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

Source: brain polar body

This article is about 5112 words. It is recommended to read it for 15 minutes.

This paper introduces that Huawei machine vision has targeted key industrial changes under the strong driving force of the "new infrastructure" cycle.

Among the many industrial upgrading values brought by artificial intelligence, machine vision will undoubtedly contribute the largest space.

Whether in industry, agriculture, service industry, and financial industry, intelligent solutions based on visual interaction account for more than 80% of the intelligent system. And vision is often the bearing way of knowledge graph, semantic segmentation, machine learning and other technologies.

With the development of AI industrialization, a consensus has been reached on the importance of machine vision inside and outside the industry. The most important problem in this industrial cycle is how machine vision, as a systematic general technical capability, can be combined with vertical industries with obvious differences and relatively complex industrial laws. Let machine vision move from the security industry to the depths of more industrial scenes, and promote machine vision from the external auxiliary force of the industry to the core of the real production system.

In March this year, we reported on the strategic upgrade of Huawei HoloSens, which established the strategic goal of using machine vision as the entrance to the industry to promote the intelligence of the industry. This upgrade attracted close attention in many fields such as artificial intelligence, security and industrial intelligence at that time. There is no doubt that the focus of Huawei machine vision is how to solve the fusion problem of machine vision technology from generalization and simplification to scene and industry. It can be said that Huawei machine vision, which is at the forefront of the industry, must take the lead to complete this breakthrough from 0 to 1.

Just now, Huawei Machine Vision released its annual answer. On May 25th, Huawei HoloSens next-generation camera and 2020 new product conference with the theme of "Vision creation | opening the eyes of the Smart World" was held in Zhejiang Province. Six new products were launched, including the new next-generation "Siwu" ecological camera and ultra-high-density intelligent video storage platform, AI ultra-low-light bayonet camera, dual-spectral thermal imaging camera, AI dual-mode ball camera and 5G camera. At the same time, ten scenario solutions have been released to accurately cover key scenarios.

Of course, if you extract these products or solutions separately, you can also interpret the very obvious value of innovation. If we want to peek into the deep fusion of the machine vision industry, we may still need to see through the trend of industrial logic and technology chain behind this release. Huawei's new products in the field of machine vision are not only commercial innovations, but also reveal Huawei's industry thinking and industry pain point judgment at this critical point.

Duan Aiguo, president of Huawei machine vision, said: "Huawei's machine vision dream is to 'perceive everything with smart eyes and light up the intelligent world'. It will continue to be introduced into cool techs in various fields of Huawei. To provide customers with competitive 'holographic perception' and 'edge cloud' collaborative solutions, to become a machine vision industry leader. The next generation of camera opportunities launched this time will bring the industry into a new era, while other new products and new scenes will also continue to broaden the depth and width of AI and detonate the machine vision industry. "

Following this judgment, our next question is the epoch-making value of the new products represented by Huawei's next generation of cameras. Today, with the full display of industrial value by artificial intelligence and the opening of the 5G era, the industrial symbiosis of industry scene to machine vision is increasing rapidly. In order to achieve the further coupling of machine vision and industrial demand, it is necessary to complete a series of redefinitions from insight, to product technological innovation, to solutions and industrial ecology, and to achieve a new "perfect relationship" between thousands of industries and machine vision technology.

Huawei calls this process, opening the eyes of the smart world.

The expected "perfect relationship": machine Vision as a Digital grip

Just as vision is the first sense of human beings, guiding the most human actions and behaviors, machine vision is also the first sense that AI technology system brings to digital devices. From recognition, understanding to feedback, the intelligent interaction with the largest number, the deepest degree and the most obvious value is basically realized through machine vision. When this ability enters the industry, it can obviously bring great productivity activation value.

Whether it is active identification in the field of security, intelligent quality inspection in the industrial field, traffic flow and vehicle identification in the field of traffic, the landing of the AI industry that we first recognize is basically related to vision. However, with the continuous deep blending between the technical capability supply of machine vision and the demand of the industry, a large number of new industrial bottlenecks have emerged that need to be broken.

Zhou Yuefeng, president of Huawei's data storage and machine vision product line, said: data infrastructure will play a key supporting role in the development of the digital economy, and machine vision is a very important part of the data infrastructure. 65% of the industry's digital information comes from video, and the deep combination of machine vision with life, entertainment, communications, production and security will be a huge market worth hundreds of billions of dollars. In the future, Huawei will continue to increase its R & D and investment in the field of machine vision, making Huawei HoloSens the number one brand of machine vision and pushing the industry to a new height.

