Shulou(Shulou.com)06/02 Report--

Big water and big fish, using this sentence to describe the current intelligent security industry may not be more appropriate.

With the continuous improvement of urban fine governance, urban security is becoming more and more important. From the beginning of public safety, road traffic safety and financial, campus, hospital and other key industries security, gradually developed to the park, community, family and other regional, personal security areas, is forming a complete security scene system.

The maturity and application of AI technology brings new opportunities for the upgrading of the security industry. Similarly, the security industry is also a mature field of AI technology, especially machine vision and knowledge graph, which is the earliest mature landing and large-scale development.

According to the 2019 China Security Industry Survey report, the total output value of the national security industry in 2019 was 826 billion yuan, with an annual growth rate of 15%. Intelligent security occupies a more and more important share. It is estimated that the intelligent security market will exceed the market value of 100 billion yuan in 2020, of which the market size of AI+ security software and hardware will reach 45.3 billion yuan.

The industrial development of intelligent security can be divided into six stages: algorithm, computing power, data, scheme, implementation, operation and maintenance.

The first three stages can be called the first half of "technical support". No matter they are head security manufacturers, integrators, Internet, AI enterprises, and cloud service providers, they are all ready to complete the technology layout. The latter three stages can be called the second half of "scene application", which involves putting products and technologies into the application of specific landing scenarios, which need to face more fierce competition and more realistic challenges, and the war curtain is about to begin.

The security industry itself is characterized by a huge market but serious fragmentation of the scene, so it can not only cultivate 10 billion revenue security giants such as Haikangwei and Dahua, but also support tens of thousands of small and medium-sized security enterprises.

Among the contestants who have poured into the intelligent security track in recent years, we can also see traditional security tycoons like Hai Daewoo and Internet giants such as BAT, and we can also see the latest high-profile launch of a new intelligent security brand by Huawei, Huawei HoloSens (good Hope), AI startups such as Shantang, Yitu, and absent vision, as well as hundreds of companies ranging from underlying hardware to terminal solutions.

It is conceivable that the wrestling intelligent security track will be a very fierce competition. We must see the desperate struggle for hundreds of millions of projects of To G and To B, and we will also see the price war that continues the traditional security market, and we will also see frequent infighting among the head players.

However, after these "deja vu" normal narratives, we hope to delve into the intelligent security industry, because what new market opportunities can be brought about by technological upgrading and scene sinking? In this gluttonous feast in the second half of intelligent security, in addition to the qualifying of those who compete for the main guest seat, can those eager small and medium-sized players also eat in it?

Climb the "three mountains" of AI to find the technical development point

With the development of AI, big data and cloud computing technology, security is moving from traditional video surveillance to intelligent security. Among them, AI technology, especially machine vision technology, is widely used in the field of video surveillance, upgrading from "visible" to "understandable" and then to "able to make decisions".

When AI technology is applied in the field of security, it is also breaking through the "three mountains" of computing power, algorithm and big data, helping the security industry to transition from the stage of "weak artificial intelligence" to the stage of real strong intelligence.

The improvement of computing power depends on the improvement of the processing power of AI chips. However, the low computing chip built into the traditional camera makes the performance of perception and recognition not ideal.

The upgrade of the algorithm needs to be supported by the decoupling of software and hardware and the open algorithm ecology. The traditional camera replacement algorithm needs to change the camera at the same time, so once the algorithm is changed, it is equivalent to the reconstruction of the whole network. If the AI camera supports software and hardware decoupling and the algorithm is loaded on demand, it will quickly adapt to business changes and greatly reduce the cost of repeated construction. The progress of deep learning algorithm and open algorithm training platform can provide rich algorithms for end-to-side multi-scene recognition.

Data integration needs to realize the standardization and interconnection of multi-dimensional data. The standardization of multidimensional data comes from the structural analysis of video data and the fitting of multidimensional data to improve the recognition rate of AI. The interconnection and interworking of data requires the data collaboration of the end-side cloud.

In the past 2019, the "three mountains" that hinder the upgrading of the intelligent security industry are being broken one by one.

