WO2019196500A1

WO2019196500A1 - Self-service retail system and control method therefor, vending apparatus and cloud server

Info

Publication number: WO2019196500A1
Application number: PCT/CN2018/122778
Authority: WO
Inventors: 季涛; 吕秀凤
Original assignee: 合肥美的智能科技有限公司
Priority date: 2018-04-11
Filing date: 2018-12-21
Publication date: 2019-10-17
Also published as: CN108564442A

Abstract

A self-service retail system and a control method therefor, a vending apparatus (100) and a cloud server (200). The self-service retail system comprises the vending apparatus (100) and the cloud server (200), wherein the vending apparatus (100) comprises: multiple goods shelves (110); multiple weight sensors (120) for respectively performing weight detection on the multiple goods shelves (110), so as to generate weight detection signals; multiple cameras (130) for respectively acquiring images of goods on the multiple goods shelves (110); a communicator (140) for communicating with the cloud server (200); and a first controller (150) for sending the weight detection signals and the images of the goods to the cloud server (200); and the cloud server (200) is used for acquiring, according to the weight detection signals, a target goods shelf, the weight of which changes, and performing settlement according to a weight detection signal of the target goods shelf and a corresponding goods image. The self-service retail system improves the accuracy of recognizing goods to be paid for, and can effectively solve the problem of the losses of a merchant or a user caused by an error during money deduction and can also bring about a better usage experience for the user.

Description

Unmanned retail system and its control method, vending device and cloud server

Cross-reference to related applications

This application claims that the invention titled "Unmanned retail system and its control method, vending device and cloud server" submitted by Hefei Midea Intelligent Technology Co., Ltd. on April 11, 2018, the Chinese patent application number is "201810322658.2" priority.

Technical field

The present application relates to the field of automation equipment technologies, and in particular, to an unmanned retail system, a control method for an unmanned retail system, a vending device, and a cloud server.

Background technique

In the related art, there is a new type of unmanned retail container, which can be opened by using a mobile phone scan code, and the door is automatically paid for. There are various ways to identify the product: 1 RFID (Radio Frequency IDentification) tag recognition mode; 2 in the shelf The way the gravity sensor is installed at the bottom; 3 the way of image recognition.

However, although the first method for identifying a product has a good experience, the label cost is high; the second method for identifying the product has a low cost and a good sense of experience, but once the user misplaces the product, the background settlement is easy. Error; the third way to identify the product is because the current recognition rate is not as high as the previous two, so it does not have the conditions for completely separate use.

Summary of the invention

The present application aims to solve at least one of the technical problems in the above-mentioned techniques to some extent.

To this end, the first object of the present application is to provide an unmanned retail system, which improves the accuracy of identifying the goods to be paid, and can effectively solve the problem of loss of the merchant or the user due to the wrong payment, and can Users bring a better experience.

A second object of the present application is to propose a control method for an unmanned retail system.

A third object of the present application is to propose a vending apparatus.

A third object of the present application is to propose a cloud server.

To achieve the above objective, the first aspect of the present application provides an unmanned retail system, including a vending device and a cloud server, wherein the vending device includes: a plurality of shelves; a plurality of weight sensors for Performing weight detection on the plurality of shelves to generate a weight detection signal; a plurality of weight sensors for respectively performing weight detection on the plurality of shelves to generate a weight detection signal; and a communicator for communicating with the cloud server Communicating; a first controller, configured to send, by the cloud server, a weight detection signal of the plurality of weight sensors and an image of the product acquired by the plurality of cameras; The weight detection signals of the weight sensors acquire the target shelf whose weight changes, and settle according to the weight detection signal of the target shelf and the corresponding product image.

According to the unmanned retail system of the embodiment of the present application, a plurality of shelves are respectively subjected to weight detection by a plurality of weight sensors in the vending device to generate a weight detection signal, and respectively, through a plurality of weight sensors in the vending device The plurality of shelves perform weight detection to generate a weight detection signal, and send the weight detection signals of the plurality of weight sensors and the product images acquired by the plurality of cameras to the cloud server through the first controller in the vending device, and then the server according to the The weight detection signal of the weight sensor acquires the target shelf whose weight changes, and settles according to the weight detection signal of the target shelf and the corresponding product image. Thereby, the accuracy of identifying the to-be-paid item is improved, and the problem of loss of the merchant or the user due to the error of the deduction can be effectively solved, and the user can be brought a better experience.

In addition, the unmanned retail system proposed according to the above embodiment of the present application may further have the following additional technical features:

In an embodiment of the present application, the cloud server acquires a weight difference between the target shelf and the front and rear of the vending device, and acquires a single piece weight of the product corresponding to the target shelf, and detects and confirms the The weight difference is an integral multiple of the weight of the single piece, and is settled based on the weight difference and a multiple of the weight of the piece.

In an embodiment of the present application, the cloud server is further configured to detect and confirm that the weight difference is not an integer multiple of the weight of the single piece, and identify an item image corresponding to the target shelf to obtain the target shelf. Above the merchandise, the merchandise above the target shelf includes merchandise of other shelves, obtaining a single piece weight of the merchandise of the other shelf, and according to the weight difference, the weight of the single piece of the other shelf, and the The weight of the single item of the target shelf is settled.

In an embodiment of the present application, the cloud server is further configured to detect and confirm that the goods on the target shelf include goods of other shelves, and control the vending device to issue prompt information to prompt the purchaser.

In an embodiment of the present application, the cloud server is further configured to: confirm that the purchaser does not restore the goods of other shelves above the target shelf, and update the product list of the target shelf.

In an embodiment of the present application, the cloud server extracts an item from an item image corresponding to the target shelf, and generates a confidence level of each item, and the item with a confidence greater than a preset threshold and a preset item model Matching is performed to obtain the merchandise above the target shelf.

In one embodiment of the present application, the predetermined merchandise model is obtained by learning or training.

In an embodiment of the present application, the cloud server is further configured to: identify an item on the target shelf, including an item of a non-sale device, and notify the purchaser through the vending device, and maintain the product The personnel will give an alarm.

In an embodiment of the present application, the vending apparatus further includes: a gating module for controlling a door of the vending device; and a scan code module for confirming that the user scans the code, the door The control module controls the door of the vending device to be turned on.

In an embodiment of the present application, the vending apparatus further includes: a shopping monitoring module, configured to monitor behavior of the shopper.

In an embodiment of the present application, the vending apparatus further includes: a lighting module, configured to confirm that the door of the vending device is turned on, perform illumination, and fill light the opening of the plurality of cameras.

