Video Cloud for Integration
The rate of deployment for Safe City solutions is growing in part due to advances in video surveillance storage systems.
By Xia Dasheng
Architect, Government Solutions
Enterprise Business Group, Huawei Technologies Co., Ltd.
Traditional Digital Video Recorder (DVR) and Network Video Recorder (NVR) distributed storage modes hinder resource sharing. To help, Huawei built its Video Cloud solution for file sharing across surveillance system departments and regions.
Cloud access nodes are deployed at police stations and other official sites. These virtual storage machines load Huawei or Huawei partner surveillance software to record, stream, fast-forward, and search. The access nodes support open stream storage, rapid convergence, and video data protection to ensure mature, stable, and reliable level-1 storage.
Also, centralized storage nodes are deployed at the data-link level in a shared pool for analysis and archival access. Storage resources are integrated with upper-layer service systems to perform multiple real-time video invocation and facial recognition services, among others. The central storage nodes ensure adequate single-system capacity and N+M backup to a resource pool of logical nodes.
Featuring centralized, multi-level networks, Huawei’s Video Cloud solution supports multiple, flexible deployment modes. For example, small- and medium-sized cities can adopt centralized deployments while larger cities can establish control centers in several locations.
Wanted: Intelligent Video Analysis
Vast amounts of original video provide essential information to solve cases; however, traditional investigative methods waste labor and material resources. For example, in China, more than 500 police officers spent thousands of hours over a 24-hour period reviewing video in multiple formats to obtain 15 seconds of relevant footage. While this example may seem extreme, human screening has long been the only available technology to do the work. Given the overwhelming numbers of available public and private video recordings, the security industry is motivated to acquire the most intelligent and efficient video analysis possible.
The Huawei Video Content Management (VCM) platform performs real-time behavioral analysis and structured processing on images sourced from intelligent video cameras and other front-end data acquisition devices. The analysis covers a range of issues from object abandonment to crowd density. This VCM platform functionality improves the effectiveness of the system to identify incidents live. The image investigation and control functions deployed at video cloud access and central nodes support the following:
- Efficient case management incorporates data asset inputs, comparative analysis, and statistical management.
- Intelligent search provides a series of search conditions — location, time, action, and type — to filter large quantities of seemingly unrelated information, making it possible to quickly and accurately spot target patterns within massive video data stores.
- Video abstraction technology separates moving objects from their backgrounds and then condenses the video into clips to shorten browsing time and reveal the value of surveillance video during a criminal investigation.
- Virtual checkpoint technology enables standard cameras with a software assist to perform automated checkpoint control of license plates and human faces.
Road safety systems in cities can generate a variety of data, including license plate information and surveillance video. ePolice and intelligent traffic system checkpoints are used to provide critical traffic management information and case detection assets. The intelligent analysis function of the ePolice system detects, captures, and records a wide range of traffic violations. The system records vehicle locations and times around the clock.
Head-on cameras automatically capture an image of each oncoming car’s license plate for upload to the command center, which then generates the identifying information for storage in a checkpoint information database for query and statistics. The command center system reports on vehicles involved in serious infractions.
Although video clouds resolve many data-sharing and intelligent analysis issues, the public security system still faces the following challenges:
- Video information cannot automatically be matched with crime cases.
- Performance of control platform nodes has not kept pace with increased numbers of checkpoints.
- Traditional checkpoint systems support only local storage, so many municipal police departments are only able to conduct historical data analysis.
Because information is scattered across databases, multi-database operations are required to complete the analysis. To address this, Huawei has utilized Big Data technologies to perform comprehensive analysis of structured data generated from VCM and vehicle records, photos, and related data. This analysis allows for more clues to be found quickly.
The use of ‘video in the cloud’ grows more complex as its scale expands. The Huawei Video Cloud platform offers distinct advantages in handling complicated transitions. With a cross-region video cloud platform featuring multi-layer sharing and Big Data analytics, Huawei has pioneered and continues to be a leader in the field of cloud surveillance.