Yuntian Lifei to create AI police officially launched "he" is improving traffic chaos

[Introduction]: This weekend, Yuntian Lifei created the AI ​​police officially on the line. The previous traffic inspection targets were all cars. This time it was for people. The red light was immediately captured and displayed in the front screen, making you the focus. Is this a good time to violate the rules? Such an AI policeman is said to be good at improving traffic chaos.

It is understood that on the first day of the AI ​​police's induction, a total of 58 special red light lamps and 67 non-motor vehicles entered the motor vehicle lane were captured. The so-called "special industry personnel" refers to the objects that can be identified by the AI ​​face capture system applied by the Shenzhen traffic police. So far, the system is only for the special industry groups that have been in the warehouse before, such as express delivery, take-out and other practitioners, and temporarily can not do accurate, real-time supervision and capture of the whole people, but this is still a symbolic step.

After the illegal traffic lights of special industry people are captured by the equipment, the Shenzhen traffic police automatically compares them with the face recognition system, and the relevant identity information is pushed to the law enforcement platform for manual review. After the confirmation is completed, the system will automatically extract the mobile phones registered by the relevant personnel in the traffic police for registration. Number and send a penalty notification message. If the parties have no objection, they can make their own contributions through the Shenzhen Traffic Police WeChat Service Number, APP Client, WeChat and Alipay “Urban Service”. If you have any objections, you can go to the relevant department for further investigation.

In the case of good results in the pilot work, the next step, Shenzhen traffic police "brush face" law enforcement will gradually expand to the dump truck, unlicensed fake car management, traffic violation window supervision, traffic accident "top package" and other aspects . The Shenzhen traffic police realized the man-car correspondence through brushing the law enforcement, and built an “invisible but ubiquitous” law enforcement network in Shenzhen to achieve precise traffic management for key people and key cars in Shenzhen.

The AI ​​police on the line in Shenzhen mainly focused on the red light behavior of “special industry personnel” such as take-out and express delivery. As early as April last year, Shenzhen piloted a set of face recognition red light capture device at the Xinzhou Lotus intersection. The device automatically captures the pedestrian red light behavior with the camera and plays it on the big screen on the safe island.

It is understood that the system is built by Yuntian Lifei. By detecting pedestrian red light behavior, the face is extracted and identified. Through real-time search comparison, it is found whether the same person has red light multiple times, and the class is connected through data to implement the identity of the illegal person. Display the corresponding information on the large screen of the offender.

Although this system has caused controversy to protect pedestrian privacy, it is undeniable that it is effective in stopping pedestrians from red light and ensuring traffic safety. The data shows that during the half-year period of the system, a total of 13,930 illegal pedestrians were arrested; and the pedestrian red light violations dropped from about 150 per hour to 8 per hour, and the traffic order was effectively improved.

In addition to effectively solving the "people" problem in public transportation, AI technology also has a set of methods for the "car" problem.

Interlocking has always disturbed road traffic. Although most car owners will receive a ticket from the traffic police department after parking, the problem is that after-the-fact punishment can't immediately restore road traffic, and many drivers are unconscious. Traffic violations, that is, they did not know that the location could not stop when they parked.

In response to such a situation, the Wuhan Railway Station has launched a set of “autonomous traffic violation reminder system” created by Yushi. It is understood that the system is directly connected to the LED display by a front-end camera equipped with an AI chip. After capturing the illegal driving and parking behavior, the camera can directly process the front end and feed the result back to the LED in the second time. On the screen, intervene in advance and suspend violations. The system has been effective since its application to the Wuhan Railway Station. The traffic violations have been reduced from about 100 times a month to about 20 times a month. At the same time, the traffic volume during peak hours has increased from 1,800 per hour. To 2,200 vehicles.

What else can AI do?

Takeaways, couriers, pedestrians, red lights and vehicles are parked. It is just the tip of the iceberg in public transportation. The pain points of public transportation are traffic congestion, traffic safety, traffic violations, etc., and with the development of the times, new Traffic problems are also constantly emerging. For example, the emergence of shared bicycles brings convenience to the masses. At the same time, the problems of random parking and irregular use have brought many negative effects.

But these traffic problems that have long plagued cities have been gradually being solved by AI.

At present, traffic congestion in many cities in China is serious. Many traffic lights at intersections are not optimal. The AI ​​is based on the deep sensing of vehicle accurate sensing detection, which can accurately perceive the number, flow and density of vehicles in all directions of traffic intersections. Provide an accurate basis for the optimal timing of traffic intersections. If such vehicle detection technology is used at various intersections, traffic congestion will be greatly alleviated.

AI can also detect traffic accidents and incidents. Based on continuous video, it can analyze the behavior of vehicles, detect behaviors such as vehicle parking and retrograde, and find traffic accidents and traffic jams. This technology has been widely used on highways before, but Due to the existence of more false positives, there is still a gap between the actual effect and the real demand. In the future, with the help of deep learning technology, a truly accurate traffic incident detection system can be realized, and the traffic operation department can be provided to provide accurate and timely alarm information.

Of course, in the future, when AI and related laws are mature enough, car-assisted driving and even driverless driving can be realized. At that time, vehicle information can be connected with urban traffic data in real time, and human beings will usher in an unprecedented travel experience. The transportation system will also It has been reshaped, but the arrival of this day is still very far, so security companies should pay more attention to the immediate problems.

The traffic problems mentioned above are all “in front of” problems that need to be solved urgently. However, the above mentioned only partial public transportation, or solutions to some problems in public transportation, but to develop smart cities, it is necessary to have specific traffic. The solution to the problem, and the urban traffic brain is such a total solution.

But the overall solution at the city level is not easy to achieve. First of all, security companies should have an in-depth understanding of each security AI product, choose the right product according to different scenarios; also be able to optimize the algorithm according to different environments; after the system is built, it must be able to continue to support the subsequent operation services... ... It can be said that it is difficult to complete these heavy tasks by the strength of one company alone.

It has become an industry trend to provide a total solution from the provision of a single product. Therefore, security companies must strengthen cooperation with their peers and build their own ecosystem. For example, Huawei has been able to enter security and has achieved good results. The reason is that Huawei has built its own ecosystem by cooperating with various AI companies and security companies.

Although traditional security companies do not have the influence and appeal of Huawei, they can still find long-term partners in various ways through years of accumulation in the security industry. First, strategic alliances, similar types of enterprises form a strategic alliance to enhance competitiveness; second, upstream and downstream alliances; third, partner alliances, such as alliances between product manufacturers and distributors, agents, and engineering companies. relationship.

Large and powerful companies can even consider the acquisition of resources in the market, technology, capital, etc. through mergers and acquisitions. After all, under the era of “great security” becoming a trend, it is difficult to achieve a single climate, only through the establishment of partnerships, mutual use The advantages of the resource network can form new advantages and continuously develop new markets.