In order to achieve the above object, the second aspect of the present application provides a control method for an unmanned retail system, where the unmanned retail system includes a vending device and a cloud server, and the vending device communicates with the cloud server. The vending apparatus includes a plurality of shelves, and the control method includes the following steps: the vending apparatus separately performs weight detection on the plurality of shelves to generate a weight detection signal; and the vending apparatus respectively acquires the plurality of shelves An image of the merchandise on the shelf; the vending device transmits the weight detection signal and the merchandise image to the cloud server; the cloud server acquires a target shelf whose weight changes according to the weight detection signal, and The settlement is performed according to the weight detection signal of the target shelf and the corresponding product image.

According to the control method of the unmanned retail system of the embodiment of the present application, first, the vending apparatus performs weight detection on the plurality of shelves to generate a weight detection signal, and respectively acquires images of the plurality of shelves, and the weight detection signals and The product image is sent to the cloud server, and then the cloud server acquires the target shelf whose weight changes according to the weight detection signal, and settles according to the weight detection signal of the target shelf and the corresponding product image. Thereby, the accuracy of identifying the to-be-paid item is improved, and the problem of loss of the merchant or the user due to the error of the deduction can be effectively solved, and the user can be brought a better experience.

In addition, the control method of the unmanned retail system proposed according to the above embodiment of the present application may further have the following additional technical features:

In an embodiment of the present application, the cloud server acquires a target shelf whose weight changes according to the weight detection signal, and performs settlement according to the weight detection signal of the target shelf and the corresponding product image, including: the cloud Obtaining, by the server, a weight difference between the target shelf and the front and rear of the vending device, and acquiring a single piece weight of the product corresponding to the target shelf, detecting and confirming that the weight difference is an integral multiple of the weight of the single piece And settle according to the weight difference and a multiple of the weight of the single piece.

In an embodiment of the present application, the cloud server detects and confirms that the weight difference is not an integer multiple of the weight of the single piece, and identifies an item image corresponding to the target shelf to obtain an item above the target shelf. The merchandise above the target shelf includes merchandise of other shelves, obtaining a single piece weight of the merchandise of the other shelf, and according to the weight difference, the weight of the single piece of the other shelf, and the commodity of the target shelf Single piece weight is settled.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the cloud server detecting and confirming that the item on the target shelf includes goods of other shelves, and controlling the vending device to issue a prompt Information to prompt the buyer.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the cloud server confirms that the purchaser does not restore the items of the other shelves above the target shelf, and updates the target shelf. List of products.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the cloud server identifying that the merchandise above the target shelf includes a merchandise of the non-sale device, through the vending device The purchaser alerts and alerts the maintenance personnel.

In an embodiment of the present application, the vending apparatus further includes a gating module, the control method further comprising: the vending device confirms that the user scans the code, and the gating module controls the vending device The door is opened.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the vending device monitoring the behavior of the shopper.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the vending device confirms that the door is opened, performs illumination, and separately acquires images of the plurality of shelves above the product, and performs illumination.

In order to achieve the above objective, the third aspect of the present application provides a vending apparatus, including: a plurality of shelves; and a plurality of weight sensors for respectively performing weight detection on the plurality of shelves to generate a weight detection signal; a plurality of cameras for respectively acquiring images of the commodities on the plurality of shelves; a communicator for communicating with the cloud server; and a first controller for detecting weights of the plurality of weight sensors and Sending the product image acquired by the plurality of cameras to the cloud server, so that the cloud server acquires a target shelf whose weight changes according to the weight detection signals of the plurality of weight sensors, and according to the weight of the target shelf The detection signal and the corresponding product image are settled.

According to the vending apparatus of the embodiment of the present application, the weight detection is performed on the plurality of shelves by the plurality of weight sensors to generate the weight detection signal, and the images of the products on the plurality of shelves are respectively acquired by the plurality of cameras, and the first The controller sends the weight detection signal of the plurality of weight sensors and the product image acquired by the plurality of cameras to the cloud server, so that the cloud server acquires the target shelf whose weight changes according to the weight detection signals of the plurality of weight sensors, and according to the target The weight detection signal of the shelf and the corresponding product image are settled. Thereby, the accuracy of identifying the to-be-paid item is improved, and the problem of loss of the merchant or the user due to the error of the deduction can be effectively solved, and the user can be brought a better experience.

In addition, the vending apparatus proposed according to the above embodiment of the present application may further have the following additional technical features:

In an embodiment of the present application, the vending apparatus further includes: a gating module, configured to control a door of the vending device; and a scan code module, configured to confirm that a user scans the code, the gating module The door of the vending device is controlled to be turned on.

In order to achieve the above objective, a fourth embodiment of the present application provides a cloud server, including: a receiving module, configured to receive a weight detection signal and a product image sent by a vending device; and a settlement module, configured to detect a signal according to the weight Obtaining a target shelf whose weight has changed, and performing settlement according to the weight detection signal of the target shelf and the corresponding product image.

According to the cloud server of the embodiment of the present application, the receiving module receives the weight detection signal and the product image sent by the vending device, and the settlement module acquires the target shelf whose weight changes according to the weight detection signal, and detects according to the weight of the target shelf. The signal and the corresponding product image are settled. Thereby, the accuracy of identifying the to-be-paid item is improved, and the problem of loss of the merchant or the user due to the error of the deduction can be effectively solved, and the user can be brought a better experience.

In addition, the cloud server proposed according to the foregoing embodiment of the present application may further have the following additional technical features:

In an embodiment of the present application, the settlement module acquires a weight difference between the target shelf before and after the door opening and closing of the vending device, and acquires a single piece weight of the product corresponding to the target shelf, and detects and confirms the The weight difference is an integral multiple of the weight of the single piece, and is settled based on the weight difference and a multiple of the weight of the piece.

In an embodiment of the present application, the settlement module is further configured to detect and confirm that the weight difference is not an integer multiple of the weight of the single piece, and identify an image of the product corresponding to the target shelf to obtain the target shelf. Above the merchandise, the merchandise above the target shelf includes merchandise of other shelves, obtaining a single piece weight of the merchandise of the other shelf, and according to the weight difference, the weight of the single piece of the other shelf, and the The weight of the single item of the target shelf is settled.

In an embodiment of the present application, the settlement module is further configured to: detect and confirm that the commodity on the target shelf includes goods of other shelves, and control the vending device to issue a prompt message to prompt the purchaser.

In an embodiment of the present application, the settlement module is further configured to confirm that the purchaser does not restore the goods of the other shelves above the target shelf, and update the product list of the target shelf.

In an embodiment of the present application, the settlement module extracts an item from the product image corresponding to the target shelf, and generates a confidence level of each item, and the item with a confidence greater than a preset threshold and a preset item model Matching is performed to obtain the merchandise above the target shelf.

