| The traditional video monitoring system is limited to the local area network or wide area network to operate the monitoring data in the way of artificial assistance,and to store the data in the traditional storage architecture such as the storage area network.Generally,this kind of architecture used for data storage had the characteristics of non intelligence,low reliability and low storage efficiency,which makes the current traditional video monitoring system unable to keep up with the development of present situation Stage.Therefore,aiming at the problems that the traditional storage architecture can not guarantee the high reliability of data and the waste of storage resources,this paper designs the interaction mechanism between the intelligent data processing center and the trusted data support combining with computer vision and distributed database,and realizes the intelligent and high trust cloud storage platform VCeph,which is dedicated to monitoring video data storage.This paper starts with a series of operations before data storage,makes intelligent filtering and selection,binds the characteristic information contained in the data with the storage object file,optimizes the high trustworthiness of the data and hardware storage resources,etc.the main work are as follows:1.An intelligent data processing center module is proposed,which is used to intelligently filter the key frame images of "warm and cold" surveillance video data,so as to select data sources to simplify the storage of data,reduce the multiple redundant information of data between video frames,and identify the attribute features of the selected data sources,mainly including the facial information and torso information of the characters and so on.2.A trusted data support interaction mechanism is designed to achieve the metadata information acquisition function of the data object to be stored.Combined with the information of intelligent data processing center filtering and feature attribute recognition,it is bound to the trusted support platform according to the requirements of high paradigm,so as to solve the problems of information high credibility and retrieval efficiency that traditional cloud storage cannot guarantee.3.Taking people as the starting point,only storing the key frame containing people can not only reduce the multiple redundant information existing in the monitoring video data,but also improve the efficient management of the data onto the cloud platform,optimize the resource utilization of the storage hardware,and reduce the investment cost of human and material resources.Through the construction and performance test of intelligent platform,trusted cloud assurance platform and overall cloud platform,it is found that: 1.The dedicated trusted cloud storage platform can intelligently extract the people's characters in the key frame,and recognize multiple attributes information of the characters,with the recognition accuracy of more than 91%;2.The trusted data assurance interaction mechanism of cloud platform realizes the high reliability of data and can also improve the efficiency of informationretrieval by more than 40%;3.The cloud platform can find the failure of storage equipment and automatically repair,the network bandwidth and disk read-write bandwidth support the read-write of big data,and save lots of hardware resources. |
