CN115862219A - Intelligent shelf system and its control method - Google Patents

Abstract

The present invention discloses an above-mentioned smart shelf system and its control method, including a plurality of smart shelves that can be placed in different physical locations; wherein, each shelf includes a smart door lock; the unlocking method can connect a specific user mobile device with and will trigger the payment procedure after the user closes the door lock; one or more controlled goods storage devices; shelf controllers; used to obtain information about the product category, price, and quantity of the goods obtained , record and transmit the information to the cloud cargo identifier through the network, and update the product category and quantity information in the cloud cargo identifier at the same time; the cloud cargo identifier uses deep learning technology to train models to identify various products, and synchronizes the Shelf controller to identify new products. Since the present invention adopts the above-mentioned intelligent shelf system and its control method, it can efficiently identify and integrate multiple automatic sales shelves in different locations and/or regions.

Description

Intelligent shelf system and its control method

This application is a divisional application of the Chinese invention patent application with the application number 201910198159.1, the application date is March 15, 2019, and the invention title is "Intelligent Shelf System and Its Control Method".

technical field

The invention belongs to the field of intelligent shelf systems, in particular to an intelligent shelf system for identifying target goods through artificial intelligence and a control method thereof.

Background technique

For a long time, sellers have needed to efficiently arrange vending devices, so as to know the sales situation in time and replenish goods as soon as possible; but for the vending devices in different locations and/different regions, there is a lack of An efficient way to process and analyze the information, and update the data management system in time.

A variety of smart shelves are disclosed in the prior art. For example, the Chinese invention patent application 201711249049.0 discloses the use of sensors to monitor the movement of the shelves, and the monitoring equipment and smart shelves that feed back the product information associated with the target object in the movement track; or in Chinese invention patent application 201810381313.4 discloses that there are several storage boards on the shelf body, and the smart mat is arranged on the storage board, and the smart mat is used to measure the weight of the goods on the smart mat and the contact between the smart mat and the goods area, the control module is electrically connected to the smart mat and a communication module, and the communication module is connected to the cloud server in communication.

The above-mentioned existing technologies basically adopt the method of physically sensing the sale status of goods, which improves the method for the seller to obtain information, thereby reducing the cost. However, the above method still has the defect of being unable to integrate information from multiple locations and multiple regions. Even though the cloud server is mentioned in the above-mentioned second patent application, the role of the cloud server is to receive and record the identification characteristics and location data of the goods, and collect the coordinates of the corresponding smart mat. That is to say, the recognition process is completed through the pressure change of the smart gasket; the function of the cloud server is to store and record the relevant recognition results. Therefore, there is a need for a solution that efficiently identifies and integrates automated sales racks in multiple different physical locations.

Contents of the invention

The object of the present invention is to provide an intelligent shelf system and a control method thereof that can efficiently identify and integrate multiple different locations.

An intelligent shelf system of the present invention includes: a plurality of intelligent shelves that can be placed in different physical locations; wherein, each shelf includes an intelligent door lock; Contact; and will trigger the payment procedure after the user closes the door lock; one or more controlled goods storage devices; shelf controller; used to obtain relevant information such as product category, price, and quantity of the goods obtained, record and pass The network transmits the information to the cloud cargo identifier, and simultaneously updates the product category and quantity information in the cloud cargo identifier; the cloud cargo identifier uses deep learning technology to train the model to identify various products, and synchronizes the shelf controller, to identify new products.

