were attained by the model in comparison to YOLOv5s, with an increment of 1.1 million parameters. Moreover, a 0.11% improvement in mAP and 0.05% improvement in F1 score was achieved by the proposed model compared to YOLOv8s while having 3 million fewer parameters and 12 gigabytes fewer Giga Floating Point Operation per Second (GFlops).https://www.mdpi.com/1999-4893/16/9/452road damage detectionpavement detectionYOLOv5ECA-NetFocal LossK-means<sup>++</sup> |
| spellingShingle | Abdullah As Sami Saadman Sakib Kaushik Deb Iqbal H. Sarker Improved YOLOv5-Based Real-Time Road Pavement Damage Detection in Road Infrastructure Management Algorithms road damage detection pavement detection YOLOv5 ECA-Net Focal Loss K-means<sup>++</sup> |
| title | Improved YOLOv5-Based Real-Time Road Pavement Damage Detection in Road Infrastructure Management |
| title_full | Improved YOLOv5-Based Real-Time Road Pavement Damage Detection in Road Infrastructure Management |
| title_fullStr | Improved YOLOv5-Based Real-Time Road Pavement Damage Detection in Road Infrastructure Management |
| title_full_unstemmed | Improved YOLOv5-Based Real-Time Road Pavement Damage Detection in Road Infrastructure Management |
| title_short | Improved YOLOv5-Based Real-Time Road Pavement Damage Detection in Road Infrastructure Management |
| title_sort | improved yolov5 based real time road pavement damage detection in road infrastructure management |
| topic | road damage detection pavement detection YOLOv5 ECA-Net Focal Loss K-means<sup>++</sup> |
| url | https://www.mdpi.com/1999-4893/16/9/452 |
| work_keys_str_mv | AT abdullahassami improvedyolov5basedrealtimeroadpavementdamagedetectioninroadinfrastructuremanagement AT saadmansakib improvedyolov5basedrealtimeroadpavementdamagedetectioninroadinfrastructuremanagement AT kaushikdeb improvedyolov5basedrealtimeroadpavementdamagedetectioninroadinfrastructuremanagement AT iqbalhsarker improvedyolov5basedrealtimeroadpavementdamagedetectioninroadinfrastructuremanagement |