Design Of Itelligent Digital Audio And Video Surveillance System Based On DM3730

With the development of video surveillance technology and the vediosurveillance system requirement improvement, the surveillance system has not onlystay in the ordinary course of network video monitoring. People need the computeraid analysis of vedio content to complete the target detection, tracking and behavioranalysis. So that the fatigue, the hard to real-time monitoring of each video, the pooraccuracy of alarm, the long response time of alarm and the difficult video dataanalysis can be eliminated. Therefore the intelligent video surveillance system willbecome the future development trend of the video monitoring system with improvethe ability of video surveillance.In this thesis, based on the TI DM3730processor beagleboard-xm developmentkits, the design has realized the collection of audio and video abnormal eventdetection as well as alarm, parameter real-time receive, algorithm in real-time update,real-time save log files, parameter transmission and storage before encryption, whichintegrates intelligent digital audio and video surveillance system.This thesis discusses the Beagleboard-xm hardware platform and the audio andvideo surveillance basic principle and related algorithm. The construction, drive andtransplantation of dependent libraries of embedded Linux system are discussed in thethesis. Firstly, the thesis introduces the overall design scheme of intelligent digitalaudio and video surveillance system. The function the system can realize and thefunction of each module are described in detail. The TI davinci technology the systemused, the drivers which the audio and video rely on, the program libraries as well asthe basic technology such as network communication method are made the briefanalysis and introduction. Then, the structure and performance of hardware platformare described and evaluated in detail in view of the beagleboard-xm development kit.The thesis introduces the building process of software platform and soft framework ofthe system. So that the system can realize the basic function such as input, output andnetwork communication. Finally, the thesis introduces the current mainstream audioand video analysis algorithm. According to the characteristics of the embeddedplatform, the algorithms are transplantated, optimized and integrated into the systemsoftware. All functions of the surveillance and control system are realized in the end. Currently, the system can real-time collect video data from the network or USBcamera as well as gather audio data from pickup. The audio and video content areanalysed and the alarm information is produced. The alarm information will be sentthrough the network communication, the parameters are received and real-time logfiles are saved. The joint debugging with the server side is completed.