基于监控相机的道路设施遮挡检测、背景建模及异常检测方法
By using instance segmentation models and background modeling methods, the system automatically monitors the occlusion and anomalies of road facilities, solving the problem of low efficiency in existing technologies. It enables real-time detection and anomaly identification of facilities that do not have network communication capabilities, thereby improving traffic safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ANHUI TRANSPORT CONSULTING & DESIGN INST
- Filing Date
- 2024-07-22
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies cannot effectively and promptly detect and monitor abnormalities such as obstruction, damage, missing parts, and positional deviation of road infrastructure without network communication capabilities, such as safety barriers, lane lines, and stop lines, leading to traffic safety hazards. Furthermore, relying on manual inspections is inefficient.
By extracting information on road facilities and traffic participants from video frames of surveillance cameras using an instance segmentation model, constructing a directory index, using the minimum bounding rectangle to determine occlusion, and combining background modeling and anomaly detection methods, automated monitoring and anomaly identification of road facilities can be achieved.
It enables real-time occlusion detection and anomaly monitoring of road facilities that lack network communication capabilities, reducing personnel and time costs and improving the level of informatization and digitalization of urban roads and highways.
Smart Images
Figure CN118918544B_ABST