Automatic replenishment method and smart shelf based on the number of times the product is picked up

Abstract

The present invention relates to the technical field of smart shelves, and discloses an automatic replenishment method based on the number of times a product is picked up and a smart shelf, the method steps comprising: obtaining a surveillance image of a pickup area; identifying customers in the surveillance image and the types and quantities of corresponding products; recording the customer's stay time and the number of times each product is picked up in the current replenishment cycle, and automatically replenishing the products based on the number of times each product is picked up and the corresponding replenishment threshold; updating the total number of times a product is picked up, the first stay time, the second stay time, and the purchase conversion rate; updating the stocking threshold of the product; if the inventory of the product is less than the stocking threshold, sending a stocking reminder to the manager. The present invention realizes dynamic control and automatic replenishment of product inventory by combining Internet of Things technology, target detection algorithm and data analysis, improves replenishment efficiency and optimizes product inventory.

Description

Technical Field

[0001] The present invention relates to the technical field of smart shelves, and in particular to an automatic replenishment method based on the number of times a commodity is picked up and a smart shelf. Background Art

[0002] The retail industry currently relies mainly on traditional experience judgment, sales data analysis, and manual inspections for product replenishment. These methods have many shortcomings and are difficult to meet the needs of modern retail business for intelligence and refinement. Many retailers rely on personal experience to make replenishment decisions. Although this method takes into account factors such as seasonal changes and holiday effects to a certain extent, it often lacks accurate data support and is easily affected by personal subjective judgment, resulting in untimely replenishment or inventory backlogs, which in turn affects sales efficiency and customer satisfaction. In large supermarkets or e-commerce warehouses, due to the wide variety of goods and the wide distribution of shelves, manual inspections are not only time-consuming and laborious, but also difficult to achieve real-time monitoring and accurate statistics. They are also easily affected by factors such as subjective judgment and work fatigue of inspectors, resulting in the accuracy and timeliness of replenishment decisions being affected.

[0003] Existing replenishment systems have limitations in data collection. On the one hand, many systems only focus on the sales data of goods, but ignore the behavioral data of customers during the shopping process, such as the number of times the goods are picked up and the length of time customers stay. On the other hand, due to technical limitations and cost considerations, some systems are unable to achieve real-time data collection and processing, resulting in a lack of timeliness and accuracy in replenishment decisions. Existing replenishment systems also often lack a dynamic adjustment mechanism and are unable to adjust replenishment strategies in real time based on market changes, customer demand, and product sales. This static replenishment strategy is difficult to adapt to the rapidly changing market environment and can easily lead to inventory backlogs or out-of-stock problems.

[0004] In the prior art, the replenishment of goods is generally regular replenishment or targeted replenishment when the goods are out of stock. Regular replenishment is prone to require secondary replenishment due to seasonal factors, or the remaining quantity of each product needs to be counted in advance. For smart shelves, the remaining quantity of goods needs to be counted multiple times. Multiple counts may require manual inventory, resulting in increased costs. Although it is accurate to update the data based on each checkout data, it does not have forward-looking (i.e., prediction function). At present, if the deep learning model is used in this scenario, on the one hand, the data dimension is small and the model cannot converge. On the other hand, the training cost of the deep learning model is high, and shopping also has a regional bias, resulting in the need to customize a deep learning model for each region or even each supermarket, which has great disadvantages.

[0005] For example, a Chinese patent with authorization announcement number CN109741519B discloses an unmanned supermarket shelf monitoring system and a control method thereof, the method comprising: after the camera is turned on, the camera executes a preset camera program to take photos and upload the photos taken to a background server connected to the camera; the background server stores the photos taken by the camera; the background server performs image recognition on the latest stored photo, and when it is recognized that the number of items on the shelf is less than the preset number, the background server sends a replenishment reminder message to a management terminal connected to the background server; the camera automatically shuts down after verifying that the uploaded photo is successful; an RTC alarm built into the camera re-awakens the camera after a preset standby time interval, and the unmanned supermarket shelf monitoring system repeats the above method steps in the above order; the invention realizes ultra-low power consumption shelf item quantity monitoring, and has a replenishment reminder function, with the advantages of low power consumption, accurate recognition, and convenient installation.