In the process of machine vision and industry scene from shallow integration to deep coupling, from auxiliary capacity to production core competence, we can really see many factors that industry and technology are in urgent need of upgrading. In the process of moving towards the giant market, machine vision is not a golden key, but an industrial starting point that needs to be honed, promoted and rebuilt. For example, we can see that in the process from machine vision to the core of the industry, there are several outstanding problems that need to be overcome:

1. After the machine vision ability covers the basic scene quickly, and provides the basic capabilities such as active identification and active marking for various industries. Its value is not enough to match the deeper industrial needs. In the deep industry demand, we see that machine vision capabilities need to be deeply integrated with more technological contexts, such as sensor technology, thermal imaging technology, photovoltaic technology, etc., in order to adapt to more complex, diversified and challenging industry scenarios, and let visual intelligence move towards holographic intelligence.

2. Today, the software and hardware environment of machine vision is in the stage of rapid development. To integrate the needs of the industry with machine vision, it is also necessary to update the system of products and solutions in the trend of technological development. For example, the arrival of 5G provides large bandwidth and low latency network infrastructure iteration, which will bring a new imagination to the industry machine vision. 5G machine vision equipment has become a new market demand space.

3, machine vision to the industry, must meet the specific vertical demand and customized scene, any manufacturer can not meet the huge demand. Therefore, it is necessary to build the same algorithm model circulation space in the mobile application market, and meet the needs of vertical scenarios with the help of the creativity of ISV and AI developers in various industries.

The integration of software and hardware technology, the follow-up of new technology trends and the construction of multiple open ecology have become the three hard elements for machine vision to move towards the industry. In the process of entering the industry, Huawei machine vision has successively come up with benchmarking products and solution innovations such as software-defined cameras, intelligent 1tuo N, HoloSens Mall and so on. While moving forward and stepping into the deep waters of machine vision in the industry, Huawei's strategic idea of machine vision is to continue to rely on Huawei's technology system, introduce cool techs in various fields of Huawei, take Huawei cloud&AI BG as the support, broaden the width and depth of machine vision applications, continue to introduce technology and product innovation, and promote the deep coupling of machine vision and industry demand.

And the key role newly unlocked on this road is the new camera product released this time.

Unlocking and opening: the era of "Super Fusion" of cameras

If we say that the machine vision algorithm is integrated into the hardware to form an artificial intelligence camera, it is the first fusion in the proposition of machine vision industrialization. Then, when the industry continues to move forward, and machine vision capabilities need to be deeply integrated into industry scenarios, industry needs, and industry open platforms, machine vision capabilities, camera products, and new capabilities and technologies needed by the industry need to be integrated again, which can be called the "super-fusion" period of AI cameras.

Signs of super integration can already be clearly identified in Huawei Machine Vision's next-generation camera products, as well as innovative products such as 5G cameras. Let the machine vision technology further integrate with the industry demand, special operation scene, complex installation and operation and maintenance environment, and form a new industry wind direction.

For example, the most eye-catching next-generation "Siwu" ecological camera can be said to have achieved conceptual machine vision product innovation, breaking the time limit in the application of machine vision products. The reason why this red dot award-winning work in 2020 is called "four without design". Because it realizes flexible deployment of "no net" and "no electricity" through the VideoX sub-machine self-organizing network + optical storage integration module; SuperColor+, realizes day and night viewing in the "no light" environment from full-color pictures to full-color video; and the design of "no on-site operation and maintenance" uses intelligent operation and maintenance technology to reduce operation and maintenance costs and improve industrial efficiency. In addition, the "ecological warehouse" design of the next generation camera can support all types of sensor interfaces, can flexibly achieve ecological interconnection, and integrate into real holographic perception with the help of a variety of machine vision algorithms and sensor systems.

This product, which exceeds the imagination of most users, obviously redefines the deployment environment, capability experience, and operation and maintenance costs of machine vision products. More complex industry scenes, especially outdoor and remote scenes, can be covered by machine vision, realizing the real "next generation".

For example, the start-up of 5G commercialization and enterprise-level services are the biggest industrial dividends in the "new infrastructure" cycle. How to integrate the 5G dividend into the machine vision system is indeed a difficult problem. The 5G camera launched by Huawei Machine Vision fills a key gap in this field. Based on Huawei's leading end-to-end 5G technology and a number of innovative technologies jointly developed with Huawei 5G Wireless Application scene Lab (XLab), 5G cameras can be flexibly covered in the wireless network environment. The intelligent flow control technology and intelligent coding technology of end-to-end network cooperation increase the number of camera access by 2-3 times. FillP patent technology ensures that when the packet loss rate is 10%, the bandwidth utilization is still more than 95%, and the video is more smooth. Based on the 5G camera, the industry scene can be the first to apply the 5G dividend in the end-to-side scene to achieve efficient and stable 5G + industry + machine vision.

In addition, Huawei machine vision also integrates a variety of technologies into products such as cameras to form an end-to-side infrastructure that can meet the needs of various industries. For example, the ultra-high-density intelligent video storage platform HoloSens IVS3800S120 brings extreme space, with a disk utilization of 92%. It can provide industrial support for massive data storage and edge view; AI ultra-low-light bayonet camera, with the help of Huawei's powerful AI chip ability, through computing power to exchange images, intelligent and accurate lighting, reduce night light pollution, and realize the usual deployment and application of machine vision equipment in low-light environment. Dual-spectral thermal imaging camera, the integration of AI visible light and thermal imaging technology, can obtain ultra-long-distance, diversified machine vision capabilities in the natural environment; AI dual-mode spherical camera, using wide-angle and telephoto dual-lens integrated design, to achieve binocular coordination, bringing unique value in a wide range of industry scenes.

