The rapid rise of edge computing _ fusion "AI + security"

At the moment, “AI+Security” technology has become a prominent school. In the era of security artificial intelligence, cloud computing is undoubtedly one of the key technologies. Its advantage is that it can realize resource sharing and make computing power more flexible. At the same time, because cloud computing aggregates all data into the cloud for processing, Therefore, problems such as unresponsiveness, high power consumption, and high network bandwidth requirements are also present.

For the above problems of cloud computing, Hikvision took the lead in introducing the edge computing concept into security, injecting AI capabilities into edge devices such as front-end cameras, enabling edge devices through high-performance computing chips and image recognition intelligent algorithms, and implementing video images at the edges. The detection, extraction, modeling and analysis of the target, the large amount of computational pressure of image parsing is evenly distributed to the large-scale edge computing resources of small particles, and only the refined structured valid data is uploaded to the cloud for processing, which can effectively reduce the transmission of the video stream. With storage costs, the computing and storage pressure of the cloud center is shared to maximize efficiency while enabling the terminal to achieve low latency and fast response.

So, what is the strangeness of reading and the edge calculation of the mouth? According to Wikipedia, Edge CompuTIng is also translated as "edge computing", which is a decentralized computing architecture that moves applications, data and services from network central nodes to network logic. Edge nodes to handle.

The "edge" in edge computing is a relative concept, referring to any computing resource and network resource from the data source to the cloud computing center data path. "Edges" are not limited to edge nodes such as edge servers, but also include cameras at the edge of the network, smartphones, gateways, wearable computing devices, and sensors. The basic idea is to use the existing computing power of the edge device to migrate all or part of the computing tasks of the application service program from the cloud center to the edge device terminal to reduce energy consumption.

In fact, edge computing is not new, and its principle of operation has been proposed many years ago. The reason why edge computing has risen in the near future is inseparable from the development of chips, algorithms, and data. The rapid rise of edge computing also provides artificial intelligence (AI) hardware and intelligent robots with the computing services required for high-speed interaction. Embedded artificial intelligence technology characterized by edge computing has begun to receive attention, and the role of edge computing can optimize resources and improve efficiency.

According to a recent report released by market research firm Research and Markets, the market average annual growth rate of edge computing is as high as 35.2%. In the field of security monitoring, edge computing also breeds a huge market. It is reported that in addition to Hikvision, Dahua, Huawei, Yushi and other leading companies have also followed up to deploy edge computing technology.

At the end of last month, Hikvision's "AI+: Insight into the Industry, Helping Change" theme forum, Hikvision first publicly released the company's AI Cloud architecture.

The AI ​​Cloud based on "cloud-side fusion" consists of three parts: cloud center, edge domain and edge node. It realizes the organic integration of edge computing + cloud computing, injects AI computing power into the edge, and empowers edge intelligence to achieve fast and efficient perception. The cloud focuses on the global analysis and cognition of AI data. The two complement each other with video AI data as the core, which can give full play to the advantages of edge computing agility and cloud big data computing.

It is reported that Hikvision configures the image analysis capability at the edge of the security system, so that the collected video stream is structured in the near direction, and the structured data enters the cloud for big data understanding collision. In the AI ​​technology and deep learning ability, the reasoning and human interaction To achieve front-end awareness and network-wide cognitive capabilities. Through this process, it is equivalent to decomposing the computational pressure of big data into large-scale local edge devices, and then transferring important information to the cloud center through structured processing, and doing big data correlation analysis in the cloud center. By understanding and disposing of different requirements, local data can be placed locally, and the requirements of big data correlation analysis processing can be placed in the cloud center. This hierarchical computing form can better satisfy customers' resource input and procurement costs. , network robustness, diversity of requirements and other factors.

Through the product demonstrations at the AMB meeting, Hikvision, which has been deeply involved in the AI ​​field for many years, has completed its deployment and technical reserves in the “AI Cloud” strategy. Hikvision's products such as squats, hunters, eagle eyes, and passenger flow statistics have placed calculations at the edge nodes. At the same time, some algorithms are deployed in the edge domain, such as AI server device blade, falcon, Facebook, Super Brain, etc., the cloud center is equipped with high-density GPU servers, integrated cloud storage products and cloud management platforms and other software and hardware products, which together constitute the AI ​​Cloud architecture.

If you look at it, in the era of security big data, the concept of "cloud-side integration" undoubtedly has unique advantages and charm. Cloud computing provides powerful global structured data reasoning analysis and resource management, while edge computing provides fast, agile, efficient, and accurate real-time response. The complementarity of the two will drive the application of artificial intelligence in the security industry to a new level.

According to IDC's forecast, by 2020, the Internet of Things will have 50 billion sensing devices, and 50% of computing will be on edge devices. Cloud edge convergence is a common model in the future.

It is reported that Hikvision has launched the application of AI Cloud+ industry solutions, and the project of “cloud-side integration” innovation concept has been applied in many industries, such as emergency command, people's livelihood service, urban operation, traffic management, business decision-making, etc. The field provides AI solution services.

Nowadays, not only Hikvision, but also some monitoring vendors and VCs have integrated deep learning algorithms into smart cameras. For example, license plate recognition and face recognition have been implemented at the front end.