An aspect of the present invention also includes a local goods sensing device; used to obtain information such as the quantity and type of goods increased or decreased. The shelf controller is a dedicated chip for intelligent product identification; the cloud cargo identifier is a cloud AI training engine. Wherein the local goods perception device is a camera, which is used to obtain images and/or videos of goods displayed on the smart shelf, or images and/or videos of goods added or removed from the smart shelf. Wherein said local cargo sensing device is a pressure sensor; used to sense the increase or decrease of cargo. Temperature control devices; and/or lighting devices; and/or touch control panels and advertising displays. The encoding module converts the obtained image and/or video into a binary coded image of the image; the convolution module learns the binary coded image using a convolutional neural network ("CNN"), which is composed of one or more The convolutional layer and the top fully connected layer (corresponding to the classic neural network) also include associated weights and pooling layers; its artificial neurons respond to surrounding units within a part of the coverage; the convolutional neural network Each convolutional layer is composed of several convolutional units, and the parameters of each convolutional unit are optimized through the backpropagation algorithm, and different features of the input product category information are extracted through convolution operations. The product network can iteratively extract complex features from the low-level features obtained in the low-level convolution, so as to obtain the feature map; further adopt the Maxpooling method, and use the same form of nonlinear pooling function to form the feature map The downsampling of the input image is divided into several rectangular areas, and the maximum value is output for each sub-area; the feature map obtained by the dimensionality reduction is generated into a flatten layer (Flatten), that is, the input of the two-dimensional image is one-dimensional to generate a dimensional array, to further realize the transition from the convolutional layer to the fully connected layer (fully connected); the identification module, after the training of the neural network model by the cloud cargo recognizer module, classifies the information in the fully connected layer, thereby identifying a certain 1. Commodity, brand and quantity. It also includes a display and locking module, which displays the amount, type and quantity involved in this goods sales activity on the user's mobile device; The payment method accepted by the above-mentioned intelligent shelf system shall be used to pay the fees required for the sale of goods.

The method for controlling the smart shelf system of the present invention includes the following steps: placing a plurality of smart shelves at different physical locations; wherein each shelf is provided with a smart door lock; and will trigger the payment procedure after the user closes the door lock; set up one or more controlled goods storage devices; set up the shelf controller; obtain relevant information such as the product category, price, and quantity of the goods obtained, record and The information is transmitted to the cloud cargo identifier through the network, and the product category and quantity information in the cloud cargo identifier is updated at the same time; the cloud cargo identifier is set, and the model is trained to identify various products through deep learning technology, and the shelf control is synchronized device to identify new products.

Since the present invention adopts the above-mentioned intelligent shelf system and its control method, it can efficiently identify and integrate multiple automatic sales shelves in different locations and/different regions.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present invention more clearly, the following will briefly introduce the drawings that are used in the embodiments. Apparently, the drawings in the following description are only some examples of the present invention, and those skilled in the art can also obtain other drawings according to these drawings without any innovative work.

Fig. 1 is a schematic diagram of the intelligent shelf system of the present invention.

Fig. 2 is a flow chart of selling goods through the intelligent shelf system of the present invention.

Fig. 3 is a schematic flow chart of controlling the shelf controller of the intelligent shelf system of the present invention.

FIG. 4 is a schematic flow chart of the cloud cargo identifier of the smart shelf system of the present invention.

Fig. 5 is a schematic diagram of identifying a certain brand of instant noodles by the fully connected layer of the present invention.

Fig. 6a is a flow chart of using the intelligent shelf system of the present invention to control inventory and logistics.

6b and 6c are flowcharts of analyzing inventory and logistics big data by using the cloud-based goods identifier of the smart shelf system according to an embodiment of the present invention.

Fig. 6d is a flow chart of using the intelligent shelf system of an embodiment of the present invention to conduct accurate sales by using big data analysis.

FIG. 7 is a schematic diagram of the goods sensing device 104 in an embodiment of the smart shelf system of the present invention.

Fig. 8 is a structural diagram of the shelf controller of the intelligent shelf system of the present invention.

FIG. 9 shows a schematic diagram of a single-depth slice through the pooling means of step 406 in the process shown in FIG. 4 .

Detailed ways

Specific embodiments of the present invention will now be described in conjunction with the corresponding drawings. However, this invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. These embodiments are provided only so that the present invention will be thorough and complete so that those skilled in the art can fully describe the scope of the present invention. Wording used in the detailed description of the embodiments illustrated in the drawings should not limit the invention.

Fig. 1 is a schematic diagram of the intelligent shelf system of the present invention. Wherein, the smart shelf 101 is one or more (not shown in the figure) smart shelves that can be placed in different physical locations. Each smart shelf 101 includes a smart door lock 103 ; one or more controlled goods storage devices 102 ; a goods outlet (not shown in the figure); and a shelf controller 105 . Through the shelf controller 105 , the smart shelf 101 can be controlled by a cloud item identifier (not shown in the figure) included in the cloud 110 . The cloud 110 can be Alibaba Cloud, Tencent Cloud, Amazon AWS, etc., but is not limited thereto. It can be any storage method that can be jointly accessed and updated by terminals located in different physical locations in a network environment. The cloud 110 includes The cloud-based cargo identifier can instantly control and update the shelf controllers in smart shelves located in different physical locations.

