Research On Spatio-temporal Context Algorithm For Video Surveillance

Based on the analysis of the recent theory and the algorithms of target tracking,An improved algorithm is proposed according to the spatio-temporal context tracking algorithm.The spatio-temporal context algorithm takes the advantages of the context information of the surrounding time and space to track the target This algorithm has good real-time performance and strong robustness,and is promising for applications.However,the conventional spatio-temporal context tracking algorithm suffers from the difficulties to recognize the target after it is hidden for a long time.In this paper,the characteristics of the spatio-time context algorithm is combined with the speeded-up robust features algorithm to propose an improved spatio-temporal context algorithm,which can be adopted to detect effectively the error updating of the the spatio-time context and correct it.In addition,when the target is seriously blocked,the target is retrieved and the accuracy of the algorithm is improved compared with the conventional algorithm.The experiment proves that the improved algorithm has been effective.A video wireless monitoring system,embedded with the improved algorithm,is constructed for tracking suspicious targets.The core module of the hardware is the development board of Raspberry Pi,combined with Logitech C270 camera make up the video acquisition side.The purpose is achieving acquisition of the monitoring system image acquisition.Further,the embedded processor architecture and operating system were analyzed and configured,through the WIFI wireless network to achieve the Raspberry Pi dispatch between the development board and the computer Socket communication.The use of wireless networks achieve the client and server data between the real-time transmission.Based on OpenCV visual library to achieve the wireless video compression,in the computer application of OpenCV visual library to achieve the system video reception and playback.Utilizing the MFC,an easy operating human-computer interaction interface is developed.Combined with the above technology,the video wireless monitoring and suspicious target tracking can beimplemented.Verification results show that the test system is reliable and stable,indicating that it is practical for applications.