At the computing level, the computing power of mainstream AI cameras is changing from a low computing power of 0.66T to a high computing power of 4T, 8T or even 16T. This will greatly improve the front and rear intelligent processing ability, and greatly improve the efficiency of face, vehicle capture, intelligent analysis and so on.

At the algorithm level, in order to deal with complex security scenarios, different algorithms are being developed, including multi-feature extraction and recognition, cross-shot tracking, end-cloud collaborative intelligent analysis and so on. The iterative development of the algorithm under the subdivision scenario provides a space for many small and medium-sized AI enterprises to develop.

At the data level, the cloud and cloud-side collaboration of data has become the mainstream trend. Relying on the cloud platform and big data, in addition to the software and hardware decoupling of the terminal, hardware, platform, data, and algorithm applications will be decoupled at different levels, so that a set of intelligent security system can be supported by multiple professional suppliers. It provides development space for industrial diversification.

The formal commercial use of 5G technology in 2019 and the popularity in the next few years will also become a strong driving factor for the development of intelligent security. The large bandwidth capability of 5G satisfies the HD transmission of video, the low delay capability satisfies the real-time manipulation of monitoring, and the massive connection provides the possibility of multi-dimensional perception for intelligent security.

The combination of 5G and AI will greatly improve the efficiency of intelligent security in perception, analysis and decision-making. at the same time, it can also greatly enrich the application scenarios of the security industry, expanding intelligent security from the existing public safety, transportation and other special industries to richer B-end and C-end industry scenarios.

In this process, in addition to the direct benefits of operators, cloud service vendors, chip and hardware manufacturers with their own natural advantages, new AI startups, integrators and channel vendors can also establish their own exclusive advantages in technology, ecology and channels to enter the second half of the scene application track.

Application scene sinking to find the niche of industry

From the perspective of the industries and business scenarios involved in intelligent security, non-civilian areas such as public transport systems and public safety systems are undoubtedly the most important and largest market areas. For example, public security and transportation projects accounted for nearly 40% of the total market share in 2018.

These AI security fields are characterized by large scale, considerable profits and high technical requirements, but at the same time, they have a long construction cycle, many links and a long payback cycle, which has become a high barrier field that traditional security established manufacturers, integrators and IT technology giants can "afford".

The second largest market area is the B-end security market involving smart finance, Wisdom Society (Park) District, Wisdom Campus and other niche areas. In almost every subdivision, there have been a number of traditional security enterprises and AI manufacturers actively entering the bureau, and even real estate developers like Vanke have been building their own community security platforms to create intelligent communities with their own advantages; and Haekangwei, like the traditional public security leaders, is also involved in access control, property management, building intercom, license plate recognition and other businesses through active acquisitions.

The third largest segment is the pure C-end market. Due to the general technical difficulty, it has become a sub-area that AI security startups and any hardware enterprises are trying to set foot in. In the initial stage of mixed competition, entry enterprises compete more with brands and performance-to-price ratio, trading sales for the market, but how to survive and live for a long time has always become a problem for these enterprises.

However, the "fragmentation" and "decentralization" characteristics of the entire security market determine that it is difficult to form a stable pattern of one-take-all or several dominant companies in the field of intelligent security. With the continuous extension and sinking of intelligent security industry and scene, the commercial landing of intelligent security is full of new opportunities.

In the field of public safety, the "Xueming Project" proposed in 2016 has brought a market size of tens of billions of dollars. This mass public security prevention and control project to promote safe rural construction has directly sunk the security market to the county, township and village level, becoming a "cake" for integrators with rich products and strong engineering capacity. As for AI startups that cannot compete directly, they can get a share of the benefits by playing the role of a technology or server provider and taking on some of the engineering construction.

In the B-end industry scene, intelligent security is going deep into the subdivision scene that many industries can not touch. For example, at construction sites, the original camera can only play the responsibility of anti-theft monitoring, but now the construction progress, personnel operation standards and safety protection of the project can be effectively monitored by configuring cameras on building tower cranes and vehicle mechanical equipment. Similarly, in port, mining, fire protection, environmental protection, logistics manufacturing and other B-end industries, the layout of AI cameras can greatly reduce the task of manual on-site operation and inspection. And by the security system for real-time monitoring and analysis and danger report. Both AI algorithm vendors and application solution integrators can build their professional capabilities in a large number of new scenarios.