Optionally, the smart shelf 101 also includes a local product sensing device 104; and/or a temperature control device; and/or a lighting device; and/or a touch control panel and an advertisement display device. The local product perception device 104 can be one or more cameras, which are used to obtain images and/or videos of the products displayed on the smart shelf, so as to obtain basic information such as the appearance, quantity, size, and variety of the product, thereby determining the quality of the product. Type, model and price to specify a product. The smart shelf can be divided into parts isolated from each other according to needs, and different commodities can be stored according to the different temperature requirements of the commodities. For example, beverages such as red wine and soft drinks that need to be stored at a specific temperature are separated from foods such as biscuits and bread at normal temperature and fast food under warm conditions. Wherein the temperature control device respectively controls different areas of the smart shelf, and can be but not limited to a temperature controller of a freezer, a disinfection cabinet, a warming cabinet or a display cabinet. The lighting device can be one or more various ordinary lights, LED lights, etc. suitable for smart shelves. The touch control panel and/or the advertisement display device may be a two-dimensional or three-dimensional display screen, which is used to display the products in the cabinet or the advertisement information unrelated to the products in the cabinet. In addition to the smart door lock, the touch screen and/or advertising display device can also select goods by touch, or input the unique unlocking password embodied by the personal mobile device to open the smart door lock.

Fig. 2 is a flow chart of selling goods through the intelligent shelf system of the present invention. In step 201 , the user activates the mobile device and unlocks the smart door lock 103 . The unlocking methods include but are not limited to two-dimensional code scanning; identification code input reflecting personal identity; and other methods that can unlock the smart door lock 103, as long as the unlocking method can connect the specific user's mobile device with this goods sales activity. connect. In step 202, the user obtains the desired goods. While obtaining the goods, the shelf controller 105 obtains relevant information such as the product category, price, and quantity of the goods obtained, records and transmits the information to the cloud goods identifier through the network, and simultaneously updates the product in the cloud goods identifier Category and quantity information, optionally updating and displaying the remaining quantity in the user's mobile device. In this step, the shelf controller 105 can perceive the increase or decrease of the goods through the goods sensing device 104, that is, the image or video obtained by the camera, and obtain relevant information such as the category, price, and quantity of the products sold; Or other mechanical means to sense the increase or decrease of goods. In step 203, the amount, type and quantity involved in this goods sales activity are displayed on the user's mobile device; The payment method accepted by the smart shelf pays the fee required for the sale of the goods. The smart door lock 103 described above can be any smart door lock that can be opened by means of, but not limited to, two-dimensional code scanning; personal identification code input, as long as the method can connect the specific user mobile device with the goods sales activity. and can control and open the smart door lock through the mobile device.

Fig. 3 is a schematic flow chart of controlling the shelf controller of the intelligent shelf system of the present invention. Wherein, the smart shelf 101 is one or more (not shown in the figure) smart shelves that can be placed in different physical locations. Wherein each smart shelf 101 may contain one or more local cargo sensing devices 104, which are used to obtain images and/or videos of the goods displayed on the smart shelf, and obtain basic information such as the appearance, quantity, size, and variety of the product, so as to determine The type, model and price of the product in order to specify a product. In step 301 , the local item sensing device 104 sends the obtained image and/or video signal to the shelf controller 105 in the smart shelf 101 . The shelf controller 105 can identify a dedicated chip for the smart product. The shelf controller 105 may be implemented via, but not limited to, a central processing unit (“CPU”), a circular processing unit (“GPU”), a tensor processor (“TPU”), a field programmable gate array (“FPGA”) ), neural network processor ("NPU"), application-specific integrated circuit ("ASIC") eASIC and other general-purpose or special-purpose chips, with the development of AI chip technology and the continuous reduction of costs, the intelligent shelf system of the present invention New balance points will be found constantly between cost and efficiency. In one embodiment, a heterogeneous and multi-core SoC design combining NPU (Neural Network Processor), DPS (Signal Processor) and CPU (Central Processing Unit) can be used to support a smart shelf system with one chip, and has The characteristics of low cost and low power consumption can effectively solve the cost problem of AI chips.