In an embodiment of the present application, the settlement module is further configured to: identify an item on the target shelf, including an item of a non-sale device, and notify the purchaser through the vending device, and maintain the product The personnel will give an alarm.

The advantages of the additional aspects of the present application will be set forth in part in the description which follows.

DRAWINGS

1 is a block schematic diagram of an unmanned retail system in accordance with one embodiment of the present application;

2 is a block schematic diagram of an unmanned retail system in accordance with another embodiment of the present application;

3 is a block schematic diagram of an unmanned retail system in accordance with yet another embodiment of the present application;

4 is a block schematic diagram of an unmanned retail system in accordance with yet another embodiment of the present application;

5 is a flow chart of a control method of an unmanned retail system according to a specific example of the present application;

6 is a flow chart of a method of controlling an unmanned retail system according to an embodiment of the present application;

Figure 7 is a block schematic diagram of a vending apparatus in accordance with one embodiment of the present application;

Figure 8 is a block schematic diagram of a vending apparatus in accordance with another embodiment of the present application;

9 is a block schematic diagram of a cloud server in accordance with one embodiment of the present application.

detailed description

The embodiments of the present application are described in detail below, and the examples of the embodiments are illustrated in the drawings, wherein the same or similar reference numerals are used to refer to the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are intended to be illustrative, and are not to be construed as limiting.

The unmanned retail system, the control method of the unmanned retail system, the vending device and the cloud server of the embodiment of the present application are described below with reference to the accompanying drawings.

1 is a block schematic diagram of an unmanned retail system in accordance with one embodiment of the present application.

As shown in FIG. 1 , the unmanned retail system of the embodiment of the present application includes: a vending apparatus 100 and a cloud server 200. The vending apparatus 100 includes: a plurality of shelves 110, a plurality of weight sensors 120, and a plurality of cameras 130. The communicator 140 and the first controller 150. It should be noted that, for each of the plurality of shelves 110 described in this embodiment, only one item is initially placed, and the plurality of shelves 110 can be placed in the plurality of shelves in the vending apparatus 100. Inside the rack, the shelf 110 can be a tray.

The communicator 140 communicates with the cloud server 200. The communicator 140 can perform wireless communication with the cloud server 200.

The plurality of weight sensors 120 respectively perform weight detection on the plurality of shelves 110 to generate a weight detection signal.

In the embodiment of the present application, the weight sensor 120 may be a gravity sensor, and at least one weight sensor 120 may be installed in each shelf 110 to facilitate weight detection of the plurality of shelves 110.

The plurality of cameras 130 respectively acquire images of the plurality of shelves 110.

Further, as shown in FIG. 2, the vending apparatus 100 may further include a lighting module 160 for performing illumination when the door of the vending apparatus 100 is opened, and for supplementing light when a plurality of cameras are turned on.

In the embodiment of the present application, a camera 130 may be installed above each shelf 110 to acquire images of products on the plurality of shelves 110, or a plurality of cameras 130 may be separately installed on each floor of the vending apparatus 100. The sides or top of the shelf are used to take pictures of each shelf to obtain images of the items above the plurality of shelves 110.

Alternatively, color information is susceptible to various factors such as the color and direction of the light source in view of the image being captured and transmitted. In order to avoid such color shift, after acquiring the image of the product, the camera 130 may first perform image enhancement processing on the image of the product, which may include contrast enhancement, spatial filtering, noise cancellation, image smoothing, and image sharpening.

Further, as shown in FIG. 3, the vending apparatus 100 may further include: a gating module 170 and a scanning module 180.

The door control module 170 is used to control the door of the vending apparatus 100.

The scan code module 180 is configured to control the door of the vending apparatus 100 to be turned on by the gating module 170 after the user scans the code.

In the embodiment of the present application, the user can scan or paste the two-dimensional code on the surface of the vending device 100 through a mobile terminal (for example, a mobile phone, a tablet computer, or a computer) to control the door of the vending device 100. Open. Specifically, after scanning the two-dimensional code, the mobile terminal acquires an identity verification instruction in the two-dimensional code, acquires user information according to the identity verification instruction, and sends the information to the scan code module 180, and then The scan code module 180 can send the user information to the cloud server 200, and the cloud server 200 checks the user identity, and after the verification is successful, sends a door open command to the scan code module 180, and after the scan code module 180 receives the door open command, the scan code module 180 passes the The gating module 170 controls the door of the vending apparatus 100 to be turned on.

It should be noted that the two-dimensional code described in this embodiment may include an account authorization request instruction, and when the user first scans the two-dimensional code through the mobile terminal, the mobile terminal acquires the account authorization request instruction, and according to The account authorization request instruction requests the user whether to authorize the account to automatically deduct the right to charge. If the user confirms the authority to grant the account automatic deduction, the cloud server 200 may have the right to automatically deduct the amount of each purchase from the user's account. The account authorization request instruction may include an Alipay account authorization request instruction, a WeChat account authorization request instruction, and the like.

In addition, when the user confirms the right to grant the account automatic deduction, the mobile terminal also sends the user information in the account to the cloud server 200, so as to subsequently debit the identity of the user.

The first controller 150 transmits the weight detection signals of the plurality of weight sensors 120 and the product images acquired by the plurality of cameras 130 to the cloud server 200.

The cloud server 200 acquires a target shelf whose weight has changed based on the weight detection signals of the plurality of weight sensors 120, and performs settlement based on the weight detection signal of the target shelf and the corresponding product image. Among them, the target shelf can be multiple.

Further, the cloud server 200 can obtain the weight difference of the target shelf before and after the door opening and closing of the vending device 100, and obtain the weight of the single item corresponding to the target shelf, and determine whether the weight difference is an integral multiple of the weight of the single piece, if it is an integer The double is settled based on the weight difference and the multiple of the individual piece weight.

Specifically, when the user wants to purchase goods from the vending apparatus 100 of the present application, the two-dimensional code on the surface of the vending apparatus 100 can be scanned and pasted by the mobile phone to control the door of the vending apparatus 100 to be opened. Then, the user can select the items to be purchased from the plurality of shelves 110 in the vending apparatus 100, and close the door of the vending apparatus 100 after the selection. At this time, the plurality of weight sensors 120 can respectively perform the plurality of shelves 110. The weight detection generates a weight detection signal and transmits it to the first controller 150, and the plurality of cameras 130 can respectively acquire images of the plurality of shelves 110 and send them to the first controller 150.

