基于多元数据融合的现场安全违章自动感知技术研究

打开文本图片集

中图分类号：TP274 文献标志码：A文章编号：1001-5922（2025）06-0143-04

Abstract：With the rapid development of science and technology， the requirements of operation standard control are getting higher and higher. In order to standardize the system of on-site operation，improve the operation control pro⁃ cess， establish a new safety management scheme， and improve the safety supervision system. In this paper，through the detection of illegal behavior by random forest algorithm， face recognition technology is designed to realize the in⁃ telligent management method of work ticket，and OCR recognition technology is used to automatically identify and au⁃ dit work ticket，so as to realize the whole process monitoring of operation，the automatic recognition of work ticket text image and the automatic recognition algorithm of illegal action of on-site personnel，and obtain the illegal information data in real time，automatically and accurately. The results show that this study realizes the automatic perception sys⁃ tem of safety violation based on multivariate data fusion technology. The research results of this project will lay a the⁃ oretical foundation for the development and application of intelligent safety supervision system.

Key words：safety violation；automatic perception；keyword extraction；face recognition

随着科技水平的不断提高，传统的管控方法已无法满足规范化作业要求，多元数据融合的安全违章自动感知系统是一个重要的研究方向。（剩余5721字）