[0006] For example, the Chinese patent application with publication number CN113487262A discloses an intelligent shelf material management system and method based on image recognition, which mainly includes the following components: shelves, aisles, aisle labels, industrial cameras and industrial computers. The shelves are distributed with multiple aisles; the aisle labels contain the location information and material information of the aisles. The industrial cameras are installed on the shelves. The industrial computers perform image recognition analysis on the photos taken by the industrial cameras, determine the material inventory status, update the local inventory value, and report the information to the background system. Its beneficial effect is that the invention can update the aisle material inventory value according to the detection status through the industrial camera, combined with the image recognition algorithm, and report the information to the supplier, prompting the supplier to replenish the stock in time and eliminate the risk of material shortage; the invention can accurately and automatically detect the inventory status of shelf materials, accurately and automatically identify label information, reduce material management costs, and improve the efficiency of material inventory counting.

[0007] The above patents all have the problems raised by this background technology: they ignore the behavioral data of customers during the shopping process, such as the number of times the goods are picked up, the customer's stay time, etc.; the replenishment system lacks a dynamic adjustment mechanism.

[0008] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the invention and should not be regarded as an acknowledgement or any form of suggestion that the information constitutes the prior art already known to ordinary technicians in this field. Summary of the invention

[0009] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide an automatic replenishment method and a smart shelf based on the number of times a product is picked up. On the one hand, it does not require the configuration of complex devices, and on the other hand, it can be combined with monitoring equipment for collaboration. It also has a certain prediction function, and the realization of the prediction function can be decoupled from the region and supermarket type. The same model can be used in each supermarket and each scene. At the same time, it can avoid the problem of model failure to converge due to small data dimension.

[0010] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0011] In one aspect, the present invention provides an automatic replenishment method based on the number of times a product is picked up, comprising the following steps:

[0012] S1: Monitor the shelves and goods through surveillance video, and when it is detected that the goods are picked up, capture the surveillance image of the picking area; the picking area is the shelf area corresponding to the picked goods;

[0013] S2: performing target detection on the surveillance image of the pickup area to identify the customers in the surveillance image and the types and quantities of the corresponding commodities;

[0014] S3: Continuously track and monitor the pickup area, record the customer's stay time and the number of times each product is picked up in the current replenishment cycle; automatically replenish the products based on the number of times each product is picked up and the corresponding replenishment threshold;

[0015] S4: Update the total number of times the product is picked up and record the customer's unpurchased stay time and purchased stay time;

[0016] S5: updating the first stay time of the product based on the unpurchased stay time, and updating the second stay time of the product based on the purchased stay time;

[0017] S6: Calculate the purchase conversion rate of the product, and update the stocking threshold of the product based on the purchase conversion rate and the first residence time and the second residence time; if the inventory of any product is less than the corresponding stocking threshold, send a stocking reminder to the management personnel.

[0018] As a preferred solution of the automatic replenishment method based on the number of times the product is picked up according to the present invention, the method of continuously tracking and monitoring the picking area and recording the customer's residence time is as follows: continuously capturing surveillance images from surveillance video, and detecting and tracking the customer and the product picked up by the customer in each frame of the surveillance image through a multi-target tracking algorithm; the residence time is recorded as follows: starting from the time when the product is detected to be picked up, until it is detected that the product is put back on the shelf or taken away, then stopping the timing, and the accumulated timing duration is the customer's residence time.

[0019] As a preferred solution of the automatic replenishment method based on the number of times the product is picked up, the method for determining whether the product is put back on the shelf is as follows: the range of the picking area is marked in the monitoring image; if the customer leaves the picking area and the product does not leave the picking area, the product is put back on the shelf; if the product is put back on the shelf, the specific time when the timing is stopped is the time when the customer leaves the picking area;

[0020] The method for determining whether the goods are taken away is as follows: if both the customer and the goods leave the pickup area, the goods are taken away; if the goods are taken away, the specific time to stop timing is the time when both the customer and the goods leave the pickup area.