Then, the first controller 150 transmits the received weight detection signals of the plurality of weight sensors 120 and the product images acquired by the plurality of cameras 130 to the cloud server 200 through the communicator 140. After receiving the weight detection signal and the product image, the cloud server 200 can acquire the target shelf whose weight changes according to the weight detection signal, and obtain the weight difference between the target shelf before and after the door opening and closing of the vending device 100, and acquire the target shelf. Whether the weight difference is an integral multiple of the weight of the single piece, and if it is an integral multiple, the settlement is performed according to the weight difference and the multiple of the single piece weight.

For example, if the item preset in the A shelf 110 is 10 Coca-Cola, the weight is 240g/piece, and the preset item in the B shelf 110 is 10 pieces of chewing gum, 50g/piece, after the user closes the door, if the A shelf is passed The weight sensor 120 installed in the 110 detects that the weight in the A shelf 100 is reduced by 240g, and the weight in the B shelf 110 is reduced by 50g, indicating that the weight difference of the A shelf 110 before and after the opening and closing of the vending device 100 is A in the shelf 110. The integral weight of the B shelf 110 before and after the door opening and closing of the vending device 100 is an integral multiple of the weight of the single piece in the B shelf 110, whereby the cloud server 200 can determine that the user vending device 100 has taken away A Coca-Cola and a chewing gum, and deducting the price of a Coca-Cola and a chewing gum from the user's account.

In addition, in an embodiment of the present application, the cloud server 200 may be further configured to identify an item image corresponding to the target shelf to obtain an item above the target shelf after determining that the weight difference is not an integer multiple of the weight of the single piece. When the goods on the target shelf include other shelves, the individual pieces of the goods of the other shelves are obtained, and the settlement is based on the weight difference, the weight of the individual pieces of the other shelves, and the weight of the single piece of the target shelf.

The cloud server 200 extracts the product from the product image corresponding to the target shelf, generates a confidence level of each product, and matches the product with a confidence greater than a preset threshold with the preset product model to obtain the target shelf. commodity. The preset product model is obtained through learning or training, and the preset threshold may be calibrated according to actual conditions. For example, the preset threshold may be 95%. It should be noted that the preset product model described in this embodiment may include a product database, and the database includes information of all products in the vending device 100, for example, the name, weight, and price of each product. , quantity, and so on.

In another embodiment of the present application, the cloud server 200 is further configured to control the vending apparatus 100 to issue a prompt message to prompt the purchaser when determining that the item above the target shelf includes the item of the other shelf. Wherein, if the purchaser does not restore the items of other shelves above the target shelf, the cloud server 200 may also update the item list of the target shelf.

For example, if the item preset in the A shelf 110 is 10 Coca-Cola, the weight is 240g/piece, and the preset item in the B shelf 110 is 10 pieces of chewing gum, 50g/piece, after the user closes the door, if the A shelf is passed The weight sensor 120 installed in the 110 detects that the weight in the A shelf 110 is increased by 50g, and the weight in the B shelf 110 is reduced by 100g, which means that 50g of the item is placed in the A shelf 110, and 2 pieces of chewing gum are taken in the B shelf 110. At this time, the cloud server 200 can further confirm by receiving the image of the received product, and the conclusion obtained by the identification of the product image is: A shelf 110 has Coca-Cola and chewing gum, and it is judged that a sticker is placed in the A shelf 110. The goods in the other shelves 110 are correct with the preset information. Therefore, it is judged that the user mistakenly puts a chewing gum in the B shelf 110 into the A shelf 110, and takes away a chewing gum. At this time, the cloud server 200 deducts the price of one chewing gum instead of deducting two chewing gums, thereby improving the recognition. The accuracy rate of paid products can effectively solve the problem of loss of merchants or users due to wrong deductions, and bring a better user experience.

At the same time, the cloud server 200 will remind the user and guide the user to put the gum in the A shelf 110 back to the B shelf 110. If the user is not executed within a certain time, the cloud server 200 will preset the goods in the A shelf 110. Adding a chewing gum; that is, there are 10 Coca-Cola and 9 chewing gums remaining in the vending device 100, wherein 10 Coca-Cola and 1 chewing gum are placed in the A shelf 110, and 8 chewing gums are placed in the B shelf 110. When the user next purchases, if the gum in the A shelf 110 is taken out, the cloud server 200 judges according to the current preset item in the tray 1.

Further, if the weight sensor 120 recognizes that the A shelf 110 is reduced by 530g after the shopping is closed again, since the cloud server 200 has corrected the preset items in the A shelf 110 to 10 Coca-Cola and 1 chewing gum, it is assumed to be taken out. X Coca-Cola and Y chewing gum, list the equation:

240X+50Y=530,

-10≤X≤10, -9≤Y≤1, X, Y are integers;

The above equation is calculated by the formula exhaustive method (enumeration method), and X=2, Y=1 is obtained; that is, the user takes out two Coca-Cola and one chewing gum from the A shelf 110, and the cloud server 200 can perform the process. A reasonable checkout deduction will prevent the user from inadvertently placing the wrong item in the wrong position, resulting in loss of the merchant or user.

It should be noted that when a plurality of products are placed in an incorrect position, the method is analogized according to the above method, but the position of the placed product is not too much wrong. When the number of items placed in the wrong position exceeds a certain number, the cloud server 200 can send an alarm message. For the staff, the staff is required to regularly check the position of the goods.

In addition, the cloud server 100 stores in advance the product information of the vending apparatus 100 (for example, including but not limited to the following information: product name, brand, category, quantity, weight, net content, shelf life, supplier, and shelf date). When the user opens the door by scanning the two-dimensional code of the vending device 100 through the mobile terminal, the mobile terminal can access the cloud server, and acquire the product information of the vending device 100 stored in advance in the cloud server, and provide the user with the product information, so as to The user edits and views.

Optionally, the cloud server 200 is further configured to notify the purchaser through the vending device 100 if the item above the target shelf includes the item of the non-sale device, and alert the maintenance personnel.

Optionally, as shown in FIG. 4, the vending apparatus 100 may further include a shopping monitoring module 190 for monitoring the behavior of the shopper. Specifically, the shopping monitoring module 190 can monitor in real time whether the user has bad behavior during shopping (for example, estimating the destruction of the shelf 110) through the built-in camera, and if so, send an alarm message to the relevant worker.

In order to make the present application clear to those skilled in the art, FIG. 5 is a flowchart of a control method of an unmanned retail system according to a specific example of the present application, it should be noted that the tray described in the flowchart may be The applied shelf 110, as shown in FIG. 5, the control method of the unmanned retail system may include the following steps:

S101, whether the gate signal is related, if yes, executing step S102; if not, repeating step S101.

S102, whether the weight of the tray m changes, wherein m is a positive integer not zero, if yes, step S103 is performed; if not, the flow ends.