The shelf controller 105 is also controlled by the cloud item identifier 305 in the cloud 110 . The cloud cargo identifier 305 can be a cloud AI training engine, which uses deep learning technology to train models to identify various products, and the training method will be described in detail below. The cloud item identifier 305 may be implemented by, but not limited to, a central processing unit (“CPU”), a circular processing unit (“GPU”), and a tensor processing unit (“TPU”). In step 304, the cloud item identifier 305 will synchronize the shelf controller 105 to identify new products. At step 302 a new product is identified and recorded 303 . At step 306, the shopping list displayed in the user's mobile device 307 is updated with the identified product information.

In another embodiment, the shelf controller 105 can also store a large amount of goods information in advance. Similarly, when new goods are updated to the smart shelf, the cloud goods identifier 305 receives the basic information of the new goods in the cloud. data, identify the new product, and update the information in the control shelf controller 105 synchronously. When the new product is taken away by the user, the shelf controller will directly feed back the type, quantity and price information of the new product to the user's mobile device.

FIG. 4 is a schematic flow chart of the cloud cargo identifier of the smart shelf system of the present invention. The cloud cargo recognizer can train the engine for cloud AI; train the model through deep learning to identify various products. In one embodiment of this case, the cloud AI training engine obtains image and/or video data sets 401 related to products, and roughly divides the data according to product categories, for example, classifying drinking water and alcoholic beverages, and bread and pastries. etc.; divide the image and/or video data belonging to the same product category into a plurality of different sub-sets 402 . As far as each image and/or video information is concerned, an image binary code map (image binary code) 403 is obtained, that is, a digital image 403 is obtained; in step 404, a convolution neural network (convolution neural network) is used for the binary code map 403 , CNN) for learning; CNN is composed of one or more convolutional layers and the top fully connected layer (corresponding to the classic neural network), and also includes associated weights and pooling layers; its artificial neurons Can respond to surrounding units within a portion of the coverage area. Each convolutional layer in the convolutional neural network is composed of several convolutional units, and the parameters of each convolutional unit are optimized through the backpropagation algorithm. The purpose of the convolution operation is to extract different features of the input. The first convolutional layer may only extract some low-level features such as edges, lines, and corners. More layers of networks can iteratively extract more complex features from low-level features. features; thereby obtaining a feature map (feature map) 405. In step 406, a pooling method, especially a max pooling method, is adopted, that is, the maximum value is extracted from the corrected feature map as the pooled value of the region. Divide the input image into several rectangular regions, and output the maximum value for each subregion. Intuitively, this mechanism works because, after a feature is discovered, its precise location is far less important than its relative location to other features. The pooling layer will continuously reduce the space size of the data, so the number of parameters and the amount of calculation will also decrease, which also controls overfitting to a certain extent. Generally speaking, pooling layers are periodically inserted between the convolutional layers of CNN. Pooling layers typically act on each input feature separately and reduce its size. The most commonly used form of pooling layer is to divide the image into 2×2 blocks every 2 elements, and then take the maximum value of the 4 numbers in each block. This will reduce data volume by 75%.

A single-depth slice as in the example in Figure 9.

After the maximum pooling step 406, the dimensionality-reduced feature map 407 is obtained. Since the dimensionality-reduction feature map 407 is still a two-dimensional picture, in step 408, a flattening layer (Flatten) is generated, that is, the input of the two-dimensional picture is one-dimensional to generate a one-dimensional array 409 for the transition from a convolutional layer to a fully connected layer. The flattening operation on the two-dimensional image does not affect the size of the batch (Batch), where the batch is used to better handle non-convex loss functions; and make reasonable use of memory capacity.

In step 410, train a fully connected feedforward neural network (full connected feedforward network). In the example shown in FIG. (Neuron) accepts a large number of nonlinear input information, the input information is called input vector. The second layer 412, the third layer 413, and the fourth layer 414 are hidden layers (Hidden layer), referred to as "hidden layers", which are various layers composed of many neurons and links between the input layer and the output layer; The hidden layer, for example, in the case of three layers, means multiple activation functions; the fifth layer 415 is the output layer (Output layer), information is transmitted, analyzed, and weighed in neuron links to form an output result, and the output message is called output vector.

