Hybrid Machine Learning for Automated Road Safety Inspection of Auckland Harbour Bridge DOI Open Access
Munish Rathee, Boris Bačić, Maryam Doborjeh

и другие.

Electronics, Год журнала: 2024, Номер 13(15), С. 3030 - 3030

Опубликована: Авг. 1, 2024

The Auckland Harbour Bridge (AHB) utilises a movable concrete barrier (MCB) to regulate the uneven bidirectional flow of daily traffic. In addition risk human error during regular visual inspections, staff members inspecting MCB work in diverse weather and light conditions, exerting themselves ergonomically unhealthy inspection postures with added weight protection gear mitigate risks, e.g., flying debris. To augment inspections an using computer vision technology, this study introduces hybrid deep learning solution that combines kernel manipulation custom transfer strategies. video data recordings were captured conditions (under safety supervision industry experts) involving high-speed (120 fps) camera system attached vehicle. Before identifying hazard, unsafe position pin connecting two 750 kg segments MCB, multi-stage preprocessing spatiotemporal region interest (ROI) involves rolling window before frames containing diagnostic information. This ResNet-50 architecture, enhanced 3D convolutions, within STENet framework capture analyse data, facilitating real-time surveillance (AHB). Considering sparse nature anomalies, initial peer-reviewed binary classification results (82.6%) for safe (intervention-required) scenarios improved 93.6% by incorporating synthetic expert feedback, retraining model. adaptation allowed optimised detection false positives negatives. future, we aim extend anomaly methods various infrastructure enhancing urban resilience, transport efficiency safety.

Язык: Английский

Depth feature fusion based surface defect region identification method for steel plate manufacturing DOI
Dongxu Bai, Gongfa Li, Du Jiang

и другие.

Computers & Electrical Engineering, Год журнала: 2024, Номер 116, С. 109166 - 109166

Опубликована: Март 11, 2024

Язык: Английский

Процитировано

9
LGGFormer: A dual-branch local-guided global self-attention network for surface defect segmentation DOI
Gaowei Zhang, Yang Lu, Xiaoheng Jiang

и другие.

Advanced Engineering Informatics, Год журнала: 2025, Номер 64, С. 103099 - 103099

Опубликована: Янв. 7, 2025

Язык: Английский

Процитировано

1
An efficient detector for detecting surface defects on cold-rolled steel strips DOI

Shuzong Chen,

Shengquan Jiang,

Xiaoyu Wang

и другие.

Engineering Applications of Artificial Intelligence, Год журнала: 2024, Номер 138, С. 109325 - 109325

Опубликована: Сен. 13, 2024

Язык: Английский

Процитировано

6
An Intelligent Industrial Visual Monitoring and Maintenance Framework Empowered by Large-scale Visual and Language Models DOI
Huan Wang, Chenxi Li, Yan‐Fu Li

и другие.

IEEE Transactions on Industrial Cyber-Physical Systems, Год журнала: 2024, Номер 2, С. 166 - 175

Опубликована: Янв. 1, 2024

Язык: Английский

Процитировано

4
Unsupervised surface defect detection using dictionary-based sparse representation DOI Creative Commons
Fanwu Meng, Tao Gong, Di Wu

и другие.

Engineering Applications of Artificial Intelligence, Год журнала: 2025, Номер 143, С. 110020 - 110020

Опубликована: Янв. 11, 2025

Язык: Английский

Процитировано

0
A computer vision system for recognition and defect detection for reusable containers DOI Creative Commons
Vincent Wahyudi,

Cedric C. Ziegler,

Matthias Frieß

и другие.

Machine Vision and Applications, Год журнала: 2025, Номер 36(2)

Опубликована: Янв. 18, 2025

Язык: Английский

Процитировано

0
A convex Kullback-Leibler divergence and critical-descriptor prototypes for semi-supervised few-shot learning DOI
Yukun Liu, Daming Shi

Applied Intelligence, Год журнала: 2025, Номер 55(5)

Опубликована: Янв. 21, 2025

Язык: Английский

Процитировано

0
An Auto Hierarchical Clustering Algorithm to Distinguish Geometries Suitable for Additive and Traditional Manufacturing Technologies: Comparing Humans and Unsupervised Learning DOI Creative Commons
Baris Ördek, Éric Coatanéa, Yuri Borgianni

и другие.

Results in Engineering, Год журнала: 2025, Номер unknown, С. 104418 - 104418

Опубликована: Фев. 1, 2025

Язык: Английский

Процитировано

0
An ESG-ConvNeXt network for steel surface defect classification based on hybrid attention mechanism DOI Creative Commons
Ning Zhang, Ziyang Liu,

Enxu Zhang

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Март 29, 2025

Defect recognition is crucial in steel production and quality control, but performing this detection task accurately presents significant challenges. ConvNeXt, a model based on self-attention mechanism, has shown excellent performance image classification tasks. To further enhance ConvNeXt's ability to classify defects surfaces, we propose network architecture called ESG-ConvNeXt. First, the processing stage, introduce serial multi-attention mechanism approach. This method fully leverages extracted information improves retention by combining strengths of each module. Second, design parallel multi-scale residual module adaptively extract diverse discriminative features from input image, thereby enhancing model's feature extraction capability. Finally, downsampling incorporate PReLU activation function mitigate problem neuron death during downsampling. We conducted extensive experiments using NEU-CLS-64 surface defect dataset, results demonstrate that our outperforms other methods terms performance, achieving an average accuracy 97.5%. Through ablation experiments, validated effectiveness module; through visualization exhibited strong Additionally, X-SDD dataset confirm ESG-ConvNeXt achieves solid results. Therefore, proposed shows great potential classification.

Язык: Английский

Процитировано

0
PBD-YOLO: Dual-Strategy Integration of Multi-Scale Feature Fusion and Weak Texture Enhancement for Lightweight Particleboard Surface Defect Detection DOI Creative Commons

Haomeng Guo,

Zheming Chai, Huan Dai

и другие.

Applied Sciences, Год журнала: 2025, Номер 15(8), С. 4343 - 4343

Опубликована: Апрель 15, 2025

Surface defect detection plays an important role in particleboard quality control. But it still faces challenges detecting coexisting multi-scale defects and weak texture defects. To address these issues, this study proposed PBD-YOLO (Particleboard Defect-You Only Look Once), a lightweight YOLO-based algorithm with feature fusion enhancement capabilities. In order to improve the ability of extract features, SPDDEConv (Space Depth Difference Enhance Convolution) module was introduced study, which reduced loss information down-sampling process through space-to-depth transformation enhanced edge difference convolution. This approach improved mAP (mean average precision) weakly featured but edge-sensitive (such as scratches) by much 20.9%. algorithm’s detect defects, ShareSepHead (Share Separated Head) C2f_SAC (C2f Switchable Atrous modules. fused maps from different scales neck network adding convolutional layer shared weights, adaptively multi-rate receptive fields switching mechanism. The synergistic effect accuracy 10.6–20.8%. experimental results demonstrated that achieved 85.6% at 50% intersection over union (IoU) 81.4% recall, surpassing YOLOv10 5.5% 13%, respectively, while reducing parameters 11.3%. summary, could be better meet need accurately surface on particleboard.

Язык: Английский

Процитировано

0