S103, whether the weight change of the tray m is an integral multiple of the current product weight on the tray m, and if yes, step S104 is performed; if not, step S106 is performed.

S104, whether the weight of the tray m is reduced, if yes, step S105 is performed; if no, step S111 is performed.

S105, the product is taken out in the tray m.

S106: Uploading a cloud server to each shelf, the cloud server divides the image according to the tray interval, and respectively identifies the product image in each tray.

S107, the product preset in each tray forms a product set T=[T1, T2···Ts], and Ts corresponds to the preset product weight Ts0, wherein s can be a positive integer not zero. n products are identified in the tray m, and each product forms a product set according to the reliability of image recognition, and a product set Sn=[Sn1, Sn2···Snx] with a confidence rate exceeding 90% is selected, wherein x is not A positive integer of zero, taking the intersection Rn=Sn∩T, the n products in the tray m are Rm1, Rm2···Rmn, respectively.

S108, it is determined whether Rmn Sn∩T is ≠0, if yes, step S109 is performed, and if no, step S102 is performed.

In S109, the identified Rmn set is the product identified by the image, and the product set after the last closed door is subtracted from the closed product set, and is greater than 0. If yes, step S110 is performed, and if no, step S111 is executed.

S110, the equation is listed, and the enumeration method (exhaustive method) is used to calculate the number of items to be taken out, and the product difference set before and after closing is the taken item.

S111, the goods in the tray are put into other trays in the vending device, reminding the user to put the goods back to the original place, and recording the misplaced position as the reference basis for the deduction of the next purchase.

In S112, a product other than the vending device is placed in the tray, and the user is prompted not to put the product in the non-sale device, and the alarm information is sent to the staff.

S113, calling the price preset for each commodity, and performing the total price settlement deduction.

Further, in order to better illustrate the present application, the above steps are described in detail below:

In the above steps, the same product is placed in each tray, including the same product name, weight and price, and the same tray is not allowed to place different types of goods.

In the above steps, the products placed in the unmanned retail cabinets are not allowed to have the same weight if they are of different product names or different prices.

In the above step S103, whether the weight change of the tray m is an integral multiple of the current product weight on the tray m, and if not, it indicates that the product that is not in the tray preset is placed in the tray m, and further passes the image. Identify and confirm the wrong product name and quantity to correct the calculations that follow.

The specific example is as follows. Example 1: If there is only one preset product Q on the tray m, the weight change ΔKm is a times the product Q (a is a natural number), and it means that a product Q is taken in the tray m. Example 2: If there is more than one preset item Q in the tray m (this step is recorded by the image recognition after the last time the door is closed), there are other preset items in the tray, for example, the tray m has the product Q, Further, when the product V is preset on the tray j, it is judged whether or not the weight change ΔKm is an integral multiple of the weight Qz of the product Q and the weight Vz of the product V, that is, ΔKm=p*Vz+l*Qz (p, both It is a natural number). If it is, it means that p items V and 1 item Q are taken off on the tray m; if not, the misplaced items are further confirmed by image recognition; if there are multiple items on the tray m, And so on.

In the above step S106, the camera may take a picture according to the manner of one camera per tray, or may take a separate photograph of each shelf, and then use the spacing feature between the trays to divide the space to further identify the products on each tray. For example, the width of each tray is known, and the picture of this layer can be divided according to the known width.

In the above step S107, there are known s pallets, and the product name preset in each tray is Ts, and the set is T=[T1, T2···Ts], for example, s=5, assuming the product set T=[ Uniform iced tea, Coca-Cola, Wangzi yoghurt, unified instant noodles, Wahaha], through image recognition, after identifying the door closed, there is one kind of goods in the tray 1, the possibility of this product is unified ice black tea is 98%, for Wang Laoji The possibility of herbal tea is 95%, the possibility of milk for Wangzi is 80%···, and the product with the confidence rate of more than 90% in the tray 1 is S1=[unified iced black tea, Wang Laoji herbal tea], and the intersection Rn is taken. =Sn∩T=[Unified iced black tea, Wang Laoji herbal tea]∩[Unified iced black tea, Coca-Cola, Wangzi yoghurt, unified instant noodles, Wahaha]=Unified iced black tea, it is determined that there is a unified ice black tea in the tray 1 before closing. Thereby, it is possible to effectively utilize the product information already placed in the vending device, reduce the image recognition range, and indirectly improve the reliability of the image recognition.

Example 1: After scanning the code, the weight sensor detects that the weight of the product in the tray 1 has changed, and the weight is 500g. If the weight of a unified iced tea is preset to 250g, then 500/250=2, 2 is The integer is greater than zero, so the user is determined to take out 2 bottles of uniform iced tea.

Example 2: The user scans the code and closes the door after shopping. The weight sensor detects that the weight of the product in the tray 4 changes to 100 g, and the uniform instant noodle is placed in the tray 4, and the preset weight is 100 g. It is detected that the weight of the product in the tray 1 is changed to 400g, and the uniform ice black tea is placed in the tray 1, the preset weight is 250g, and 400/250=1.6, 1.6 is not an integer, so the image recognition is further determined by the image recognition. After the merchandise and the vending device are closed, in the same manner as above, it is recognized that there are two kinds of merchandise in the tray 1. After the first merchandise is taken in the above manner, it is determined that the ice black tea is unified, and the second merchandise is intersected as described above. After that, it is determined that the uniform instant noodles, the preset weight of the unified instant noodles is 100 g, and the number of instant noodles recognized by the image is 1, it is determined that the user mistakenly puts a bucket of instant noodles in the tray 4 into the tray 1, which is known in each tray. The number of preset items is the maximum value, 0 is the minimum value, and the quantity of each item is an integer, which is calculated by the enumeration method (exhaustive method) (400+100)/250=2, which is taken by the user. The food is 2 uniform iced black teas, the unit price of unified iced black tea is called, the total price is calculated, and the deduction is charged; at the same time, the user is prompted to put the unified instant noodles in the tray 1 back into the tray 4, and if the user is not returned to the original position, , The information is sent to the correction module, for recording, a plurality of inner product detected in step S103 tray m, i.e., derived from this data.

After receiving the information, the cloud server adds a 100g unified instant noodle to the product preset in the tray 1. When there is another user shopping, the weight change of the product in the tray 1 will preferentially include 100g instant noodles, for example, the weight of the product in the tray 1. When the change is 350 g, it is determined that the user takes out a unified iced black tea and a unified instant noodle from the tray 1, and can further confirm by the missing items in the tray 1 recognized by the image.

