Artificial intelligence plays an important role in the fields of medical and health care, energy power, transportation and aerospace, language and image recognition, and it is also worth looking forward to in fields such as security. The application of transformative technologies such as artificial intelligence, deep learning, video structuring technology, Internet of Things technology, and big data analysis has made security video surveillance more and more powerful. AI-based intelligent identification and analysis technology has basically become the mainstay of video surveillance. standard.
Based on the AI video intelligent analysis capability of the intelligent analysis gateway + EasyCVR video fusion platform, it can intelligently identify and analyze the video images in the monitoring scene, and provide recognition, capture and comparison of faces, human bodies, vehicles, fireworks, objects, behaviors, etc. Yes, alarm and other services. At the same time, based on the technical architecture of cloud-edge-device collaboration, the following four capabilities can be realized to meet the diverse scenario needs of the industry.
1) Cloud-edge-device collaboration capabilities
By expanding the computing power of the cloud, the edge cloud can go deep into edge application scenarios that cannot be covered by traditional clouds. Terminal side and edge data can be reversely connected to the cloud center, supporting data collection from different regions, multiple business systems, multiple heterogeneous data sources, and devices, and easily realizing cross-level data exchange and application collaboration. Data travels from the device to the edge to achieve "device-to-edge", and from the edge to the cloud to achieve "data into the cloud".
- The device side focuses on multi-dimensional perception data collection and front-end intelligent processing;
- The edge side focuses on perception data aggregation, storage, processing and intelligent applications;
- The cloud center focuses on the fusion of multi-dimensional data including IoT data, as well as multi-dimensional analysis applications based on AI and big data.
2) AI intelligent recognition ability
- Face recognition/comparison
Support real-time video monitoring and capture, extract and analyze facial features in pictures and video sources, and realize multi-face detection and capture, face recognition, traffic statistics, face comparison search, face database management, etc. It can also support recording and real-time comparison with the control list and alarm. It can be used for control and verification at various checkpoints, such as shopping malls, buildings, communities, stations, streets, airports, ports, entertainment and other public places and important entrances and exits. Among them, it meets the needs of personnel identification in different scenarios.
- Vehicle detection/license plate recognition
It supports the capture, detection and recognition of motor vehicles/non-motor vehicles in the video, including vehicle type (characteristics), brand (vehicle logo), color, license plate and other information, and can recognize multiple license plates in the image at the same time. When a vehicle is detected in the specified area (area where parking is not allowed), the vehicle illegal parking alarm will be triggered immediately, and the alarm information will be pushed to the staff through various methods such as phone call, text message, email, and WeChat notification.
- Helmet detection/reflective clothing detection
Through real-time video monitoring and early warning whether on-the-job workers have taken safety precautions as required, such as: whether to wear safety helmets correctly, whether to wear reflective clothing, protective clothing and other work clothes. If anomalies are detected, an alarm is issued.
- Person/area intrusion detection
Real-time detection of established areas, such as: people intruding into restricted areas, dangerous areas, important areas, etc., timely capture the intruders and call the police, and can also be linked with on-site voice prompts to facilitate timely stop.
- Smoke/Flame Identification
Accurately identify the smoke and flames in the monitoring area in real time. Once smoke and fire are detected, an alarm will be issued immediately, and the alarm information will be pushed to the relevant management personnel for processing in time. It can also be linked with fire-fighting facilities for sprinkler and fire extinguishing operations.
- Wearing a mask for identification
Whether the person wearing a mask is detected in the monitoring area, once a person not wearing a mask is detected, the voice player can be linked to give a voice reminder.
- other identification
Other behavior identification includes: personnel on-duty and off-duty identification, dangerous behavior identification, such as: stay, suspicious wandering, etc.
3) Computing resource scheduling capability
The intelligent analysis gateway can support on-demand aggregation of data, as well as flexible and fine-grained scheduling capabilities of AI computing resources. Through the establishment of AI algorithm model specifications, various AI algorithms can be managed and scheduled in the algorithm warehouse. At the same time, it can also manage and schedule the computing storage resource pool, data resource pool and AI algorithm warehouse resources in the domain to improve the resource utilization efficiency of AI computing. Realize the flexible access of algorithms, the unified scheduling and allocation of AI computing power resources, and the unified display of intelligent analysis results, etc.
-
Support unified management and control of cloud computing nodes, edge computing nodes, computing power nodes, network resources (including computing node internal network and wide area network), etc., and perform unified scheduling of computing power resources, corresponding network resources, storage resources, etc. according to business needs.
-
Real-time monitoring of business traffic, dynamic adjustment of computing power resources, efficient processing of various tasks and integrated output, and elastic scaling of resources and optimization of computing power allocation under the premise of meeting business needs.
4) Video data sharing capability
The video fusion platform can solve the problems of poor overall planning, non-uniform standards, insufficient sharing, and single application of video resource construction at the current stage, gather video resources at all levels, and form video sharing capabilities based on intelligent big data. At the same time, AI technology is also gradually sinking from the central cloud to the edge and end. With the support of AI, the video fusion platform can realize intelligent search, filtering, early warning, data aggregation and visualization of data information, and improve the accuracy of event early warning. and reduce the rate of false alarms, thereby improving regulatory efficiency and greatly shortening the traceability time.
The platform has powerful data access, processing and distribution capabilities, high openness, easy deployment, flexible expansion of functions, and easy integration and docking with third parties to meet the needs of more advanced scenarios.
The AI intelligent video surveillance solution (intelligent analysis gateway + EasyCVR video fusion platform) based on cloud-edge-end collaborative management, using AI, Internet of Things, cloud computing, big data and other technologies, can meet data perception, intelligent detection, and intelligence based on video services. Analysis, intelligent alarm and other requirements can be widely used in intelligent security, communities, campuses, scenic spots, parks, gas stations, chemical plants, construction sites, factory areas, electric power and other scenarios. The AI project of TSINGSEE Qingxi Video also supports trial and error in small batches, and everyone is welcome to learn about it.