| 1,606 | 3 | 864 |
| 下载次数 | 被引频次 | 阅读次数 |
近年来随着社会经济和建筑行业的快速发展，安全问题日渐凸显，工人死亡和受伤事件屡见不鲜。钢结构焊接作为建筑行业施工中不可缺少的环节之一，需要一种高效快速的方法来检测工人在钢结构焊接中的施工风险。基于深度神经网络方法，构建图像文字描述模型，对钢结构焊接施工现场监控视频中提取的图像进行文字生成，并基于生成的文字识别现场工人的安全穿戴风险。通过网络爬虫和施工现场拍照收集图片数据，对图片数据进行数据增强和标注，制作成数据集；构建图像文字描述机器学习模型，使用建立的数据集对模型进行训练和验证，结果表明模型训练和验证识别准确度分别达到88%和85%;在文字识别结果的基础上采用关键词识别的方法进行风险结果判定。施工现场应用结果表明模型识别效果良好，并做出了准确直观的判定。
Abstract：In recent years, with the rapid development of social economy and construction industry, safety problems have become increasingly prominent, and workers' deaths and injuries are not uncommon. Steel structure welding is one of the indispensable links in construction industry, and an efficient and rapid method is needed to detect the construction risk of workers in the welding process. Based on the deep neural network method, this study constructs an image text description model, which generates texts from the images extracted from the surveillance videos of the steel structure welding site and identifies the safety wearing risks of workers on the site based on the generated texts. Image data is collected through web crawler and construction site photos, and the data is enhanced and labeled to make a data set. The machine learning of the image text description model is constructed and the model is trained and verified with the established data set. The results show that the recognition accuracy of the model training and verification reaches 88% and 85% respectively. On the basis of the text recognition results, the method of keyword recognition is used to judge the risk results. The application results on the construction site show that the recognition effect of the model is good, and accurate and intuitive judgment is made by the model.
[1] PARK M W,ELSAFTY N,ZHU Z H.Hardhat-wearing detection for enhancing on-site safety of construction workers[J].Journal of Construction Engineering and Management,2015,141(9):04015024.
[2] KIM B,CHO S.Automated vision-based detection of cracks on concrete surfaces using a deep learning technique[J].Sensors,2018,18(10):3452.
[3] LUO H B,LIU J J,FANG W L,et al.Real-time smart video surveillance to manage safety:A case study of a transport mega-project[J].Advanced Engineering Informatics,2020,45:101100.
[4] FANG Q,LI H,LUO X C,et al.Detecting non-hardhat-use by a deep learning method from far-field surveillance videos[J].Automation in Construction,2018,85:1-9.
[5] 石凌晨.基于计算机视觉的钢筋排布检测[D].重庆：重庆大学，2022.SHI Lingchen.Rebar arrangement detection based on computer vision[D].Chongqing:Chongqing University,2022.
[6] 梅杰，李庆斌，陈文夫，等.基于目标检测模型的混凝土坯层覆盖间歇时间超时预警[J].清华大学学报(自然科学版),2021,61(7):688-693.MEI Jie,LI Qingbin,CHEN Wenfu,et al.Over warning of concrete pouring interval based on object detection model[J].Journal of Tsinghua University(Science and Technology),2021,61(7):688-693.
[7] 杨静，张育飞，毛晓琦.施工作业面安全帽的深度学习检测方法[J].计算机应用，2020,40(S2):178-182.YANG Jing,ZHANG Yufei,MAO Xiaoqi.Deep learning detection method of safety helmet on construction working surface[J].Journal of Computer Applications,2020,40(S2):178-182.
[8] BHALEKAR M.D-CNN:A new model for generating image captions with text extraction using deep learning for visually challenged individuals[J].Engineering,Technology & Applied Science Research,2022,12(2):8366-8373.
[9] HE K M,ZHANG X Y,REN S Q,et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016:770-778.
[10] XU K,BA J,KIROS R,et al.Show,attend and tell:Neural image caption generation with visual attention[J].arXiv Preprint arXiv,2015:1502.03044.DOI:10.48550/arXiv.1502.03044.
[11] QU Z,MEI J,LIU L,et al.Crack detection of concrete pavement with cross-entropy loss function and improved VGG16 network model[J].IEEE Access,2020,8:54564-54573.
[12] CHEN Y D,ZHAO C,YU Z,et al.On the relation between sensitivity and accuracy in in-context learning[J].arXiv Preprint arXiv,2022:2209.07661.DOI:10.48550/arXiv.2209.07661.
[13] PAPINENI K,ROUKOS S,WARD T,et al.Bleu:A method for automatic evaluation of machine translation[C]//Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL),Philadelphia,2002:311-318.
[14] BROWN P F,PIETRA V J D,SOUZA P V D,et al.Class-based N-gram models of natural language[J].Computational Linguistics,1992,18(4):467-479.
[15] KINGMA D P,BA J.Adam:A method for stochastic optimization[J].arXiv Preprint arXiv,2014:1412.6980.DOI:10.48550/arXiv.1412.6980.
基本信息:
中图分类号:TP183;TU758.11
引用信息:
[1]王群力,张凯,刘丕养.基于深度神经网络的钢结构焊接施工风险识别方法研究[J].青岛理工大学学报,2025,46(01):1-7.
基金信息:
山东省自然科学杰出青年基金(JQ201808)
2025-02-20
2025-02-20