With the innovation and development of mobile payment technology, intelligent face recognition payment has come to the attention of the public after the increasingly mature and popularized QR code payment, especially the rapid development of face recognition technology in the past two years. The effective improvement of recognition accuracy and speed has provided a foundation for the application of face recognition in many fields.
The application of face recognition began to increase gradually. With the official launch of the Jinhua Bus Face Payment Project in January 2019, it became the first city in China to use the “face recognition” system, beginning a new era of face payment in the bus industry.
How about the practice and application of smart face recognition technology in the field of public transportation? In 2019, more than 20 cities in the bus industry have conducted face recognition tests or used face recognition systems. Face recognition has become popular in the bus industry. However, there are a few bus routes that can support face recognition in secondary cities. Most of them are in the trial operation stage. Therefore, it can be seen that the acceptance of face recognition by passengers needs to be improved. There are certain limitations.
Problems with facial recognition in public transportation
The main processes of face recognition technology include face image acquisition, face detection, preprocessing, face feature extraction, face matching/recognition, etc.
There are some problems in the practice and application of face recognition technology in the field of public transport, as follows:
1.High recognition accuracy
Bus payment belongs to the payment industry. The payment industry requires high recognition accuracy. Identification errors will cause a large number of complaints, which is a big challenge to user experience and bus operation. They can seriously affect users’ acceptance of face recognition.
2.Large user base
It has a large user base. Compared with the financial industry and the railway industry, the bus industry has a relatively larger user base. The bus system of a city covers almost all the people in the city. Secondary cities often have millions of users.
3.High traffic efficiency
Due to the large number of passengers, it is required to achieve fast passing in order to avoid congestion and traffic jams, especially during peak hours.
4.Different operating environment
Every city has buses. There are not only high buildings, viaducts, and tunnels, but also remote suburbs. Therefore, the operating environment is relatively complex. At the same time, different bus stop locations and times will also have different lights. There are different lighting environments such as backlight, low light, and strong light. There are also different 4G network environments in different areas and different locations in the same area. The online identification of places with weak network signals is greatly limited.
5.High construction costs
Urban public transportation is a social welfare. Most bus companies are facing financial constraints. At present, the face recognition terminals on the market are mostly face recognition terminals. There is no tapping and QR code payment, so buses need to install multiple terminals. On the one hand, it increases construction costs and causes a waste of social resources. On the other hand, multiple sets of equipment often correspond to multiple sets of management systems, which increases operation and management costs and also brings problems in data integration.
Facial Recognition Solutions of Public Transportation
Hangzhou Golong Technology provides a complete set of solutions in the field of public transportation based on face recognition technology. Advanced technology is adopted to ensure the accuracy and recognition speed of face recognition, which will help the development of face recognition technology in the field of public transportation.
1.Improve recognition accuracy
Hangzhou Golong Technology’s face recognition products use 3D face recognition technology. 3D face recognition technology has a higher adaptability to various situations.
It can achieve higher recognition accuracy, and effectively solve the problems of face recognition.
In terms of recognition efficiency, Golong Technology adopts the offline and online dual recognition method of headend equipment and backend servers, which can achieve millisecond offline recognition. Furthermore, it can solve the majority of user recognition during the peak hour. In other words, it can accelerate the recognition speed, increase the accuracy, reduce communication data, and decrease the recognition pressure of backend servers.
The bus operating environment is complex and varied, and the light environment is different. Hangzhou Golong Technology captures face data through a combination of optical and infrared cameras, which more effectively ensures the speed and accuracy of face recognition in strong light, weak light, and backlight. It can provide users with a better experience when traveling.
2.Solve large user base
First of all, with the development of computing capacity, existing computers have been able to identify million-scale databases and compare information from tens of millions of people. Secondly, the wide application of cloud services also provides a more effective deployment model and a more reasonable resource utilization rate. Thirdly, with the widespread promotion of microservices and other architectures, database and multi-node operations have become a reality to provide better solutions. Fourthly, multi-layer comparisons can also be made. Passengers can be divided into different libraries according to the travel rules, etc. Small databases are screened by large databases. Then, accurate recognition can be realized. The above technology can be used to solve a large user base.
3.Improve the efficiency of traffic flow
There are several solutions to the problem of traffic efficiency:
First of all, offline identification can be achieved to solve the efficiency problem caused by network latency. Offline identification is now mainly limited by the computing power of headend devices and the size of the local database of stored faces. Therefore, a set of algorithms is needed to screen the common passengers of each bus. travel information of each person can be analyzed through big data.
Secondly, with the development and gradual use of 5G technology, high-speed transmission networks will also become an excellent solution. Finally, capture cameras are deployed to pre-identify passengers in real-time in some places, such as BRT platforms, subway gates, etc. After successful identification, passenger information is synchronized to the charging terminal to achieve rapid identification.
4.Solve the problem of large differences in operating environments
In the actual operation of public transportation, the environment is complex. For different light problems, 3D recognition mentioned in the recognition accuracy can solve this problem. Whether it is low light, backlight, or strong light, faces can be well recognized. Secondly, face recognition can be combined with QR codes. Different payment methods correspond to the same account, which can not only improve the user experience but also solve the problems caused by the environment in face payment.
5.Save costs
In terms of cost savings, Golong uses a system integration method instead of replacing the existing payment methods. At present, IC cards, QR codes, face payment, and other payment methods are integrated into the full payment terminal, rather than a single face payment terminal. Hangzhou Golong Technology integrates the current popular QR code technology with face recognition technology on the basis of improving the efficiency and accuracy of face recognition. An App and an account can support multiple payment methods. Users only need to register in the App once. They can selectively open two application methods, which more effectively improves the user experience and speeds up users’ acceptance.
6.Wide application of face recognition
While the above problems can be solved, Hangzhou Golong Technology has also extended the face payment to various applications.
The first is the application of security, which can enable face payment devices to exert another layer of value.
Secondly, face recognition can also solve the losses caused by the borrowing of special groups' bus cards (elderly people, military personnel, students, etc.), and realize the online annual check of preferential cards at the same time.
Third, it also combines bus scheduling and driving safety to achieve multiple applications such as recording attendance through the drivers' faces, driving behavior analysis, and passenger flow statistics.
Fourth, during the epidemic, the application of facial recognition can provide better data statistics for the prevention and control of the epidemic, and reduce the workload of manual statistics.