一种基于区间视频流的地铁隧道渗水风险预警方法
By calculating the integral parameters of global image brightness gradient variation and separating tunnel wall texture using a decoupled encoder, combined with a dual-scale time window, the false alarm and inaccurate benchmark update problems of existing water seepage identification methods are solved, and accurate early warning of water seepage risk in subway tunnels is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHAANXI HONGQI TESTING TECH CO LTD
- Filing Date
- 2026-05-09
- Publication Date
- 2026-07-17
AI Technical Summary
Existing video recognition methods struggle to effectively distinguish between transient optical disturbances caused by trains passing each other, headlight sweeping, and reflections from wet surfaces and the actual edges of water stains when dealing with water seepage in subway tunnels, leading to false alarms or inaccurate benchmark updates.
Suspend interception by calculating the integral parameters of global image brightness gradient variation, use decoupled encoder to separate the steady-state hidden vector sequence of tunnel wall texture, and form a long-term background update window and a short-term observation window based on dual-scale time windows to determine the topological sphere center reference coordinates and output seepage prevention instructions.
It enables accurate early warning of water seepage risk under conditions of train passing and direct sunlight, reduces false alarms, ensures the accuracy and stability of benchmark updates, and provides a continuous data processing path.
Smart Images
Figure CN122157131B_ABST