| 基于 KanBIM 和物联网的电力工程现场安全行为识别方法研究 |
| |
| 引用本文： | 贾云博,张永炘,黄伟文,何东城,蔡岱翰.基于 KanBIM 和物联网的电力工程现场安全行为识别方法研究[J].工程管理学报,2022,36(5):154-158. |
| |
| 作者姓名： | 贾云博 张永炘 黄伟文 何东城 蔡岱翰 |
| |
| 作者单位： | 广东创成建设监理咨询有限公司 |
| |
| 摘 要： | 针对电力施工现场的安全行为问题，研究了基于KanBIM 和物联网技术的电力工程现场安全行为识别方法。对传统Faster R-CNN 模型进行改进，提高其视频监控识别的准确率，通过增加模型锚点数量、改进损失函数和特征图像插值预测的方式，提高视频监管效果。实验证明，该算法在对于安全帽管理和不安全行为的识别方面，准确率和精度均大幅提高，优势明显，可用于电力施工现场的安全行为识别工作。
|
| 关 键 词： | KanBIM 电力工程 物联网 安全行为识别 |
Research on Identification Method of Safety Behavior in Power EngineeringSite Based on KanBIM and Internet of Things |
| |
| JIA Yun-bo,ZHANG Yong-xin,HUANG Wei-wen,HE Dong-cheng,CAI Dai-han.Research on Identification Method of Safety Behavior in Power EngineeringSite Based on KanBIM and Internet of Things[J].Journal of Engineering Management,2022,36(5):154-158. |
| |
| Authors: | JIA Yun-bo ZHANG Yong-xin HUANG Wei-wen HE Dong-cheng CAI Dai-han |
| |
| Affiliation: | Guangdong Chuangcheng Construction Supervision Consulting Co. Ltd. |
| |
| Abstract: | Aiming at the problem of safety behavior in a power construction site，this paper studies the method of powerengineering site safety behavior identification based on KanBIM and Internet of things technology. The traditional fast r-cnn model isimproved to enhance the accuracy of video surveillance recognition. The video surveillance effect is strengthened by increasing thenumber of model anchors，improving the loss function, and refining feature image interpolation prediction. Experiments show thatthe algorithm has obvious advantages in terms of accuracy and precision in helmet management and unsafe behaviors identification,and can be used to identify safe behaviors in power construction sites. |
| |
| Keywords: | KanBIM power engineering internet of things safety behavior identification |
|
| 点击此处可从《工程管理学报》浏览原始摘要信息 |
|
点击此处可从《工程管理学报》下载全文 |
