In the past two years, in addition to 5g, the hottest technical term is the Internet of things. It seems that any company that does not mention the Internet of things is embarrassed to say that it is a company in the science and technology circle. Internet leaders take the Internet of things as the next outlet and invest heavily in the layout.
In fact, the concept of Internet of things has been put forward for 20 years, which is not “new”. The concept of interconnection of all things has been repeated, but it has not really started to be applied until these two years.
Previously, due to technical restrictions and various industry barriers, there were many difficulties in landing. Now, with the breakthrough of lpwa technology, especially the commercial landing application of 5g in 2019, it provides a good technical foundation for the landing of the Internet of things. The chips and algorithms that have developed rapidly taking the opportunity of neural network artificial intelligence have also become the booster for the development of the Internet of things, and the interconnection of all things is no longer a concept, It will become a practical application.
The “thing” in the interconnection of all things is a huge and complex system. Generally, the whole system is divided into application layer, platform layer, network layer and perception layer. There are countless sensors, voice receiving equipment, video acquisition equipment and so on in the perception layer alone.
Among them, video, as the eye of the Internet of things, is not only the most important part of the perception layer, but also the most complex part to obtain the most information in the whole system.
How to make good use of the information collected by video has always been a research topic for video practitioners, and the loading of artificial intelligence is only the beginning. Therefore, how to play the role of eyes in the Internet of things is an important key to whether the Internet of things can “see clearly”.
In fact, as an important part of traditional security (performance may account for 30%), in addition to its own video elements, video capture equipment has long been “Internet of things”, such as RFID camera, 4G camera, 2.4g/5.8g wireless camera, UWB camera, Lora camera, floor display camera, etc.
It can even be said that the security platform integrating various sensor data such as fire fighting, various noise and environmental meteorological elements is a small prototype system of the Internet of things, but the scope is limited and more applications have not been explored.
Now, under the overall trend of the Internet of things, how can security front-end equipment technology better cut into the application of the Internet of things? It can be seen from the following evolution directions:
A front-end device is not only a video acquisition point, but also the smallest edge unit in the Internet of things system according to the scene requirements, which not only saves various distribution resources, but also saves networking resources.
Of course, the fusion is also divided into many aspects. On the one hand, it is the fusion of the camera itself, on the other hand, it is the integration of a variety of sensor positioning, and others, such as the integration of optical port 485 interface.
The first two aspects are undoubtedly the most important development directions, and the integration of the camera itself has also begun to take shape. For example, the brilliant fusion of invisible light and visible light in thermal imaging in this epidemic, such as the binocular thermal imaging visible light integrated equipment and the multi camera products of the watcher series released by Dahua, are similar to multi camera integration applications. Such applications are also in line with the government’s demand to build a “beautiful” city, replace the “sparrow pole” image and reduce the number of urban poles.
The integration of video and sensors has begun since the era of simulation. At the beginning, smoke detector cameras, and later various wireless cameras, 4G, GPS, RFID, etc. have integrated cameras. Later, with the concretization of various scene requirements, users’ demands to reduce the amount of equipment control, and the camera is no longer just a camera, It is a collection of various perception layer devices.
More intelligent decision-making
In recent years, it has been a process from the pilot of CNN intelligent application of front-end acquisition equipment to its gradual implementation. Nowadays, intelligent applications in all walks of life have played an important role, such as mask detection and alarm in the epidemic, video analysis and monitoring of alarm events, and intervention and prevention in the event, saving personnel on duty and avoiding personnel contact, This is the primary decision-making role of front-end video capture equipment.
In fact, most of the existing intelligent front-end devices only stay in the primary intelligent stage of forensics, such as face capture cameras, which only screen the best face pictures in each frame and upload them to the back-end for further analysis by the back-end, which not only occupies a lot of network resources, but also increases the cost of the back-end processor and wastes the processing performance of the front-end chip.
Therefore, gradually putting the decision-making algorithm in front and making full use of the smallest computing unit at the edge can better adapt to the limited resources in the era of Internet of things.
On the other hand, according to the statistics of relevant institutions, 75% of smart cameras in China use Hisilicon chips. The computing power of the terminal chip shengteng 310 newly released by Hisilicon in 2019 has reached 16t, and the computing power of the chip has far exceeded the needs of the existing front-end intelligence. The forward movement of decision-making algorithms will make full use of the computing power of the chip and give full play to the maximum role of the smallest unit of front-end video acquisition equipment.
Smart home has always been an important landing scene of the Internet of things, but its development is very slow. A big problem is that the agreements between various household products are not feasible. Each household appliance manufacturer fights its own way. It is difficult to get through the agreements of various manufacturers for various complex interests.
At present, the access volume of tmall elf devices at the Internet of things entrance has reached 100 million. Looking around the world, there are also Internet of things entrance devices such as Amazon echo and Google home. With the participation of top Internet companies of these non-traditional electronic hardware devices, the interconnection of electronic products is just around the corner, and video, as an important sensing device in the Internet of things, especially household cameras, It must be a trend to follow a unified interconnection protocol in the future.
The commercial landing of 5g provides more possibilities for 10000 “things” to access the huge Internet of things.
Of course, in addition to the above, more front-end acquisition technology routes, such as all-weather adaptability and all intelligent element extraction, are also very important, which will not be carried out one by one here.
The development of the Internet of things not only provides more convenience for the evolution of security front-end equipment, but also makes front-end equipment subject to more tests, such as the following two aspects:
First, data security. After all things are interconnected, the most important thing is to ensure the security of data. The 360 small water drop incident is a lingering fear. Therefore, for security video monitoring, the state has also issued 35114 security standard, but the current implementation level is limited. How to ensure the security of front-end equipment under the Internet of things will be a big test.
Second, privacy protection is closely related to data security. The collected video data needs to consider security in order to better protect privacy. However, before collection, there should be corresponding laws and regulations to prevent personal privacy from being collected and abused unknowingly. Whether this can be directly avoided from the technical aspect is a higher requirement for video technology.
In short, the era of Internet of all things has come. As a video collection product of traditional security, how to use this wave of air outlet to make collection products to a higher level is a test question for every old security person.
Responsible editor: GT