Modeling Multi-Granularity Context Information Flow for Pavement Crack Detection

doi:10.1109/TITS.2024.3438883

	Modeling Multi-Granularity Context Information Flow for Pavement Crack Detection
	Pang, Junbiao 1; Xiong, Baocheng 1; Wu, Jiaqi 1; Huang, Qingming 2,3
	2025-07-01
发表期刊	IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
ISSN	1524-9050
卷号	26 期号:7 页码:9165-9174
摘要	Pavement cracks have a highly complex spatialstructure, a low contrasting background and a weak spatialcontinuity, posing a significant challenge to an effective crackdetection method. To precisely localize crack from an image, it iscritical to effectively extract and aggregate multi-granularity con-text, including the fine-grained local context around the cracks(in spatial-level) and the coarse-grained semantics (in semantic-level). In this paper, we apply the dilated convolution as thebackbone feature extractor to model local context, then we builda context guidance module to leverage semantic context to guidelocal feature extraction at multiple stages. To handle label align-ment between stages, we apply the Multiple Instance Learning(MIL) strategy to align the feature between two stages. In addi-tion, to our best knowledge, we have released the largest, mostcomplex and most challenging Bitumen Pavement Crack (BPC)dataset. The experimental results on the three crack datasetsdemonstrate that the proposed method performs well and outper-forms the current state-of-the-art methods. On BPC, the proposedmodel achieved AP 88.32% with the 16.89 M parameters underthe 45.36 GFlops runing speed. Datset and code are publiclyavailable at: https://github.com/pangjunbiao/BPC-Crack-Dataset.
关键词	Feature extraction Semantics Heating systems Context modeling Noise Convolution Asphalt Semantic segmentation YOLO Training Crack detection context information multi-scale spatial structure
DOI	10.1109/TITS.2024.3438883
收录类别	SCI
语种	英语
WOS研究方向	Engineering ; Transportation
WOS类目	Engineering, Civil ; Engineering, Electrical & Electronic ; Transportation Science & Technology
WOS记录号	WOS:001508153100001
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/42370
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Pang, Junbiao
作者单位	1.Beijing Univ Technol, Fac Informat Technol, Beijing 100124, Peoples R China 2.Univ Chinese Acad Sci, Sch Comp Sci & Technol, Beijing 101408, Peoples R China 3.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Pang, Junbiao,Xiong, Baocheng,Wu, Jiaqi,et al. Modeling Multi-Granularity Context Information Flow for Pavement Crack Detection[J]. IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS,2025,26(7):9165-9174.
APA	Pang, Junbiao,Xiong, Baocheng,Wu, Jiaqi,&Huang, Qingming.(2025).Modeling Multi-Granularity Context Information Flow for Pavement Crack Detection.IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS,26(7),9165-9174.
MLA	Pang, Junbiao,et al."Modeling Multi-Granularity Context Information Flow for Pavement Crack Detection".IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 26.7(2025):9165-9174.