In summary, according to the unmanned retail system of the embodiment of the present application, a plurality of shelves are respectively subjected to weight detection by a plurality of weight sensors in the vending device to generate a weight detection signal, and the plurality of weights are sensed through the vending device. Performing weight detection on the plurality of shelves respectively to generate a weight detection signal, and transmitting the weight detection signals of the plurality of weight sensors and the product images acquired by the plurality of cameras to the cloud server through the first controller in the vending device, and then The server acquires the target shelf whose weight changes according to the weight detection signals of the plurality of weight sensors, and performs settlement according to the weight detection signal of the target shelf and the corresponding product image. Thereby, the accuracy of identifying the to-be-paid item is improved, and the problem of loss of the merchant or the user due to the error of the deduction can be effectively solved, and the user can be brought a better experience.

6 is a flow chart of a method of controlling an unmanned retail system in accordance with one embodiment of the present application.

As shown in FIG. 6, the control method of the unmanned retail system of the embodiment of the present application includes the following steps:

S1. The vending apparatus performs weight detection on the plurality of shelves to generate a weight detection signal.

S2. The vending device respectively obtains images of the products on the plurality of shelves.

S3. The vending device sends the weight detection signal and the product image to the cloud server.

S4. The cloud server acquires a target shelf whose weight changes according to the weight detection signal, and performs settlement according to the weight detection signal of the target shelf and the corresponding product image.

In an embodiment of the present application, the cloud server acquires a target shelf whose weight changes according to the weight detection signal, and performs settlement according to the weight detection signal of the target shelf and the corresponding product image, including: the cloud server acquires the target shelf in the vending device. The weight difference between the front and the rear of the door is opened, and the weight of the single item corresponding to the target shelf is obtained, and whether the weight difference is an integral multiple of the weight of the single piece, and if it is an integral multiple, the settlement is performed according to the weight difference and the multiple of the single piece weight.

In an embodiment of the present application, after determining that the weight difference is not an integral multiple of the weight of the single piece, the cloud server identifies the product image corresponding to the target shelf to obtain the product above the target shelf, and the product above the target shelf When the products of other shelves are included, the weight of the individual items of the other shelves is obtained, and the settlement is based on the weight difference, the weight of the individual pieces of the other shelves, and the weight of the single piece of the target shelf.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: when the cloud server determines that the item on the target shelf includes the goods of the other shelf, the cloud control device sends the prompting information to the buyer. prompt.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the cloud server updating the item list of the target shelf when the purchaser does not restore the items of the other shelves above the target shelf.

In an embodiment of the present application, the cloud server extracts the commodity from the product image corresponding to the target shelf, and generates a confidence level of each commodity, and matches the commodity with a confidence greater than a preset threshold with the preset commodity model to Get the items above the target shelf.

In one embodiment of the present application, a predetermined merchandise model is obtained through learning or training.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the cloud server reminding the purchaser through the vending device when identifying that the item above the target shelf includes the item of the non-sale device And alert the maintenance personnel.

In an embodiment of the present application, the vending apparatus further includes a gating module, and the controlling method further includes: the vending device controls the door of the vending device to be turned on by the gating module after the user scans the code.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the vending device monitors the behavior of the shopper.

In an embodiment of the present application, the method for controlling the unmanned retail system further includes: the vending device illuminates when the door of the vending device is opened, and illuminates when acquiring images of the merchandise on the plurality of shelves respectively. .

It should be noted that, for details not disclosed in the control method of the unmanned retail system of the embodiment of the present application, refer to the details disclosed in the unmanned retail system of the embodiment of the present application, and details are not described herein again.

In summary, according to the control method of the unmanned retail system of the embodiment of the present application, the first selling device performs weight detection on a plurality of shelves to generate a weight detection signal, and respectively acquires images of the products on the plurality of shelves, and the weight. The detection signal and the product image are sent to the cloud server, and then the cloud server acquires the target shelf whose weight changes according to the weight detection signal, and settles according to the weight detection signal of the target shelf and the corresponding product image. Thereby, the accuracy of identifying the to-be-paid item is improved, and the problem of loss of the merchant or the user due to the error of the deduction can be effectively solved, and the user can be brought a better experience.

Figure 7 is a block schematic illustration of a vending apparatus in accordance with one embodiment of the present application.

As shown in FIG. 7, the vending apparatus 100 of the embodiment of the present application includes a plurality of shelves 110, a plurality of weight sensors 120, a plurality of cameras 130, a communicator 140, and a first controller 150.

The plurality of weight sensors 120 are configured to perform weight detection on the plurality of shelves 110 to generate a weight detection signal.

A plurality of cameras 130 are used to acquire images of items on a plurality of shelves 110, respectively.

The communicator 140 is configured to communicate with a cloud server.

The first controller 150 is configured to send the weight detection signals of the plurality of weight sensors 120 and the product images acquired by the plurality of cameras 130 to the cloud server, so that the cloud server acquires the weight according to the weight detection signals of the plurality of weight sensors 120. The target shelf of the change is settled according to the weight detection signal of the target shelf and the corresponding product image.

In an embodiment of the present application, as shown in FIG. 8, the vending apparatus 100 further includes a lighting module 160 for performing illumination when the door of the vending device is opened, and for supplementing light when the plurality of cameras are turned on.

In an embodiment of the present application, as shown in FIG. 8, the vending apparatus 100 further includes a gating module 170 for controlling the door of the vending apparatus 100, and the scan code module 180 is used after the user scans the code. The door of the vending apparatus 100 is controlled to be turned on by the gating module.

In one embodiment of the present application, as shown in FIG. 8, the vending apparatus 100 further includes a shopping monitoring module 190 for monitoring the behavior of the shopper.

It should be noted that, for details not disclosed in the vending apparatus of the embodiment of the present application, refer to the details disclosed in the unmanned retail system of the embodiment of the present application, and details are not described herein again.

In summary, according to the vending apparatus of the embodiment of the present application, the weight detection is performed on a plurality of shelves by a plurality of weight sensors to generate a weight detection signal, and images of the products on the plurality of shelves are respectively acquired by the plurality of cameras, and Sending, by the first controller, the weight detection signals of the plurality of weight sensors and the product images acquired by the plurality of cameras to the cloud server, so that the cloud server acquires the target shelf whose weight changes according to the weight detection signals of the plurality of weight sensors, And according to the weight detection signal of the target shelf and the corresponding product image for settlement. Thereby, the accuracy of identifying the to-be-paid item is improved, and the problem of loss of the merchant or the user due to the error of the deduction can be effectively solved, and the user can be brought a better experience.

As shown in FIG. 9, the cloud server 200 of the embodiment of the present application includes: a receiving module 210 and a settlement module 220.

