摘要
视频图像技术在智能安防中的应用已从传统监控转变为行为建模与风险干预,研究重点聚焦结构提取、时序表达与响应机制。构建融合图像接入、边界提取与长短期记忆(Long Short-Term Memory,LSTM)行为建模的系统架构,在风险判断中引入三维评分函数与非线性惩罚机制,形成多通道融合与闭环响应链。系统部署于轨道交通枢纽 区,验证方案在多方面的性能优势,具备良好的实用适应性与工程推广价值。
The application of video image technology in intelligent security has shifted from traditional monitoring to behavior modeling and risk intervention.The research focus is on structure extraction,temporal expression and response mechanism.This paper constructs a system architecture that integrates image access,boundary extraction and Long Short-Term Memory(LSTM)behavior modeling.In risk judgment,three-dimensional scoring functions and nonlinear penalty mechanisms are introduced to form multi-channel fusion and closed-loop response chains.The system is deployed in the rail transit hub area to verify the performance advantages of the scheme in multiple aspects,and it has good practical adaptability and engineering promotion value.
作者
韦观泳
卢伯宪
WEI Guanyong;LU Boxian(Guangxi Urban Vocational University,Chongzuo,Guangxi 532200,China)
出处
《智能物联技术》
2025年第4期130-133,共4页
Technology of Io T& AI
关键词
视频图像技术
行为建模
风险评分
联动响应
智能安防
video image technology
behavioral modeling
risk score
linkage response
intelligent security