[0021] As a preferred solution of the automatic replenishment method based on the number of times a commodity is picked up in the present invention, each commodity is preset with a corresponding replenishment threshold; the method for automatically replenishing commodities based on the number of times each commodity is picked up and the corresponding replenishment threshold is as follows: if the redundancy of any commodity in the current replenishment cycle is less than the corresponding replenishment threshold, the commodity is replenished; the redundancy is the cumulative replenishment quantity of the commodity in all replenishment cycles minus the cumulative sales of the corresponding commodity in all replenishment cycles; in the current replenishment cycle, if the number of times any commodity is picked up is not less than the preset pick-up threshold, the replenishment threshold is adjusted; the replenishment threshold of any commodity is adjusted based on the corresponding pick-up number, and the formula is as follows:

[0022] ;

[0023] Where D represents the adjusted replenishment threshold of any commodity; Indicates the initial value of the replenishment threshold of the corresponding product; Indicates the maximum value of the replenishment threshold of the corresponding product; n is the number of times the corresponding product is picked up in the current replenishment cycle; is the adjustment factor.

[0024] As a preferred solution of the automatic replenishment method based on the number of times a product is picked up, when any replenishment cycle begins, the initial value of the replenishment threshold based on the number of times any product is picked up is To assign values, the formula is as follows:

[0025] ;

[0026] in, Indicates the replenishment threshold of the corresponding product at the end of the previous replenishment cycle; Indicates the number of times the corresponding product was picked up in the previous replenishment cycle. express The number of times the corresponding product was picked up in the previous replenishment cycle of the corresponding replenishment cycle; Represents the weight coefficient.

[0027] As a preferred solution of the automatic replenishment method based on the number of times a product is picked up, the total number of times a product is picked up is the sum of the number of times the product is picked up in all replenishment cycles;

[0028] The calculation method of the first residence time is as follows: for any kind of commodity, the customer's residence time is recorded when it is detected that the commodity is picked up; if the commodity is put back on the shelf, the recorded customer's residence time is marked as unpurchased residence time; all unpurchased residence times of the corresponding commodity are queried and the average is calculated to obtain the first residence time of the corresponding commodity.

[0029] As a preferred solution of the automatic replenishment method based on the number of times the product is picked up as described in the present invention, the calculation method of the second residence time is as follows: for any product, the customer's residence time is recorded when it is detected that the product is picked up; if the product is taken away, the recorded customer's residence time is marked as the purchased residence time; all purchased residence times of the corresponding product are queried and the average is calculated to obtain the second residence time of the corresponding product.

[0030] As a preferred solution of the automatic replenishment method based on the number of times a product is picked up according to the present invention, the calculation formula of the purchase conversion rate is as follows:

[0031] ;

[0032] Among them, C represents the purchase conversion rate of any product; Indicates the sales volume of the corresponding product; Indicates the total number of times the corresponding product is picked up.

[0033] As a preferred solution of the automatic replenishment method based on the number of times a product is picked up according to the present invention, the formula for updating the stock threshold of the product is as follows:

[0034] ;

[0035] Among them, B represents the updated stocking threshold of the product; Indicates the first dwell time of the corresponding product; represents the second dwelling time of the corresponding product; C represents the purchase conversion rate of the corresponding product; represents the adjustment coefficient of the first residence time, k represents the adjustment coefficient of the second residence time, represents the adjustment coefficient of purchase conversion rate, represents the reference value of the first dwell time, represents the reference value of the second dwell time, Indicates the benchmark value of purchase conversion rate. Indicates the reference stocking threshold of the corresponding product.

[0036] In a second aspect, the present invention provides an automatic replenishment smart shelf based on the number of times a product is picked up, including a shelf body, a camera module, a database, a processing module, and a communication module; wherein:

[0037] The shelf body includes M grids, where M is a positive integer; each grid is a shelf area, and each shelf area is used to place a specified type of goods;

[0038] The camera module includes surveillance cameras installed above and above the side of each shelf area, and is used to capture surveillance images of the corresponding shelf area;

[0039] The database is used to store the commodity information of each commodity, including the inventory quantity, sales quantity, number of pick-ups, total number of pick-ups, unpurchased stay time, and purchased stay time of each commodity;

[0040] The processing module is configured with a target detection algorithm and a multi-target tracking algorithm, which are used to detect and track customers and the goods picked up by customers from the surveillance images; the processing module is also configured with a replenishment threshold for each product, and automatically replenishes the products based on the number of times each product is picked up and the corresponding replenishment threshold in each replenishment cycle, and generates a replenishment list;

[0041] The processing module is also configured with a stocking threshold for each product and an update formula for the stocking threshold. The stocking threshold for each product is updated by calculating the purchase conversion rate, the first dwell time, and the second dwell time. If the inventory of any product is less than the corresponding stocking threshold, the processing module generates a stocking reminder.