In order to distinguish a certain commodity from other commodities, the neural network model is trained by the cloud AI training engine, and the information is classified in the fully connected layer to identify a certain commodity and brand. Fig. 5 is the schematic diagram that the fully connected layer of the present invention recognizes a certain brand of instant noodles; Red/yellow/green dots (in Fig. 5 represent different colors with different points of depth) represent that this neuron is found, namely activated; the same layer Other unhighlighted neurons in , indicating that either the features of instant noodles are not contained in the involved one-dimensional array, or the features of instant noodles are not obvious. Then output the final result, judge the type and brand of the product, that is, determine it as a certain brand of instant noodles.

The above description is only one implementation of the cloud-based goods identification, and other known or unknown deep learning technologies can be used, as long as the technology can train the shelf controller of the smart shelf system from the cloud, so as to monitor and understand the sales of goods in real time The situation, as well as basic information such as the type, quantity and brand of goods that need to be supplemented.

Fig. 6a is a flow chart of controlling inventory and logistics using the smart shelf system according to an embodiment of the present invention. In step 601, the goods are placed on the shelves. The shelves in the drawings are open shelves, and the payment system has nothing to do with the lock of the shelf itself. The payment system can be set at the exit of the store; the intelligent shelf system described in the present invention includes but It is not limited to an open shelf, and may be a closed smart shelf; in step 602, sensors are set on the smart shelf to determine the quantity and variety of the goods. In addition to the aforementioned pressure sensor in the manual, the sensor can be an infrared sensor, a volumetric Displacement sensors, light curtain sensors, etc., among which infrared sensors can distinguish between the user's hand and the goods; volumetric displacement sensors can be used to distinguish the size and mechanical displacement of the goods, so as to perceive the state of the goods. The light curtain sensor emits modulated infrared light through the light emitter, which is received by the light receiver to form a protective net. When an object enters the protective net, and when there is light blocked by the object, the light receiver circuit responds immediately through the internal control circuit. . These sensors know the situation that the goods are taken away by sensing the stock of the goods; in step 603, the customer takes away the goods from the shelf; 605 , the cloud server processes the information that the goods have been taken away, and sends the information of lack of goods inventory to the display device at step 606 to warn the display device that the shelf inventory is low; and at step 607 the service personnel replenish the goods.

6b and 6c are flowcharts of analyzing inventory and logistics big data by using the cloud-based goods identifier of the smart shelf system according to an embodiment of the present invention. The intelligent shelf system of the present invention performs integrated management on commodities and supply chains. Specifically, in step 610 and step 620, the goods and the channels for supplying goods are respectively optimized, and the step 610 of optimizing goods includes step 611 of optimizing the combination of goods; In step 612, optimize the price of the goods; and in step 613, optimize the layout and design of the goods. Wherein, in step 611, optimizing the combination of goods includes classifying the goods centered on customers, for example, into products suitable for women, products suitable for children, etc.; and classifying all channels, such as into Various sources of goods; and classify the localization of goods, such as identifying and obtaining local goods; in step 612, factors for optimizing prices include but are not limited to quantitative analysis of customer sentiment; Dynamically adjust pricing; provide budget and predict customer behavior; conduct promotion analysis; in step 613, the step of optimizing the layout and design of goods includes the analysis of product selection; and the analysis of product layout on the shelf, etc. The step 620 of optimizing the supply channel includes step 621 optimizing the goods in stock; step 622 optimizing the distribution and logistics of the goods; and step 623 optimizing the management of the storage space of the goods. Wherein, in step 621, the step of optimizing the inventory of the goods includes analyzing the storage status of the goods, and giving a warning to the inventory level lower than the predetermined level; predicting the customer's demand according to the inventory situation; and managing the inventory cost , in order to obtain the optimal solution between stock and inventory cost; in step 622, the steps of distribution and logistics optimization of goods include analyzing the performance of different suppliers, managing identity information of different suppliers and for Manage the delivery status of each supplier to obtain the optimal supplier and logistics arrangement; in step 623, the management of the goods storage space includes the selection of stored goods and the analysis of customer purchase patterns to obtain the best storage.

