基于传感数据融合与SIFT特征匹配的楼宇安防智能监控方法  

Intelligent Monitoring Method for Building Security Based on Sensor Data Fusion and SIFT Feature Matching

作　　者：姜业栋 JIANG Yedong(Honeywell(Tianjin)Co.,Ltd.Beijing Branch,Beijing 100015,China)

机构地区：[1]霍尼韦尔(天津)有限公司北京分公司,北京100015

出　　处：《科技和产业》2025年第7期26-30,共5页Science Technology and Industry

摘　　要：为提高安防监控结果的可靠性,实现对楼宇的智能监控,基于传感数据融合与尺度不变特征变换(SIFT)特征匹配,开展楼宇安防智能监控方法的设计研究。安装摄像头、红外探测器、震动等传感器,采集楼宇安防数据,引进加权平均法,对预处理后的数据进行融合与点云数据压缩;将彩色图像转换为灰度图像,提取关键点(特征点)和对应的描述符,设计基于SIFT特征匹配的楼宇安防监控图像自动拼接;在拼接后的视频帧中,使用帧间差分法识别运动目标,设定阈值以区分前景(运动目标)和背景,实现视频运动目标的跟踪与智能监控。对比实验结果表明,设计的方法实际应用效果良好,可以精准识别到监控界面中出现的所有人物,满足楼宇安防智能监控需求。To improve the reliability of security monitoring results and achieve intelligent monitoring of buildings,research and design of building security intelligent monitoring methods were carried out based on sensor data fusion and scale invariant feature transform(SIFT)feature matching.Cameras,infrared detectors,vibration sensors were installed,building security data was collected,weighted average method was introduced to fuse and compress preprocessed data with point cloud data.Color images was converted into grayscale images,key points(feature points)and corresponding descriptors were extracted,and automatic stitching of building security monitoring images was designed based on SIFT feature matching.In the spliced video frames,the inter frame difference method was used to identify moving targets,and a threshold was set to distinguish between foreground(moving targets)and background,achieving tracking and intelligent monitoring of video moving targets.The comparative experimental results show that the designed method has a good effect in practical applications.This method can accurately identify all characters appearing in the monitoring interface,meeting the requirements of intelligent monitoring for building security.

关 键 词：传感数据融合 点云数据压缩 智能监控 安防 楼宇 尺度不变特征变换(SIFT)特征匹配 

分 类 号：TN949.21[电子电信—信号与信息处理]