The receiving module 210 is configured to receive the weight detection signal and the product image sent by the vending device.

The settlement module 220 is configured to acquire a target shelf whose weight changes according to the weight detection signal, and perform settlement according to the weight detection signal of the target shelf and the corresponding product image.

In an embodiment of the present application, the settlement module 220 obtains the weight difference of the target shelf before and after the door opening and closing of the vending device, and obtains the weight of the article corresponding to the target shelf, and determines whether the weight difference is an integral multiple of the weight of the single piece. If it is an integer multiple, it is settled based on the weight difference and the multiple of the weight of the piece.

In an embodiment of the present application, the settlement module 220 is further configured to: after determining that the weight difference is not an integer multiple of the weight of the single piece, identify the product image corresponding to the target shelf to obtain the product above the target shelf, when the target shelf When the goods on the top include goods of other shelves, the individual pieces of the goods of the other shelves are obtained, and the settlement is based on the weight difference, the weight of the individual pieces of the other shelves, and the weight of the single piece of the target shelf.

In an embodiment of the present application, the settlement module 220 is further configured to control the vending device to issue a prompt message to prompt the purchaser when determining that the item on the target shelf includes the goods of the other shelf.

In an embodiment of the present application, the settlement module 220 is further configured to update the item list of the target shelf when the purchaser does not restore the items of the other shelves above the target shelf.

In an embodiment of the present application, the settlement module 220 extracts an item from the product image corresponding to the target shelf, and generates a confidence level of each item, and matches the item with a confidence greater than a preset threshold with the preset product model. To get the goods on the target shelf.

In an embodiment of the present application, the settlement module 220 is further configured to notify the purchaser through the vending device if the item above the target shelf includes the item of the non-sale device, and alert the maintenance personnel.

It should be noted that, for details not disclosed in the cloud server of the embodiment of the present application, refer to the details disclosed in the unmanned retail system of the embodiment of the present application, and details are not described herein again.

In summary, according to the cloud server of the embodiment of the present application, the receiving module receives the weight detection signal and the product image sent by the vending device, and the settlement module acquires the target shelf whose weight changes according to the weight detection signal, and according to the target shelf. The weight detection signal and the corresponding product image are settled. Thereby, the accuracy of identifying the to-be-paid item is improved, and the problem of loss of the merchant or the user due to the error of the deduction can be effectively solved, and the user can be brought a better experience.

In the description of the present application, it is to be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", " Rear, Left, Right, Vertical, Horizontal, Top, Bottom, Inner, Out, Clockwise, Counterclockwise, Axial The orientation or positional relationship of the "radial", "circumferential" and the like is based on the orientation or positional relationship shown in the drawings, and is merely for the convenience of describing the present application and the simplified description, and does not indicate or imply the indicated device or The elements must have a particular orientation, are constructed and operated in a particular orientation, and are therefore not to be construed as limiting.

Moreover, the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, features defining "first" and "second" may include one or more of the features either explicitly or implicitly. In the description of the present application, the meaning of "a plurality" is two or more unless specifically and specifically defined otherwise.

In the present application, the terms "installation", "connected", "connected", "fixed" and the like shall be understood broadly, and may be either a fixed connection or a detachable connection, unless otherwise explicitly stated and defined. , or integrated; can be mechanical connection, or can be electrical connection; can be directly connected, or can be indirectly connected through an intermediate medium, can be the internal communication of two elements or the interaction of two elements. For those skilled in the art, the specific meanings of the above terms in the present application can be understood on a case-by-case basis.

In the present application, the first feature "on" or "below" the second feature may be the direct contact of the first and second features, or the first and second features are indirectly through the intermediate medium, unless otherwise explicitly stated and defined. contact. Moreover, the first feature "above", "above" and "above" the second feature may be that the first feature is directly above or above the second feature, or merely that the first feature level is higher than the second feature. The first feature "below", "below" and "below" the second feature may be that the first feature is directly below or obliquely below the second feature, or merely that the first feature level is less than the second feature.

In the description of the present specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" and the like means a specific feature described in connection with the embodiment or example. A structure, material or feature is included in at least one embodiment or example of the application. In the present specification, the schematic representation of the above terms is not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. In addition, various embodiments or examples described in the specification, as well as features of various embodiments or examples, may be combined and combined.

While the embodiments of the present application have been shown and described above, it is understood that the above-described embodiments are illustrative and are not to be construed as limiting the scope of the present application. The embodiments are subject to variations, modifications, substitutions and variations.

Claims

An unmanned retail system, characterized by comprising a vending device and a cloud server, wherein

The vending device includes:

Multiple shelves;

a plurality of weight sensors for respectively performing weight detection on the plurality of shelves to generate a weight detection signal;

a plurality of cameras for respectively acquiring images of goods on the plurality of shelves;

a communicator for communicating with the cloud server;

a first controller, configured to send a weight detection signal of the plurality of weight sensors and an image of the product acquired by the plurality of cameras to the cloud server;

The cloud server is configured to acquire a target shelf whose weight changes according to the weight detection signals of the plurality of weight sensors, and perform settlement according to the weight detection signal of the target shelf and the corresponding product image.
The unmanned retail system of claim 1 , wherein the cloud server acquires a weight difference between the target shelf before and after the opening and closing of the vending device, and acquires a single item corresponding to the target shelf The weight of the piece is detected and confirmed to be an integral multiple of the weight of the single piece, and is settled according to the weight difference and a multiple of the weight of the piece.
The unmanned retail system of claim 2, wherein the cloud server is further configured to detect and confirm that the weight difference is not an integer multiple of the weight of the single piece, and the product corresponding to the target shelf The image is identified to obtain an item above the target shelf, the item above the target shelf includes items of other shelves, and a piece weight of the item of the other shelf is obtained, and according to the weight difference, the other The weight of the single item of the shelf and the weight of the single piece of the target shelf are settled.
The unmanned retail system of claim 3, wherein the cloud server is further configured to detect and confirm that the merchandise above the target shelf includes merchandise of other shelves, and control the vending device to issue a prompt message. To prompt the buyer.
The unmanned retail system of claim 4, wherein the cloud server is further configured to: confirm that the purchaser has not restored items of other shelves above the target shelf, and update the target shelf Product list.
The unmanned retail system of claim 3, wherein the cloud server extracts merchandise from the merchandise image corresponding to the target shelf, and generates a confidence level of each merchandise, and the confidence is greater than a preset threshold. The item is matched with a preset item model to obtain an item above the target shelf.
The unmanned retail system of claim 6 wherein said predetermined merchandise model is obtained by learning or training.
The unmanned retail system of claim 3, wherein the cloud server is further configured to: identify an item on the target shelf comprising an item of a non-sale device, and purchase the item through the vending device Remind and alert the maintenance personnel.
The unmanned retail system of claim 1 wherein said vending apparatus further comprises:

a door control module for controlling a door of the vending device;