Fig. 6d is a flow chart of using the intelligent shelf system of an embodiment of the present invention to conduct accurate sales by using big data analysis. The purpose of using the smart shelf system according to the embodiment of the present invention to conduct big data analysis to conduct precise sales is to obtain a smart shopping experience. In step 631, intelligently analyze the customer, including identifying the identity of the customer in step 632; and analyzing the behavior of the customer in step 635. In step 634, carry out omni-channel marketing for goods, which includes performing marketing based on customer location in step 633; conduct precise marketing to customers in step 636, and improve customer experience in omni-channel in step 637. Specifically, in the customer identification step of step 632, omni-channel customer information is obtained; customer active cycle analysis; multi-dimensional segmentation of customers, for example, focusing on multiple dimensions of segmentation, then after applying the post-event segmentation model , the model will label each sample or customer, so that the gender difference, age difference, income difference, etc. of the customer can be seen through this label, and the target customer can be quickly found; the customer behavior identification step in step 635 includes The customer's cross-channel behavior, cross-shopping and customer sentiment are analyzed respectively; in the step 633 of marketing based on the customer's location, the customer's scene, real-time behavior and location are analyzed respectively; and in step 636, according to the goods and Relationship between customers, customer mining and behavior prediction analysis and marketing effect analysis to obtain accurate sales to customers, for example, according to customers' preferences for lipstick brand colors and customer analysis of fixed age groups and fixed genders, to obtain sales for a certain age group Precise sales information for customers in a certain occupation or income range; in step 637, improve the customer's omni-channel experience based on the analysis of the customer's online and offline behavior, personalized service and channel process.

FIG. 7 is a schematic diagram of the goods sensing device 104 in an embodiment of the smart shelf system of the present invention. The cargo sensing device 104 may include a camera and/or a storage unit, the storage unit may be built into the cargo sensing device or external to the cargo sensing device, and the cameras may be one or more, which are suitable for photographing The image or video information of the product appearance, the image or video information may be two-dimensional, but not limited to two-dimensional.

Fig. 8 is a structural diagram of the shelf controller of the intelligent shelf system of the present invention. Such as shelf controller server 801. The shelf controller server includes a shelf controller processor 802 , where the processor may be a general-purpose or special-purpose chip as described above, and a computer program product in the form of a memory 803 or a computer-readable medium. Memory 803 may be electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), EPROM, hard disk, or ROM. The memory 803 has a storage space 804 for program code 805 for performing any method steps in the methods described above. For example, the storage space 804 for program codes may include respective program codes 805 for respectively implementing various steps in the above methods. These program codes can be read by the cloud AI training engine or written into the shelf controller processor. The program code can eg be compressed in a suitable form. These codes, when executed by the server, cause the server to perform the steps of the methods described above.

Reference herein to "one embodiment," "an embodiment," or "one or more embodiments" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Additionally, please note that examples of the word "in one embodiment" herein do not necessarily all refer to the same embodiment.

The above description is only used to illustrate the technical solutions of the present invention, and any person skilled in the art can modify and change the above embodiments without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention should be determined by the scope of the claims. The present invention has been described above with reference to examples. However, other embodiments than those described above are equally possible within the disclosed scope of the present invention. The different features and steps of the invention may be combined in other ways than described. The scope of the present invention is limited only by the appended claims. More generally, one of ordinary skill in the art can readily understand that all parameters, dimensions, materials and configurations described herein are for exemplary purposes and actual parameters, dimensions, materials and/or configurations will depend on the particular application or invention Teach the application for which it is used.

