基于人工智能的工业产品缺陷检测关键技术与应用研究

doi:10.12267/j.issn.2096-5931.2025.08.010

信息通信技术与政策 ›› 2025, Vol. 51 ›› Issue (8): 71-77.doi: 10.12267/j.issn.2096-5931.2025.08.010

基于人工智能的工业产品缺陷检测关键技术与应用研究

Research on key technologies and applications of AI-enabled industrial product defect detection

丁怡心¹, 虞文明², 董昊¹

1.中国信息通信研究院人工智能研究所,北京 100191
2.联想(北京)有限公司,北京 100080

收稿日期:2025-07-10 出版日期:2025-08-25 发布日期:2025-09-02
作者简介:
丁怡心,中国信息通信研究院人工智能研究所工程师,主要从事边缘人工智能相关技术及产业等方面的研究工作
虞文明,联想(北京)有限公司人工智能实验室研发总监,主要负责联想边缘人工智能技术及解决方案等方面的工作
董昊,中国信息通信研究院人工智能研究所工程师,长期从事科研智能、人工智能基础设施、人工智能工程化等方面的研究工作

DING Yixin¹, YU Wenming², DONG Hao¹

1. Artificial Intelligence Institute, China Academy of Information and Communications Technology, Beijing 100191, China
2. Lenovo (Beijing) Limited, Beijing 100080, China

Received:2025-07-10 Online:2025-08-25 Published:2025-09-02

摘要/Abstract

摘要：

随着人工智能技术的快速发展,工业产品缺陷检测正从传统人工目检和图像处理方法向智能化、通用化方向演进。系统梳理了工业产品缺陷检测的类型与技术发展路径,对传统图像处理到深度学习再到基础模型阶段的技术演进进行了深入分析,重点探讨了当前在数据增强、小样本学习、模型轻量化及云边端协同等方面的关键技术突破,结合3C、钢铁、新能源电池、轮胎、纺织等行业的实际应用案例,展示了人工智能赋能缺陷检测所带来的效率提升、成本降低和质量保障等方面的显著成效,为制造业智能化转型提供了技术支撑与实践参考。

关键词: 缺陷检测, 基础模型, 数据增强, 小样本学习, 云边端协同

Abstract:

With the rapid development of artificial intelligence (AI) technology, defect inspection for industrial products is evolving from traditional manual visual inspection and image processing methods towards intelligent and generalized approaches. This paper systematically reviews the types of industrial product defect detection and the path of technological development, makes an in-depth analysis of the technological evolution from traditional image processing to deep learning and then to the basic model stage, focuses on the current key technological breakthroughs in the areas of data augmentation, few-shot learning, model lightweight and cloud-edge-end collaboration. Through practical application cases in industries including consumer electronics, steel, new energy batteries, tires, and textiles, the paper demonstrates the significant effects of efficiency improvement, cost reduction and quality assurance brought by AI-enabled defect detection, providing technical support and practical reference for the intelligent transformation of the manufacturing industry.

Key words: defect inspection, foundation models, data augmentation

中图分类号:

TP18
TN929.5

丁怡心, 虞文明, 董昊. 基于人工智能的工业产品缺陷检测关键技术与应用研究[J]. 信息通信技术与政策, 2025, 51(8): 71-77.

DING Yixin, YU Wenming, DONG Hao. Research on key technologies and applications of AI-enabled industrial product defect detection[J]. Information and Communications Technology and Policy, 2025, 51(8): 71-77.

导出引用管理器 EndNote|Ris|BibTeX

链接本文:

http://ictp.caict.ac.cn/CN/10.12267/j.issn.2096-5931.2025.08.010

http://ictp.caict.ac.cn/CN/Y2025/V51/I8/71

参考文献 29

[1]	JAIN S, SETH G, PARUTHI A, et al. Synthetic data augmentation for surface defect detection and classification using deep learning[J]. Journal of Intelligent Manufacturing, 2020, 33(4):1007-1020.
[2]	HE X, CHANG Z, ZHANG L, et al. A survey of defect detection applications based on generative adversarial networks[J]. IEEE Access, 2022,10:113493-113512.
[3]	GAO Y, LI X, WANG V, et al. A review on recent advances in vision-based defect recognition towards industrial intelligence[J]. Journal of Manufacturing Systems, 2022,62:753-766.
[4]	RASOUL A, CHUNG-CHIAN H, SHAHAB B. A systematic review of deep learning approaches for surface defect detection in industrial applications[J]. Engineering Applications of Artificial Intelligence, 2024,130:107717.
[5]	RIQUELME C, PUIGCERVER J, MUSTAFA B, et al. Scaling vision with sparse mixture of experts[J]. arXiv Preprint, arXiv:2106.05974, 2021.
[6]	SHAO J, CHEN S Y, LI Y G, et al. INTERN: a new learning paradigm towards general vision[J]. arXiv Preprint, arXiv:2111.08687, 2022.
[7]	DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16×16 words: transformers for image recognition at scale[J]. arXiv Preprint, arXiv:2010.11929, 2020.
[8]	BAO H B, DONG L, PIAO S H, et al. BEiT: BERT pre-training of image transformers[J]. arXiv Preprint, arXiv:2106.08254, 2022.
[9]	张燚钧, 张润清, 周华健, 等. 视觉基础模型研究现状与发展趋势[J]. 中国图象图形学报, 2025, 30(1):1-24.
[10]	GU Z, ZHU B, ZHU G, et al. AnomalyGPT: detecting industrial anomalies using large vision-language models[J]. arXiv Preprint, arXiv:2308.15366, 2023.
[11]	ZUO Z, DONG J, WU Y, et al. Clip3d-ad: extending clip for 3d few-shot anomaly detection with multi-view images generation[J]. arXiv Preprint, arXiv:2406.18941, 2024.
[12]	YANG T, CHANG L, YAN J, et al. A survey on foundation-model-based industrial defect detection[J]. arXiv Preprint, arXiv:2502.19106, 2025.
[13]	ZOU Y, JEONG J, PEMULA L, et al. Spot-the-difference self-supervised pre-training for anomaly detection and segmentation[C]// European Conference on Computer Vision. Cham: Springer Nature Switzerland, 2022:392-408.
[14]	WANG C, ZHU W, GAO B, et al. Real-iad: a real-world multi-view dataset for benchmarking versatile industrial anomaly detection[C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024: 22883-22892.
[15]	CAPOGROSSO L, GIRELLA F, TAIOLI F, et al. Diffusion-based image generation for in-distribution data augmentation in surface defect detection[J]. arXiv Preprint, arXiv:2406.00501, 2024.
[16]	TAI Y, YANG K, PENG T, et al. Defect image sample generation with diffusion prior for steel surface defect recognition[J]. arXiv Preprint, arXiv:2405.01872, 2024.
[17]	SNEHA K, SHENBAGA S V, SIVAMANI B, et al. Fabric defect detection using transfer learning[J]. International Journal of Research In Science & Amp Engineering, 2024: 62-71.
[18]	PANG S, ZHAO W, WANG S, et al. Permute-MAML: exploring industrial surface defect detection algorithms for few-shot learning[J]. Complex & Intelligent Systems, 2024, 10(1):1473-1482.
[19]	ZHANG Z, DONG C, WEI Z, et al. GCB-YOLO: a lightweight algorithm for wind turbine blade defect detection[J]. Wind Energy, 2025, 28(6):e70029.
[20]	LI J, CHENG M. FBS-YOLO: an improved lightweight bearing defect detection algorithm based on YOLOv8[J]. Physica Scripta, 2025,2:100.
[21]	LEI Z, ZHANG Y, WANG J, et al. Cloud-edge collaborative defect detection based on efficient yolo networks and incremental learning[J]. Sensors, 2024, 24(18):5921.
[22]	LIANG D, ZHANG H, HAN Q, et al. RasPiDets: a quasi-real-time defect detection method with end-edge-cloud collaboration[J]. IEEE Transactions on Industrial Informatics, 2025, 21(7):5525-5535.
[23]	工业互联网产业联盟, 中国信息通信研究院. 工业大模型技术应用与发展报告1.0[EB/OL]. (2023-12-26)[2025-06-28]. https://www.aii-alliance.org/resource/c331/n4797.html.
[24]	程荫. 中国AI赋能的工业质检解决方案市场分析[J]. 电子产品世界, 2021, 28(9):7-9,66.
[25]	人工智能创新应用洞察. 兴智新观察\|AI起航秀-联想乐眼工业质检方案[EB/OL]. (2023-08-31)[2025-06-28]. https://mp.weixin.qq.com/s/FkEeCklOGOadGkEOyYirFA.
[26]	中国钢铁工业协会. 宝钢股份AI转型的深度实践[EB/OL]. (2025-06-19)[2025-06-28]. https://www.chinaisa.org.cn/gxportal/xfgl/portal/content.html?articleId=e4ce7b24517ca66973dac996f437c7c47460e4c30d7990c2bc75ca88743284c6&columnId=268f86fdf61ac8614f09db38a2d0295253043b03e092c7ff48ab94290296125c.
[27]	机器视觉产业联盟. 工业质检母机软件:联想边缘大脑V3.0[EB/OL]. (2025-04-22)[2025-06-29]. https://mp.weixin.qq.com/s/3wU1Hx0fiEDgtzgQ7uUSYA.
[28]	联想研究院. 国内首台轮胎外观缺陷全检AI设备[EB/OL]. (2025-05-23)[2025-06-29]. https://mp.weixin.qq.com/s/BmsDqS0OZHaCHjRKUhV-zA.
[29]	联想研究院. 这台AI设备, 让纺织业质检效率飙升300%![EB/OL]. (2025-06-27)[2025-06-29]. https://mp.weixin.qq.com/s/8KbiIXsqJpA9HL0d2jq6fQ.

基于人工智能的工业产品缺陷检测关键技术与应用研究

Research on key technologies and applications of AI-enabled industrial product defect detection

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献 29

相关文章 1

编辑推荐

Metrics

本文评价