In some industries with complex projects and low profits, such as community intelligent security, it is changing from an awkward situation in which large manufacturers are unwilling to do it and small manufacturers cannot do to a situation in which multi-parties actively invest. Some large real estate developers are trying to carry out the overall planning of community AI security from the beginning of community planning, while Ali and Tencent are also trying to connect with community services through their own software portals and create an Internet AI community through background AI capabilities. And many AI enterprises are also through a complete set of intelligent security community solutions, which combines face recognition, video structure, big data and other technologies with community scene applications to achieve more subdivided scene functions.

Sink to avoid moving from the traffic and public safety packed with giants on the track to more dispersed but sizeable market segments, so as to gain living space; dig deeper, occupy professional technical positions or provide professional landing solutions in these industry segments, and become partners and professional technology, product and service providers of industry giants. These two directions may become a new choice for small and medium-sized enterprises to find another way in the intelligent security market.

Rich places of water and plants are found in different niches.

"in the next ten years, there will be a lot of water and a big fish to find the right position."

This is actually the point of view made by a practitioner of the security industry at the 2019 China Security Industry Empowerment Conference last year. The main logic of this view is: how to redesign their own products and systems from the perspective of user scenarios and find their own new positioning under the emerging technologies such as AI, 5G, IoT and other emerging technologies that enable the security industry?

For the development logic of the intelligent security industry, it is also with technical empowerment, customer needs, and scenario applications, the key is how to accurately translate customer requirements into landing scenario applications, and realize it through technical tools.

This logic is the same for both head security enterprises and many small and medium-sized security enterprises.

The biggest difference is that the head enterprise can independently complete the full process deployment from the underlying basic platform, network, hardware and application algorithm layer, so it can get intelligent security ecology for those highly rewarding projects and customers.

On the other hand, small and medium-sized security enterprises can focus on providing finer-grained solutions and operation and maintenance services in the subdivided areas of the industry, so that they can get orders from customers or institutions in those small and medium-sized industries, such as just a community, shopping mall or mine.

In addition, small and medium-sized intelligent security enterprises can also join the intelligent security ecosystem of giant enterprises, become industrial chain partners under platform convergence, and keep up with the pace of the "boss." give full play to their own unique advantages in algorithms, hardware, engineering, operation and maintenance, and seek better development opportunities.

Of course, for a large number of small and medium-sized intelligent security enterprises, compared with the experienced and solid security companies and cross-border giants, small and medium-sized enterprises need to face more challenges.

From technology to scene landing, these intelligent security enterprises have to deal with the following problems. For example, do a good job in the knowledge translation of customer requirements to avoid project failure caused by differences between the implementation of technical solutions and customer requirements; cost control of technical solutions, understanding the price affordability of customers, and avoiding smooth communication of solutions, while the cost cannot be covered during cooperation; to do a good job in security and credible compliance requirements, the most basic principle of the security system is data security, once customer data is lost, the loss will be irreparable.

These are the main considerations for many security customers when they choose intelligent security solutions and service providers, and they are also the key for many small and medium-sized enterprises to win customer trust.

But in any case, the second half of the intelligent security industry has come, whether in the traditional massive public security scene, or the new industry segmentation scene opened up by new technology, will provide an incomparably rich niche. To let players of different sizes and different positions win their own living places.

Of course, it is also important to remember that the second half is also the beginning of the knockout stage. Once the heroes grow up and the space becomes smaller, it may be difficult for those contestants who fail to occupy seats to avoid being squeezed, eliminated or even swallowed.

However, the competition for intelligent security has only just begun. In the wave of industrial Internet and urban intelligence driven by new technological changes such as AI, 5G and IoT, not only all the scenes of traditional security are worth doing again, but also a new wave of intelligent security scenarios will be born in To G and To B market, contributing trillions of market dividends.

The start of 2020 may bring many industries to a temporary standstill, but for the security industry, it brings a positive inspiration. Modern cities still have many problems in dealing with major public crises, and intelligent security will have a bright future.

It can be predicted that during this epidemic, those cross-border giants and AI enterprises that move quickly and can provide effective solutions will become potential players in the second half of intelligent security.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.