The scan code module is configured to confirm that the user scan code passes, and the gating module controls the door of the vending device to be turned on.
The unmanned retail system of claim 1 wherein said vending apparatus further comprises:

A shopping monitoring module for monitoring the behavior of shoppers.
The unmanned retail system of claim 1 wherein said vending apparatus further comprises:

And a lighting module, configured to confirm that the door of the vending device is turned on, perform illumination, and fill light to the plurality of cameras.
A method for controlling an unmanned retail system, characterized in that the unmanned retail system comprises a vending device and a cloud server, the vending device communicating with the cloud server, the vending device comprising a plurality of shelves The control method includes the following steps:

The vending device respectively performs weight detection on the plurality of shelves to generate a weight detection signal;

The vending device respectively obtains images of the commodities on the plurality of shelves;

The vending device transmits the weight detection signal and the product image to the cloud server;

The cloud server acquires a target shelf whose weight changes according to the weight detection signal, and performs settlement according to the weight detection signal of the target shelf and the corresponding product image.
The control method of the unmanned retail system according to claim 12, wherein the cloud server acquires a target shelf whose weight has changed according to the weight detection signal, and detects a signal according to the weight of the target shelf and a corresponding The product image is settled, including:

Obtaining, by the cloud server, a weight difference between the target shelf and the front and rear of the vending device, and acquiring a single piece weight of the product corresponding to the target shelf, detecting and confirming that the weight difference is the weight of the single piece An integer multiple of the settlement is based on the weight difference and a multiple of the individual piece weight.
The control method of the unmanned retail system according to claim 13, wherein the cloud server detects and confirms that the weight difference is not an integer multiple of the weight of the single piece, and the product image corresponding to the target shelf Identifying to obtain an item above the target shelf, the item above the target shelf includes items of other shelves, obtaining a piece weight of the item of the other shelf, and according to the weight difference, the other shelf The individual piece weight and the individual piece weight of the target shelf are settled.
The method of controlling an unmanned retail system according to claim 14, further comprising:

The cloud server detects and confirms that the goods on the target shelf include goods of other shelves, and controls the vending device to issue prompt information to prompt the purchaser.
The method of controlling an unmanned retail system according to claim 15, further comprising:

The cloud server confirms that the purchaser has not restored the items of the other shelves above the target shelf, and updates the item list of the target shelf.
The control method of the unmanned retail system according to claim 14, wherein the cloud server extracts goods from the product images corresponding to the target shelf, and generates a confidence level of each product, which is a confidence level. The item greater than the preset threshold is matched with the preset item model to obtain the item above the target shelf.
The method of controlling an unattended retail system according to claim 17, wherein said predetermined item model is obtained by learning or training.
The method of controlling an unmanned retail system according to claim 14, further comprising:

The cloud server recognizes that the commodity on the target shelf includes the commodity of the non-sale device, and reminds the purchaser through the vending device, and gives an alarm to the maintenance personnel.
The control method of the unmanned retail system of claim 12, wherein the vending device further comprises a gating module, the control method further comprising:

The vending device confirms that the user scan code passes, and the gating module controls the door of the vending device to be turned on.
The method for controlling an unmanned retail system according to claim 12, further comprising:

The vending device monitors the behavior of the shopper.
The method for controlling an unmanned retail system according to claim 12, further comprising:

The vending apparatus confirms that the door is opened, performs illumination, and separately acquires images of the items on the plurality of shelves and performs illumination.
A vending apparatus, comprising:

Multiple shelves;

a plurality of weight sensors for respectively performing weight detection on the plurality of shelves to generate a weight detection signal;

a plurality of cameras for respectively acquiring images of goods on the plurality of shelves;

a communicator for communicating with a cloud server;

a first controller, configured to send a weight detection signal of the plurality of weight sensors and an image of the product acquired by the plurality of cameras to the cloud server, so that the cloud server is configured according to the plurality of weight sensors The weight detection signal acquires a target shelf whose weight has changed, and settles according to the weight detection signal of the target shelf and the corresponding product image.
The vending apparatus of claim 23, further comprising:

a door control module for controlling a door of the vending device;

The scan code module is configured to confirm that the user scan code passes, and the gating module controls the door of the vending device to be turned on.
The vending apparatus of claim 23, further comprising:

A shopping monitoring module for monitoring the behavior of shoppers.
The vending apparatus of claim 23, further comprising:

And a lighting module, configured to confirm that the door of the vending device is turned on, perform illumination, and fill light to the plurality of cameras.
A cloud server, comprising:

a receiving module, configured to receive a weight detection signal and a product image sent by the vending device;

The settlement module is configured to acquire a target shelf whose weight changes according to the weight detection signal, and perform settlement according to the weight detection signal of the target shelf and the corresponding product image.
The cloud server according to claim 27, wherein the settlement module acquires a weight difference between the target shelf before and after the door opening and closing of the vending device, and acquires a single piece weight of the product corresponding to the target shelf And detecting and confirming whether the weight difference is an integral multiple of the weight of the single piece, and performing settlement according to the weight difference and a multiple of the weight of the single piece.
The cloud server according to claim 28, wherein the settlement module is further configured to detect and confirm that the weight difference is not an integer multiple of the weight of the single piece, and identify the product image corresponding to the target shelf. Obtaining an item above the target shelf, the item above the target shelf includes items of other shelves, obtaining a piece weight of the item of the other shelf, and according to the weight difference, the item list of the other shelf The weight of the piece and the weight of the piece of the target shelf are settled.
The cloud server according to claim 29, wherein the settlement module is further configured to: detect and confirm that the item on the target shelf includes goods of other shelves, and control the vending device to issue a prompt message to The buyer makes a prompt.
The cloud server according to claim 30, wherein the settlement module is further configured to: confirm that the purchaser does not restore items of other shelves above the target shelf, and update a list of products of the target shelf .
The cloud server according to claim 29, wherein the settlement module extracts an item from the product image corresponding to the target shelf, and generates a confidence level of each item, and the item with a confidence greater than a preset threshold The preset product models are matched to obtain the goods on the target shelf.
A cloud server according to claim 32, wherein said predetermined item model is obtained by learning or training.
The cloud server according to claim 29, wherein the settlement module is further configured to: identify that the item on the target shelf includes an item of a non-sale device, and perform the purchase to the purchaser through the vending device Remind and alert the maintenance personnel.