基于HOG+SVM的智能家居入侵检测系统设计

当前,大多数的智能家居视频监控系统仅仅完成简单的视频图像传输功能,不太智能;或是保持 PC常开,将视频图像数据传输到PC端进行相应的检测和识别,增加消耗.为解决该问题,系统以嵌入式技术为载体,利用 Cortex-A9处理器芯片搭载 Linux操作系统,结合机器视觉图像处理技术,融入 HOG+SVM机器学习方法来对视频中是否有外人入侵进行检测,并可通过 GPRS模块进行短信报警通知.同时可通过智能终端,登录指定网页或者相应的客户端进行实时查看和取证.实验结果表明,该系统数据传输稳定、算法准确度较高,可替代传统的智能家居视频监控系统.

lntelligent Home lntrusion Detection System Based on HOG+SVM

At present,most of the video monitor systems at smart home either simply perform simple video image transmission func-tions,or transmit the video image data to PC side for corresponding detection and identification, that increases the consumption.In order to solve this problem,the system uses embedded system technology as a carrier,using Cortex-A9 processor chip to equip with Linux op-erating system,combining machine vision image processing technology and the HOG+SVM machine learning method to detect whether there is outsider video in the video.Then it can send a SMS alarm notification via GPRS module.At the same time,you can login to the specified webpage or the corresponding client through the smart terminal to view and obtain evidence in real time.The experiment results show that the system has stable data transmission and is high accuracy,which can replace the traditional smart home video monitor system.