Product overview
The Video & Image Analysis System, which is in line with criteria including Video and Image Analysis System for Public Security (GA/T 1399) and Video and Image Information Application System for Public Security (GA/T 1400), is provided to users in various industries by integrating video resources, vehicle bayonet resources, face bayonet resources, RFID data resources, WIFI probe resources, among other data, and relying on hybrid cloud computing technology, intelligent video structuring analysis technology and big data technology. It aims to provide a uniform ecosystem that provide video & image analysis services and big data application services.
Functional performance
1. Meeting the public security standard on the capacity of collection and sharing of video and image information
The system is compatible with and in line with criteria including Video and Image Analysis System for Public Security (GA/T 1399) and Video and Image Information Application System for Public Security (GA/T 1400), and provides standardized and non-standardized access, convergence, integration, sharing & distribution and application services covering different manufacturers, different terminals and various video and image resources.
2. Provision and integration of public security video & image resources and social public video & image resources
The system has established the video & image big data service system with person, vehicle, object and case as the core, accessed and integrated various types of public security video & image resources, social resources and other perception and identification resource information, to provide comprehensive services and support for public security and social applications.
Diagram of access and integration of public security video & image resources and social resources
3. Intelligent identification, analysis and application of videos based on deep learning
The CPU+GPU collaborative computing capacity is fully used to achieve dynamic dispatch of resources, and the advanced intelligent analysis technology is adopted to automatically extract valuable information about human, face and vehicle details from real-time/historical/file videos & images, to form a structural description of type and color of the target, etc. This will provide data basis for the event control, comprehensive deployment and control, as well as judgment & analysis before, during and after the event., so as to ensure efficient information search and retrieval ,and improve the effectiveness of public security videos.
’Real-time structuring demonstration of videos
4. Comprehensive and multi-dimensional information deployment & control capacity
It provides comprehensive and multi-dimensional deployment & control, detection and early warning capacity for various aspects and various information types of one or more targets at important sites such as station, wharf, airport and shopping mall.
Diagram of comprehensive deployment and control
5. Multi-level video & image big data cascading application that enables network-wide response in case of inquiry at one point
Based on various levels of video & image analysis systems, it provides cross-region, cross-network and cross-system deployment & control and resource retrieval application of networking resources at various levels including provincial, municipal and county levels, so that it is able to give a network-wide response in case of inquiry or control at just one point.
6. Integrated spatio-temporal command and application of resources with map as the core
By integrating map visualization and big data retrieval analysis, it incorporates the spatio-temporal information of various resources and data analysis application into the map, to provide integrated spatio-temporal command and application capacity.
Integrated spatio-temporal resource retrieval
Demonstration of retrieval outcome
7. Full-text retrieval and searching for images with images based on various types of video & image big data
Services are extracted based on the semantic analysis and image features of the input retrieval information, to enable full-text retrieval of the video & image big data and searching for images with images, and this move helps to rapidly discover the clue information of person, vehicle, WIFI and object that meet relevant requirements among a hundred billion pieces of data.
Searching for images with images among the video & image big data
8. Tracking of human target in multiple scenarios and with several cameras
It is applicable to searching for person in complicated security scenarios, and tracking the target according to the human feature, thus human targets will be obtained in different scenarios and from different angles, tracking of movements of suspects will also be formed.
Tracking of human targets
9. Collision analysis and relation mining of video & image big data
Based on big data analysis technology and deep learning technology, and in light of demands of real applications, it provides analysis models for resources such as face, human body, vehicle, WIFI and RFID, to swiftly discover the target, track the target’s movements, conduct relation collision and judgment & mining among a hundred billion pieces of recorded information; this will provide technical support and data support for the analysis & retrieval, judgment & comparison of targets.
.jpg)
10. Chain technical application of various kinds of video and image data
It provides chain technical applications based on abundant technical models of the system, to achieve related application of technical methods, and thus ensure traceability of the investigation process.
Chain technical application of various video & image data
11. High extension, configuration and support capacity of hardware resource based on project scale
High density server, ordinary server and high-performance GPU server are selected to be configured and combined flexibly based on the project scale as well as the accessed and processed data resources, which provides low-carbon and environmentally-friendly application support that features flexible extension while ensuring the high performance operation of the system.
Technical features
1. Standardized multi-source isomerous data resource service capacity
Complying with the standard of Video and Image Information Application System for Public Security (GA/T 1400), it provides access, storage, sharing and analysis of data from multiple sources and at various levels, and has the capacity of delivering standard interface services.
2. Precise and intelligent analysis and processing of videos & images
It enables efficient automatic extraction, analysis, identification and parsing of video or image content through the intelligent identification algorithm of persons, vehicles and objects that is based on deep learning.
3. Efficient CPU+GPU collaborative computing capacity
The system automatically analyzes the task type based on the business needs, dispatches CPU and GPU resources as per necessary computing demands; by making breakthroughs in data and computing resources, it can make full use of the overall system capacity, and has the super-computing capacity.
4. Rapid big data retrieval and data mining capacity
It enables integration and collision of various types of data for the mass video & image information among the video big data, thereby improving the overall value of the resources in combination with information such as cases and WIFI, and providing the ability of processing a hundred billion pieces of data.
5. Capability of processing multiple business applications
The system has applied multiple technical methods and business analysis approaches based on business needs, to facilitate counter-terrorism and stability safeguarding, command and disposal, public security and counter-terrorism, case investigation and solving, public security, enforcement supervision and services related to livelihood improvement, etc., and cultivate the capacity of rapid and efficient processing.
