The big data platform of police affairs adopts the leading deep learning algorithm and advanced large data technology with independent intellectual property rights, which can achieve the unified access and integration of data from different sources such as mass monitoring cameras, virtual bayonet, human face bayonet, electronic tag and public security business data. It can also achieve secondary identification, intelligent query, map search through map, key object management and inspection and control functions; provide the technical tools set close to the combat for warning before the event, disposal during the event and judgment after the public security event; and achieve the visual display functions such as smart map, smart situation and view monitoring.
The system can be deployed separately to meet the needs of tagging search and map research through the map of massive bayonet pictures and videos, and can be used as the core part of intelligent video processing in the projects of smart city, safe city, intelligent transportation, three-dimensional security prevention and control, and three-in-one network.
(1) Mass and multi-data access, integration, storage and management
It can achieve the access, storage and management of massive institutionalized, semi-structured and unstructured data, and the elastic expansion of data storage and computing capacity according to the demand. It has the report data second-level query responsiveness and picture second-level information association responsiveness.
The system is compatible with the bayonet equipment of mainstream market manufacturers, which can achieve the unified integration of construction equipment of different manufacturers and different periods; the system can also access to video surveillance, electronic tags and public security business data, and realize their integration and management.
(2) Advanced model algorithm
Full-dimensional image secondary recognition: realize the real-time output of up to 25 vehicle character recognition results.
Intelligent map search: It doesn’t rely on the license plate, models, body color and other identified structured information, and adopts artificial intelligence core technology only from the vehicle shape, picture perspective, body pattern and other characteristics to extract abstract features, which is suitable for the tracking of vehicles after the accident and vehicles replaced, and the search and comparison of heterosexual vehicles with less structured training.
(3) Professional skills and tactics set
Based on years of experience in the public transport industry, it’s equipped with a professional set of technical and tactical tools to achieve the full-element image of the vehicle.
(4) Comprehensive application scenarios
Based on the needs of public security, it provides five application scenarios of trend cube, car search with God’s eyes, monitoring and investigation, big data judgment and smart map.
3. Application scenarios
(1) Public security applications
(2) Traffic police applications