基于图像识别的高空抛物预警方法、装置及系统
By using an image recognition-based method for early warning of objects thrown from heights, and processing RTSP image streams with an improved 3DUNet model, real-time warning boxes for objects thrown from heights are generated. This solves the problem of identifying and warning of objects thrown from heights, improves processing efficiency, and reduces labor costs.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- UCLOUD TECH CO LTD
- Filing Date
- 2022-05-20
- Publication Date
- 2026-07-17
AI Technical Summary
The current situation is that high-rise building littering incidents are difficult to be actively identified and alerted by ordinary cameras, resulting in difficulties in evidence collection, low processing efficiency, and high labor costs, which prevents large-scale implementation.
A high-altitude object throwing warning method based on image recognition is adopted. RTSP image streams are acquired through network cameras, and after image normalization processing, they are input into an improved 3DUNet model for inference. The probability value and location of objects thrown from high altitudes are output. Real-time warning boxes are generated using graphics operations and dynamically output in the RTSP image stream.
It enables the identification and early warning of objects thrown from high altitudes using ordinary cameras, improves identification accuracy, reduces manual intervention, and can achieve large-scale deployment without replacing older cameras.
Smart Images
Figure CN115205774B_ABST