Font Size: a A A

Research On Movement Image Detection Technology Based On Embedded Linux

Posted on:2018-03-22Degree:MasterType:Thesis
Country:ChinaCandidate:W LeiFull Text:PDF
GTID:2348330521450318Subject:Engineering
Abstract/Summary:PDF Full Text Request
Motion image detection is an important research content of video surveillance system to intelligent development.This paper studies the design and implementation of small intelligent home monitoring system based on ADSP-BF609.The system for unattended facilities,home and office space for real-time intelligent monitoring,so that manual monitoring from the boring monitoring to get rid of.In this paper,the commonly used methods of moving image detection are studied,including the basic principle of optical flow method,background difference method and adjacent frame difference method.In the Visual Studio 2010 development environment,the advantages and disadvantages of various algorithms are simulated by configuring Open CV visual library.In this paper,the BF609 dual-core processor is selected as the processing core of the moving image detection hardware.According to the modularization idea and the functional requirements of the moving image detection system,the video acquisition,image preprocessing,external storage,video output,system alarm Functional module design.Through the BF609 board on the TWI interface on the camera OV9650 control register initialization to achieve image quality control,through the parallel expansion interface EPPI1 and its dedicated DMA controller configuration parameters to complete the image acquisition and storage,through the PIXC processing unit parameter configuration,to achieve YUV To RGB image format conversion,video encoding chip(CH7024)to complete the video encoding output,audio codec(SSM2603)to achieve alarm function.In this paper,the embedded Linux operating system is used as a platform to establish a complete cross development environment on the PC,and the U-Boot and blackfin buildroot systems are transplanted into the motion image detection hardware platform in this environment.In order to make full use of DSP dual-core resources,the overall system software using asymmetric programming mode,CORE0 mainly in the control functions,including operating system operation,on-chip peripherals initialization,video image output and alarm function program,and image processing Of the complex tasks on CORE1 on the deal,including image acquisition and image processing functions,between the two cores through the MCAPI protocol to communicate.Combined with the characteristics of embedded Linux program development,the preparation of the relevant hardware drivers,video capture and output software programs,transplantation of the moving image detection program to achieve motion detection and alarm functions.The system is designed to test the moving target detection system.The result shows that the system can effectively detect the moving target and make corresponding alarm response,which has certain real-time and stability.In order to improve the applicability of the moving image detection system,it is possible to further increase the functions of face recognition and network communication.
Keywords/Search Tags:embedded Linux, image detection, smart home, ADSP-BF609
PDF Full Text Request
Related items