The six new products launched by Huawei Machine Vision can be said to have opened a new cycle of deeply integrating machine vision technology with the needs of the industry. The integration of new technological innovations such as low light level, thermal imaging, 5G, ecological warehouse and machine vision brings out a new era of "super fusion" products.

On this basis, Huawei machine vision will promote the further coupling of technology and industry, and stimulate the ecological and scene market space in the process of machine vision industrialization.

Weaving the Smart World: the scene Plan embraces thousands of Banks and all Industries

In the process of AI surging into the industry world, a core logic point is that the industry, as the recipient, must be judged by its own industrial foundation, industrial demand and industrial cycle. All new technologies and products must be combined with industry wisdom in order to truly coruscate the value of the industry.

In this case, the development of ecological and scene-based solutions has become the focus of communication between machine vision technology and the needs of the industry. In this release, based on six new products, Huawei further launched ten scenario solutions that meet the needs of specific industries, which will continue to push forward the dominoes of industrial machine vision.

The ten scenario solutions include: holographic community, holographic intersection, high-speed networking, smart water conservancy, smart network, general education network, smart electricity, smart oilfields, smart manufacturing, smart park, its content covers food, housing, work, education, finance, health care, construction, agriculture, forestry, animal husbandry, sideline and fishing and other aspects of life. It shows the evolution path of machine vision technology from security field to industry production core, from basic visual intelligence to holographic intelligence, and fully interprets the positioning of Huawei machine vision "industry digital grasp".

Deep realization of technology integration, giving full play to Huawei's technological advantages of end-to-end cloud integration, and building solutions based on industry insight can be regarded as the main features of Huawei's machine vision scene solution. For example, Huawei Smart Water solution fully releases the technology integration dividend of machine vision technology + AI infrastructure + IoT equipment + big data. In order to build a highly intelligent water affairs information guarantee system, achieve "more thorough data perception, more intelligent decision-making and command, and more intelligent facilities and services", and provide intelligent support and innovation drive for the sustainable and all-round development of intelligent water conservancy. In this solution, it includes the application of collecting water level, rainfall data and video intelligent identification of water level and urban waterlogging forecast, and benefits machine vision technology to automatically identify river floating objects and throwing objects, water ruler identification and illegal discharge and other river illegal events, and can accurately achieve accurate detection of personnel intrusion. With the application of algorithm mall, intelligent water solutions can be further upgraded and evolved to achieve industry multiple scenarios, specific requirements, and machine vision as the main body of a series of end-edge cloud capabilities.

The holographic intersection solution is to calculate the ITS800 through the traffic micro-edge, horizontally through the camera, radar, intersection signal control data, vertically support traffic business innovation, combined with high-precision map digital twins to generate the digital God perspective of the intersection.

It can provide the traffic department with accurate metadata such as license plate, speed, position, attitude, attribute, split lane flow, traffic incident detection and so on, combined with high precision map to provide traffic control users with digital holographic perspective. 40 + traffic business data such as intersection speed, headway distance, queue length, parking times and so on can be synthesized. Compared with the traffic control data collected by traditional geomagnetism, coil and flow camera, the traffic control data collected by traditional geomagnetism, coil and flow camera are more accurate and stable, and the Internet floating car data sampled in minute level are more precise and real-time. Thus, the application effect of traffic signal control optimization, traffic organization optimization, traffic hidden danger management, accident handling, command service optimization and auxiliary vehicle-road coordination V2I is better. In this way, the in-depth machine vision capability pioneered by the industry, based on the integration of systematic software and hardware systems and real scene requirements, solves the problem of traffic intersection intelligent systems being separated from each other and chimney-like data. a large number of requirements are integrated, unified and upgradeable, and the holographic intelligence of the most complex traffic scene is realized.

With the help of the development of ecological and scene-based solutions, Huawei machine vision continues to open the boundary between the needs of the industry and enter a new era in which machine vision empowers the core of productivity. This is Huawei's proposal that machine vision will lead the next era. Driven by the "new infrastructure" cycle, Huawei machine vision has targeted key industrial changes. Build industrial ecology on the basis of product and technological innovation. This is a systematic change that has never occurred in the global machine vision industry, and is constantly affecting more industrial chain enterprises, developers and industry users, ultimately benefiting every family and individual.

Machine vision is the eye of the intelligent world and the entrance to the perception of all things. A perfect relationship between technological innovation and industrial demand is emerging in the innovation of cameras.

In this new story, I believe you can also find the opportunity to explore the future world with machine vision as your eyes.

-end-

Https://www.toutiao.com/i6833597466177176067/

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.