Claims (8)

1. A device for controlling goods through an intelligent shelf system, The smart shelf system includes a plurality of smart shelves that can be placed in different physical locations; wherein each shelf includes a smart door lock; the smart door lock is unlocked in a manner that can associate a specific user mobile device with a sale of goods; and The payment procedure will be triggered after the user closes the door lock; one or more controlled goods storage devices; shelf controller; used to obtain relevant information such as product category, price, and quantity of the goods obtained, record and send all the information through the network The above information is transmitted to the cloud cargo identifier, and the product category and quantity information in the cloud cargo identifier is updated at the same time; the cloud cargo identifier uses deep learning technology to train the model to identify various products, and synchronizes the shelf controller to identify new products. product; said cargo control device includes: The payment module placed outside the smart shelf is used to provide the price to the user and record whether the payment is successful; The perception module is used to determine the quantity and variety of the goods and whether the goods are in the state of being taken away; A processing module, which transmits the information obtained by the sensing module to the cloud server, and processes the information that the goods are taken away; The warning module is used to display information on the lack of stock on the display device. 2. The control device according to claim 1, wherein the sensing module includes a pressure sensor or an infrared sensor or a volume displacement sensor or a light curtain sensor; and the smart shelf can be an open or closed smart shelf. 3. A method for controlling goods through an intelligent shelf system, comprising: Multiple smart shelves are placed in different physical locations; each shelf is equipped with a smart door lock; the unlocking method can link a specific user's mobile device with the sales activity of the goods; and the payment procedure will be triggered after the user closes the door lock ; Set up one or more controlled goods storage devices; Set up shelf controllers; Obtain relevant information such as the product category, price, and quantity of the goods obtained, record and transmit the information to the cloud goods identifier through the network, and update it at the same time The product category and quantity information in the cloud cargo identifier; the cloud cargo identifier is set, the model is trained to identify various products through deep learning technology, and the shelf controller is synchronized to identify new products; the cargo control method also includes: Set up the payment system outside the smart shelf; Perceive and determine the quantity and variety of goods and whether the goods are in the state of being picked up; Transmit the perceived information to the cloud server and process the information that the goods are taken away; Information about the lack of goods is transmitted to the display device. 4. The control method as claimed in claim 3, wherein, realizing the perception and determining the quantity and type of the goods and whether the goods are in the state of being taken away by a pressure sensor or an infrared sensor or a volumetric displacement sensor or a light curtain sensor steps; and the smart shelf can be an open or closed smart shelf. 5. A device for analyzing cargo inventory and logistics through the cloud cargo identifier in the smart shelf system, The smart shelf system includes a plurality of smart shelves that can be placed in different physical locations; wherein each shelf includes a smart door lock; the smart door lock is unlocked in a manner that can associate a specific user mobile device with a sale of goods; and The payment procedure will be triggered after the user closes the door lock; one or more controlled goods storage devices; shelf controller; used to obtain relevant information such as product category, price, and quantity of the goods obtained, record and send all the information through the network The above information is transmitted to the cloud cargo identifier, and the product category and quantity information in the cloud cargo identifier is updated at the same time; the cloud cargo identifier uses deep learning technology to train the model to identify various products, and synchronizes the shelf controller to identify new products. product; said analytical device comprising: Cargo optimization module and supply channel optimization module; where The cargo optimization module includes: Goods combination optimization module, which is used to classify goods centered on customers, and classify the channel sources of goods, so as to identify and obtain local products suitable for different customer groups; The goods price optimization module is used to quantitatively analyze customer sentiment, dynamically adjust pricing according to purchase and sale conditions, provide budgets to predict customer behavior, and conduct promotional analysis; Goods layout design optimization module, used for analysis of product selection and product layout on shelves; The supply channel optimization module includes: Inventory goods optimization module, which is used to analyze the storage status of goods, and give a warning to the inventory level below the predetermined level; predict the customer's demand according to the inventory situation; and manage the inventory cost, so that the inventory and inventory expenditure to obtain the optimal solution; The goods distribution and logistics optimization module is used to analyze the performance of different suppliers, manage the identity information of different suppliers and manage the delivery status of each supplier, so as to obtain the optimal supplier and logistics arrangement; and The cargo storage space management module is used for the management of the cargo storage space, including the selection of stored commodities and the analysis of customer purchase patterns, so as to obtain the best storage space. 