基于计算机视觉的人体危险行为智能预警方法
By acquiring video data from visual acquisition nodes in a wireless network, abnormal behavior is detected and tracked, and dangerous relay records are generated. This solves the problem of mismatch between human targets and adjacent nodes, and improves the accuracy and timeliness of early warning of dangerous behavior.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- LONGYAN UNIV
- Filing Date
- 2026-04-14
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies cannot effectively determine the mismatch results of human targets not appearing continuously between adjacent visual nodes under wireless networking conditions, resulting in insufficient effectiveness of early warning for dangerous human behavior in blind spots, occlusion areas, and abnormal movement scenarios between nodes.
By acquiring continuous video data from each visual acquisition node in the wireless network, human target detection, target tracking, and abnormal behavior recognition are performed. Exit records are extracted, and candidate successor nodes are calculated based on node adjacency relationships to generate dangerous relay records. Dangerous states are set to be continuously retained until the target is confirmed to reappear or to be in a non-dangerous state.
It improves the effectiveness of early warning for dangerous behaviors in blind spots, obstructed areas and abnormal transfer scenarios between nodes under wireless networking conditions, reduces the false alarm rate, and ensures continuous tracking of human targets and timely response to abnormal behaviors.
Smart Images
Figure CN122049992B_ABST