Since launching Amazon Go in Amazon last year, no one has become a popular topic and the object of capital favor. In fact, there are also many domestic companies that are researching unmanned sales technology, such as F5 Future Store, Boxer, quiXmart and so on. Now Alibaba has also come to join in the fun. According to Lei Feng network, Ali will launch Ali no supermarket “Amoy Coffee†on the second Taobao Creativity Festival in early July.
How does "Amoy Coffee" work?
According to Lei Feng Network, “Amoy Coffee†will be a 200-square-meter offline store that houses shopping and dining, and can accommodate more than 50 users. The actual size of the store can increase with the size of the venue.
The entire self-service shopping process is as follows:
The first step: into the store
For the first time into the store, the first time the consumer enters the store, he needs to open “Mobile Taobaoâ€, scan the code to obtain the e-admission code, and sign the data usage, privacy protection statement, Alipay deduction agreement and other terms, then go through the gate and start shopping ( In the future, it is no longer necessary to take out the mobile phone.)
Step 2: Choose goods
In Amoy Coffee, users are free to pick up any of the same items or order in the dining area. This is no different from everyday shopping until the user leaves the store.
Step 3: Pay
Before leaving the store, the user must go through a "settlement gate" or "kick-off door." It consists of two doors. When the first door senses the user's demand for departure, it automatically opens; after a few seconds, the second door will open and the "Settlement Gate" has completed the deduction. Then, the reminder next to it will say: "Alipay's total deduction of XX yuan."
Then you can leave smartly.
Ren Xiaofeng is coming here for this
In the middle of 2013-17, Ren Xiaofeng became the project leader of Amazon Go. He reinvented a new retail model using computer vision and machine learning, and launched the world's first "take away" store. The "just walk out technology" was studied by Ren Xiaofeng and his team. The principle is that after entering the supermarket, there will be face recognition at the entrance, cameras on the shelves, infrared sensors and pressure sensing devices to determine which products the customer has selected and how many products have been returned. The microphone in the store can determine the location of the consumer based on ambient sounds. . All collected information is transmitted to the information center of the Amazon Go store. When the customer leaves the store, the sensor scans the items purchased by the customer and settles automatically. There will be no delay throughout.
No major retail explosion?
The concept of unmanned convenience stores has long existed. Lei Fengnet has learned that there are mainly the following solutions:
The first kind is close to common people's unmanned vending machines, but the scale is bigger and the product category is more abundant. The biggest drawback of this solution is that consumers cannot touch the product directly before payment, and the user experience is not good.
The second is the self-service settlement system in large supermarkets (such as Tianhong) that are most familiar to everyone. After the consumer selects a good product, he/she can scan the product at the self-service cashier to settle it. The drawback of this scheme is that it is difficult to check and monitor. How can we ensure that consumers do not pay more? It is obviously unrealistic to challenge human nature.
The third type is the use of RFID chips. This solution has a long history and is relatively mature in technology, but its cost is high. In addition, there are also fatal flaws in RFID chips such as thunderstorms and induction in liquid chambers.
The fourth scheme is a machine vision intelligent identification program represented by Amazon's Amazon Go. Looking at the shopping process of Amoy Coffee, it should also use machine vision technology.
The difficulty of this solution lies in that as the scale of the shop expands, the amount of computation of the system will soar, which poses a great challenge to the GPU. Even if you leave the cost aside, the accuracy of recognition is difficult to guarantee.
Tianruo Technology CEO Chen Weilong believes that the image recognition system used in unmanned convenience stores is accurate in small-scale scenarios, such as 10 square meters, 100 commodity categories, and 2 users, and high capital and technical investment. The rate is acceptable. But if you want to further expand, you will face very big challenges, and this is the only way for unmanned convenience stores. Because even if we do not consider the various factors of commercialization, from the technology to the application point of view, this convenience store must reach 50 square meters, 300 commodity categories (and can not be specific) above, have application value. And if it's a supermarket, it's even harder to get it because if you model, train, and extract features for 100,000 items, it's a very, very large amount of work, then it's on the shelf, or on the shelf. The combination of training is also a very large system project.
In addition, credit problems are also a major challenge faced by unsold merchandise stores. For example, if customers maliciously mess up products, destroy store hygiene, or someone deliberately covers the camera, Amazon Go’s system cannot identify them. Although Mr. Chen Haibo, founder of Shenlan Technology, believes that no-consumer convenience stores can access sesame credit scores in the future, they can regulate consumer behavior through credit scoring mechanisms. However, in the absence of universal understanding of the credit system, the effect is still unknown.
In general, Amazon Go has done a very good job technically. However, at the current stage, its cost is too high. Many supermarkets are now losing money. This kind of plan will not cause fundamental changes. Today, Ma Yun launched the first solution to the problem and waited for Ma Yuan to solve the second third problem on the 8th Taobao Creativity Festival in 2017 from July 8 to 12.
Thermal Printer ,Thermal Label Printer,Thermal Receipt Printer,Bluetooth Thermal Printer
ShengXiaoBang(GZ) Material Union Technology Co.Ltd , https://www.sxbgz.com