Among the 200 million cameras in the country, there are a large number of "deaf and mute" devices. How can public safety be guaranteed?

The most taboo thing in the field of public safety is "redemption" after the fact. Under the new trend of artificial intelligence and the Internet of Things, we must break this drawback and establish a public safety active early warning and prevention system that is "promptly responsive, efficient, safe and reliable."

Public security is the key to ensuring the national economy and people's livelihood and maintaining social order. In recent years, thanks to the joint efforts of governments at all levels and all sectors of society, the public security situation has continued to improve, especially the "Snow Bright Project" and other "mass public security prevention and control projects" focusing on the networking application of public security video surveillance, which have penetrated into counties, townships, villages and other local grassroots organizations, becoming a strong backing for protecting people's lives and property safety.

However, in the era of the Internet of Everything, new security risks are coming one after another, and people have put forward new demands on the efficiency of public safety prevention. Therefore, how to use new technologies such as the Internet of Things, artificial intelligence, edge computing, and 5G to meet people's new expectations for public safety has become a top priority for the development of this field.

The current state of public safety: Straw man devices are a problem

In October 2018, residents of a residential community in Jining reported that due to the incomplete public facilities in the community, the doors of community units were stolen and the tiles on the roof eaves were also damaged. They thought that the surveillance would restore the scene, but when the police were investigating and collecting evidence, they found that the street lights in the community were not on at night, resulting in the installed surveillance being unable to capture clear images, making it like a "scarecrow" in the wheat field, just a decoration.

Coincidentally, in January 2020, a villager in Wuning County, Jiujiang City, Jiangxi Province reported that a marble mine in the area had been shut down for nearly two years, and the mining equipment that should have been left in the mine for processing was stolen at the end of last year. What was even more unexpected was that when the mine employees subsequently retrieved the surveillance footage, they discovered that the surveillance equipment in the village where the incident occurred had long been disconnected, becoming a deaf ear.

Of course, these incidents are not isolated cases. According to statistics, China has installed nearly 200 million surveillance cameras , and how to use these cameras and the data they generate more intelligently has always been a topic of discussion in the field of public security.

In the field of public safety, the most taboo thing is actually "redemption" of "closing the fold after the sheep have bolted". Under the new situation of artificial intelligence and the Internet of Things, we must break this drawback and establish a public safety active early warning and prevention system that is "timely response, efficient execution, safe and reliable". At present, two major problems have become the key to hindering public safety.

First, public security is facing the upgrade of a refined governance system. It is necessary to give full play to the role of the Internet of Things, big data, and artificial intelligence, and build a precise governance platform based on scientific and technological means. For example, in terms of scenarios, it is necessary to solve the problems of how small places can be safely connected to the public security network, industrial park safety production, and people entering and leaving hotels. In terms of technology, it is necessary to solve the problems of data validity, security, and multi-data fusion.

Secondly, public safety is a broad concept. Its coverage not only includes cities and counties, but also extends as far as the grassroots level in rural areas. Its coverage includes both large shopping malls and nine small places. Therefore, the difficulty and cost pressure faced in achieving in-depth coverage of the society's smart upgrades are naturally self-evident.

At the end of 2019, in response to the Ministry of Public Security's requirements to accelerate the transformation of scientific and technological achievements, Sunsea Intelligent Technology Co., Ltd. and the Third Research Institute of the Ministry of Public Security reached a cooperation in the field of artificial intelligence Internet of Things public security, jointly built a public security intelligent Internet of Things laboratory, and launched a professional-grade product in the public security field this year - AI Function Smart Box.

AI Function Smart Box

The AI ​​Function Smart Box is an AI edge computing terminal with scenario definition function, effective data perception and safe and reliable launched by Sunsea Intelligence and the Third Institute of Public Security for the practical empowerment of big data in the public security field. It has four core advantages:

The scene is adaptable, and the application, operation and maintenance management are perfect. Support multi-algorithm fusion, flexible function configuration; support local and remote equipment operation and maintenance, software upgrade

AIoT interconnection, edge linkage. Supports access to mainstream sensors and early warning devices in the industry; supports 4G and 5G transmission

Reuse and transform old cameras to achieve intelligent upgrades of ordinary cameras. Supports mainstream camera access methods such as RTSP, GB28181, and onvif; convenient northbound docking, provides standard API, and supports 1400 protocol

National security, stable and reliable. National security reinforcement ensures the security of equipment and data; industrial-grade design, surge protection, lightning protection, wide temperature range; Ministry of Public Security testing and certification

The “end-edge-cloud” structure is implemented to eliminate “dumb” devices

As we all know, most traditional public security prevention and control systems are built and deployed in a "cloud + terminal" manner, which means that all computing power will be realized in the cloud. The amount of data brought by the Internet of Things is incomparable to the past. The transmission of a large amount of video data to the cloud for processing will undoubtedly greatly increase the burden on the cloud; secondly, as the video clarity improves, the traffic occupied by the video will also increase exponentially in the future, which may also lead to transmission congestion and other problems.

AI function smart box application topology diagram

The AI-based smart box, by connecting to various surveillance cameras and sensors downstream and the video surveillance platform, operation and maintenance management platform and third-party applications upstream, realizes the implementation of the "end + edge + cloud" architecture model, enabling real-time, dynamic, interactive and integrated data collection and analysis capabilities in various scenarios.

For example, in the past, when front-end cameras detected the entry and exit of key personnel, they would often alarm through the cloud, but there would be a delay of several seconds during this period, and even longer delays if it was across regions. The non-real-time alarm undoubtedly led to delayed action. After the AI ​​function smart box empowers edge computing power, big data processing will take into account grassroots units, and quickly deploy control by understanding real-time dynamic information, thus solving the problem of "closing the fold after the sheep have been lost."

At the same time, the end side connected to the AI ​​function smart box does not need to be upgraded to smart, and the original surveillance camera can also be used to complete the intelligent transformation. This will undoubtedly greatly reduce the huge cost expenditure caused by equipment upgrades to a certain extent. At the same time, supporting multiple transmission methods greatly shortens the installation time and reduces the difficulty of construction.

In terms of security, there is no need to worry. Based on the respective advantages of the Third Public Security Institute and Sunsea Intelligence, the AI ​​Function Smart Box already has three major security elements. First, the AI ​​Function Smart Box has a unique trusted ID, just like a person's ID card, which is trustworthy and reliable; second, the AI ​​Function Smart Box has data encryption capabilities to prevent data from being tampered with or stolen; finally, the AI ​​Function Smart Box has secure device firmware, including software and hardware interfaces, to prevent the device from being invaded.

In addition to the security capabilities of the AI ​​function smart box itself, Sunsea Intelligence and the Third Public Security Institute have also created a complete security operation and maintenance management platform. Through the platform, managers can understand the operating status and security situation of the front-end equipment at the back end, and complete perception positioning, alarm processing, etc. Through the hardware and software operation and maintenance of a complete set of security mechanisms, a closed-loop security solution can be achieved.

At present, among the nearly 200 million front-end cameras in the country, the intelligent coverage rate is less than 20%. It can be said that the intelligentization of the public security video surveillance network system still has a long way to go.

In July this year, AI function intelligence has been deployed in more than 40 communities in a street in Shanghai, and has won the reputation of "good black box" in the local area. Through the AI ​​function smart box, intelligent perception and security reinforcement of people, vehicles, objects, events, etc. are realized at the edge, and application management and access to the upper platform are provided. Now a small black box solves the workload of the original four cameras, greatly reducing the deployment cost and shortening the original deployment cycle.