6. A method for analyzing goods inventory and logistics through a cloud goods identifier in a smart shelf system, Multiple smart shelves are placed in different physical locations; each shelf is equipped with a smart door lock; the unlocking method can link a specific user's mobile device with the sales activity of the goods; and the payment procedure will be triggered after the user closes the door lock ; Set up one or more controlled goods storage devices; Set up shelf controllers; Obtain relevant information such as the product category, price, and quantity of the goods obtained, record and transmit the information to the cloud goods identifier through the network, and update it at the same time Product category and quantity information in the cloud cargo identifier; set the cloud cargo identifier, train the model to identify various products through deep learning technology, and synchronize the shelf controller to identify new products; the analysis method includes: Optimizing goods and supply channels; Wherein said cargo optimization steps include: Optimize the mix of goods for customer-centric classification of goods, and classify the channel sources of goods to identify and obtain local products suitable for different customer groups; Optimize the price of goods, use it to quantitatively analyze customer sentiment, and dynamically adjust pricing according to the purchase and sale situation, provide budgets to predict customer behavior, and conduct promotional analysis; Optimize the layout design of goods for the analysis of product selection and product layout on the shelf; The supply channel optimization steps include: Optimizing inventory goods for analysis of the storage status of goods and warnings for inventory levels below the predetermined level; forecasting customer demand based on inventory conditions; and managing inventory costs so that inventory and inventory Get the best solution between costs; Optimize the distribution and logistics of goods to analyze the performance of different suppliers, manage the identity information of different suppliers and manage the delivery status of each supplier, so as to obtain the optimal supplier and logistics arrangement; as well as To manage the storage space of the goods, the management of the storage space of the goods includes the selection of the stored goods and the analysis of the customer's purchase pattern, so as to obtain the best storage space. 7. A device for precise sales through a smart shelf system, The smart shelf system includes a plurality of smart shelves that can be placed in different physical locations; wherein each shelf includes a smart door lock; the smart door lock is unlocked in a manner that can associate a specific user mobile device with a sale of goods; and The payment procedure will be triggered after the user closes the door lock; one or more controlled goods storage devices; shelf controller; used to obtain relevant information such as product category, price, and quantity of the goods obtained, record and send all the information through the network The above information is transmitted to the cloud cargo identifier, and the product category and quantity information in the cloud cargo identifier is updated at the same time; the cloud cargo identifier uses deep learning technology to train the model to identify various products, and synchronizes the shelf controller to identify new products. Products; said sales device includes: The intelligent analysis module obtains omni-channel customer information; analyzes customer active cycles; conducts multi-dimensional segmentation of customers; compiles category labels for customers to identify customers; and analyzes customer cross-channel behavior, cross-shopping and customer sentiment Analyze individually in order to obtain customer behavior patterns; Omni-channel marketing module, by analyzing the customer's scene, real-time behavior and location, so as to sell according to the customer's location; and, based on the relationship between goods and customers, the analysis of customer mining and behavior prediction and marketing effect analysis , in order to obtain accurate sales to customers; and based on the analysis of customers' online and offline behaviors, personalized services and channel processes, in order to improve customers' omni-channel experience. 8. A method of precise sales through a smart shelf system, Multiple smart shelves are placed in different physical locations; each shelf is equipped with a smart door lock; the unlocking method can link a specific user's mobile device with the sales activity of the goods; and the payment procedure will be triggered after the user closes the door lock ; Set up one or more controlled goods storage devices; Set up shelf controllers; Obtain relevant information such as product categories, prices, and quantities of the goods obtained, record and transmit the information to the cloud goods identifier through the network, and update at the same time Product category and quantity information in the cloud goods recognizer; set the cloud goods recognizer, use deep learning technology to train the model to identify various products, and synchronize the shelf controller to identify new products; the sales method includes: Intelligent analysis of customers; including obtaining omni-channel customer information; customer active cycle analysis; multi-dimensional segmentation of customers; and compilation of category tags for customers to identify customer identities; and through customer cross-channel behavior, cross-shopping and Customer sentiment is analyzed separately to obtain customer behavior patterns; and Conduct omni-channel sales to customers, including analyzing the customer's scene, real-time behavior and location, so as to sell according to the customer's location; and, based on the relationship between goods and customers, the analysis of customer mining and behavior prediction and Marketing effect analysis in order to obtain accurate sales to customers; and based on the analysis of customers' online and offline behaviors, personalized services and channel processes, in order to improve customers' omni-channel experience.

Images