RoadDiffBox: Automatic Road Distress Diagnosis through Controlled Image Generation and Semi-Supervised Learning.

IF 10.7 1区 综合性期刊 Q1 Multidisciplinary
Research Pub Date : 2025-08-25 eCollection Date: 2025-01-01 DOI:10.34133/research.0833
Yuanyuan Hu, Ning Chen, Hancheng Zhang, Yue Hou, Pengfei Liu
{"title":"RoadDiffBox: Automatic Road Distress Diagnosis through Controlled Image Generation and Semi-Supervised Learning.","authors":"Yuanyuan Hu, Ning Chen, Hancheng Zhang, Yue Hou, Pengfei Liu","doi":"10.34133/research.0833","DOIUrl":null,"url":null,"abstract":"<p><p>During the designed service life, road infrastructures will bear repeated loading conditions from vehicle weights and environmental conditions, resulting in the inevitable occurrence of road distresses including cracks, potholes, etc. The traditional inspection methods by transportation engineers are normally costly and labor-intensive. In recent years, artificial intelligence (AI)-based road distress detection methods have been widely used as convenient and automated approaches, while the AI-based methods heavily depend on a large amount of high-quality images, limiting the real engineering applications. To address the issues, this study introduces RoadDiffBox, a novel framework employing controlled image generation and semi-supervised learning. The framework addresses dataset imbalances through class control and accelerates image generation by utilizing the denoising diffusion implicit model's reverse process sampling method, while employing knowledge distillation techniques optimized for resource-constrained mobile devices. It generates diverse and high-quality road distress images with automatic bounding box annotations, substantially reducing manual labeling requirements. Test results show that RoadDiffBox demonstrates strong generalizability across geographic regions (Germany, China, and India) and shows cross-domain potential in medical imaging applications. Performance evaluations demonstrate RoadDiffBox's effectiveness, with classification models achieving an F1-score of 0.95 and detection models reaching a mean average precision (mAP@50) of 0.95 and an F1-score of 0.91 in controlled settings, while maintaining robust performance (an F1-score of 0.86 and a mAP@50 of 0.91) during on-site testing in real-world conditions. On server-class hardware, the model achieves generation times as low as 0.18 s per image. It is discovered that RoadDiffBox can serve as a scalable and efficient solution for real-time road maintenance with limited datasets.</p>","PeriodicalId":21120,"journal":{"name":"Research","volume":"8 ","pages":"0833"},"PeriodicalIF":10.7000,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12376291/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Research","FirstCategoryId":"103","ListUrlMain":"https://doi.org/10.34133/research.0833","RegionNum":1,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/1/1 0:00:00","PubModel":"eCollection","JCR":"Q1","JCRName":"Multidisciplinary","Score":null,"Total":0}
引用次数: 0

Abstract

During the designed service life, road infrastructures will bear repeated loading conditions from vehicle weights and environmental conditions, resulting in the inevitable occurrence of road distresses including cracks, potholes, etc. The traditional inspection methods by transportation engineers are normally costly and labor-intensive. In recent years, artificial intelligence (AI)-based road distress detection methods have been widely used as convenient and automated approaches, while the AI-based methods heavily depend on a large amount of high-quality images, limiting the real engineering applications. To address the issues, this study introduces RoadDiffBox, a novel framework employing controlled image generation and semi-supervised learning. The framework addresses dataset imbalances through class control and accelerates image generation by utilizing the denoising diffusion implicit model's reverse process sampling method, while employing knowledge distillation techniques optimized for resource-constrained mobile devices. It generates diverse and high-quality road distress images with automatic bounding box annotations, substantially reducing manual labeling requirements. Test results show that RoadDiffBox demonstrates strong generalizability across geographic regions (Germany, China, and India) and shows cross-domain potential in medical imaging applications. Performance evaluations demonstrate RoadDiffBox's effectiveness, with classification models achieving an F1-score of 0.95 and detection models reaching a mean average precision (mAP@50) of 0.95 and an F1-score of 0.91 in controlled settings, while maintaining robust performance (an F1-score of 0.86 and a mAP@50 of 0.91) during on-site testing in real-world conditions. On server-class hardware, the model achieves generation times as low as 0.18 s per image. It is discovered that RoadDiffBox can serve as a scalable and efficient solution for real-time road maintenance with limited datasets.

RoadDiffBox:通过控制图像生成和半监督学习实现道路遇险自动诊断。
在设计使用寿命期间,道路基础设施将承受来自车辆重量和环境条件的反复载荷条件,从而不可避免地出现裂缝、坑洼等道路病害。传统的运输工程师检测方法通常是昂贵和劳动密集型的。近年来,基于人工智能(AI)的道路遇险检测方法作为一种方便、自动化的方法得到了广泛的应用,但这种方法严重依赖于大量的高质量图像,限制了实际的工程应用。为了解决这些问题,本研究引入了RoadDiffBox,这是一个采用受控图像生成和半监督学习的新框架。该框架通过类控制解决数据集不平衡问题,并利用去噪扩散隐式模型的反向过程采样方法加速图像生成,同时采用针对资源受限的移动设备优化的知识蒸馏技术。它生成多样化和高质量的道路遇险图像,并带有自动边界框注释,大大减少了手动标记的要求。测试结果表明,RoadDiffBox在地理区域(德国、中国和印度)具有很强的通用性,并在医学成像应用中显示出跨领域的潜力。性能评估证明了RoadDiffBox的有效性,分类模型的f1得分为0.95,检测模型的平均精度(mAP@50)为0.95,f1得分为0.91,同时在实际条件下的现场测试中保持了稳健的性能(f1得分为0.86,mAP@50为0.91)。在服务器级硬件上,该模型实现每个映像的生成时间低至0.18秒。研究发现,RoadDiffBox可以作为一个可扩展的、高效的解决方案,用于有限数据集的实时道路维护。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Research
Research Multidisciplinary-Multidisciplinary
CiteScore
13.40
自引率
3.60%
发文量
0
审稿时长
14 weeks
期刊介绍: Research serves as a global platform for academic exchange, collaboration, and technological advancements. This journal welcomes high-quality research contributions from any domain, with open arms to authors from around the globe. Comprising fundamental research in the life and physical sciences, Research also highlights significant findings and issues in engineering and applied science. The journal proudly features original research articles, reviews, perspectives, and editorials, fostering a diverse and dynamic scholarly environment.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信