Font Size: a A A

Research And Implementation Of Multi-tracking Algorithm For Motion Picture Based On ARM11

Posted on:2016-10-14Degree:MasterType:Thesis
Country:ChinaCandidate:Y J HanFull Text:PDF
GTID:2308330479455529Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
With the continuing use of embedded technology maturely, video surveillance in embedded system is also turning intelligent video image processing sequence analysis from a simple video surveillance, the accurate, real-time, stable detection of the target information from video sequences is more and more important in indicator information. Intelligent Embedded Intelligent Video Surveillance in life has been widely used, such as: illegal punishment monitoring system, face recognition systems.In this paper, embedded target detection system platforms based on the Linux3.0.1 operating system, which hardware selection Samsung ARM11 platform equipped with USB camera, GPRS module peripherals. Embedded system capture YUV422 format video by USB camera, through a sequence of video frames preprocessing, Frame Differences to detect the target, when there is an objective in the frame there is alarm and SMS alerts and camera memory function, and in a moving video sequence detect and extract of the object, matching the target with multi-tracking characteristics of method based on histogram and centroid matching. video post-processing can be decoded after Ycbcr format decoded flow into the ARM11 that built-in multi-format hardware, H.264 compression for video coding by hardware codec, the video-coded display on the screen, the video stream encoded by the RTP protocol is transmitted to the router, by the radio network after the package sent to the remote Client PC after the Internet routing.This paper discusses the design and implementation of systematic approach. Firstly, It introduced the video sequence target detection, extraction, tracking methods and propose suitable for embedded devices to track matching algorithm; Secondly, It introduces system hardware and system software as well as various functional modules in the operating system layer to achieve multi-threaded programming; Finally, Achieving target tracking methods based on OpenCV library and do feasibility tests comparing environmental testing laboratory demonstrate that the system is designed to meet the requirements to achieve the desired effect.
Keywords/Search Tags:video surveillance, Linux operating system, target detection, target matching, OpenCV
PDF Full Text Request
Related items