一种施工安全的图像检测方法
By employing image processing techniques such as inner product singular value decomposition and the correlation matrix method, violations at construction sites are automatically identified, solving the problem of low efficiency in manual supervision in existing technologies and achieving efficient and accurate construction safety monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHENGZHOU UNIVERSITY OF LIGHT INDUSTRY
- Filing Date
- 2022-05-20
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, on-site safety monitoring relies on manual supervision, which is labor-intensive and inefficient, and makes it difficult to effectively identify minor problems. In particular, it is difficult to efficiently detect safety hazards such as violations, improper clothing, smoking, and dangerous actions in complex environments.
A feature extraction method based on inner product singular value decomposition and an association matrix method for matching violation features are adopted. Through image processing technology, the violation behavior at the construction site is automatically identified, including reducing redundant information in the image, feature extraction and matching, and using feature parameters to determine the violation category.
It has achieved automated construction safety inspection, reduced manual labor intensity, improved inspection efficiency, effectively identified and alerted to violations, and improved the efficiency and accuracy of safety monitoring at the construction site.
Smart Images
Figure CN115331043B_ABST