[0042] The communication module is used to send the replenishment list and the stocking reminder to the management personnel.

[0043] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0044] The system automatically detects when items are picked up through surveillance video and triggers a replenishment request when necessary, which reduces manual intervention, improves replenishment efficiency, ensures that shelves are adequately stocked, reduces the waiting time for customers due to out-of-stock situations, and improves the shopping experience. It also reduces the need for manual replenishment and reduces labor costs.

[0045] Based on the customer's stay time and purchase conversion rate, the stocking threshold is dynamically adjusted to make inventory management more accurate and reduce inventory backlogs and out-of-stock phenomena. Through data analysis, merchants can gain an in-depth understanding of product sales and provide data support for inventory management and product procurement. Through accurate replenishment and inventory management, the loss of expired and damaged products is reduced, and the utilization rate of products is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0047] Figure 1 A flow chart of the automatic replenishment method based on the number of times a product is picked up provided by the present invention;

[0048] Figure 2 A schematic diagram of the functional modules of the automatic replenishment smart shelf based on the number of times the product is picked up provided by the present invention;

[0049] Figure 3 Schematic diagram of application scenario of the automatic replenishment smart shelf based on the number of times the goods are picked up provided by the present invention. DETAILED DESCRIPTION

[0050] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. The embodiments of the present invention and the technical features in the embodiments may be combined with each other unless there is a conflict.

[0051] Example 1

[0052] This embodiment introduces an automatic replenishment method based on the number of times a product is picked up. Figure 1 , the method comprises the following steps:

[0053] S1: Monitor the shelves and goods through surveillance video, and when it is detected that the goods are picked up, capture the surveillance image of the picking area; the picking area is the shelf area corresponding to the picked goods;

[0054] The camera continuously captures the video stream of the shelf; the monitoring center uses the state of the shelf when no product is picked up as the background image. When products appear on the background, the system can detect these changes. By comparing consecutive frames, the monitoring center can track the movement of the product to determine whether the product has been picked up. In addition, technologies such as convolutional neural networks in deep learning can be used to classify and identify products to determine which specific product has been picked up.

[0055] S2: performing target detection on the surveillance image of the pickup area to identify the customers in the surveillance image and the types and quantities of the corresponding commodities;

[0056] Each product is placed in its own shelf area; cameras are installed on the top or side of the shelf to capture surveillance images of the shelf area. When it is detected that a product is picked up, the surveillance video shot by the camera at the corresponding position is retrieved to capture the surveillance image of the corresponding shelf area.

[0057] S3: Continuously track and monitor the pickup area, record the customer's stay time and the number of times each product is picked up in the current replenishment cycle; automatically replenish the products based on the number of times each product is picked up and the corresponding replenishment threshold;

[0058] Perform target detection on the surveillance images of the pickup area, for example, using target detection algorithms such as YOLO, Faster R-CNN, SSD, etc., to identify and mark the customers in the surveillance images of the pickup area and the types and quantities of goods picked up by the customers, so as to facilitate subsequent processing and analysis.

[0059] Each commodity is preset with a corresponding replenishment threshold; the replenishment threshold of each commodity is different, and the initial value of the replenishment threshold is set based on the historical demand of each commodity; the method of automatically replenishing commodities based on the number of times each commodity is picked up and the corresponding replenishment threshold is as follows:

[0060] If the redundancy of any commodity in the current replenishment cycle is less than the corresponding replenishment threshold, the commodity will be replenished; the redundancy is the cumulative replenishment quantity of the commodity in all replenishment cycles minus the cumulative sales of the corresponding commodity in all replenishment cycles; when the redundancy of the commodity is not less than the corresponding replenishment threshold, it means that the commodities on the smart shelf can meet the sales demand within a certain period of time, and replenishment is not required temporarily; when the redundancy of the commodity is less than the corresponding replenishment threshold, in order to prevent the commodity from failing to meet the sales demand in time, the smart shelf needs to be replenished with the commodity; when any commodity is picked up more times in the replenishment cycle, it means that the commodity has received more attention during this period of time, and there may be potential purchase opportunities. Therefore, in the current replenishment cycle, if the number of times any commodity is picked up is not less than the preset pick-up threshold, the adjustment of the replenishment threshold is triggered; the replenishment threshold of any commodity is adjusted based on the corresponding pick-up number, and the formula is as follows:

[0061] ;

[0062] Where D represents the adjusted replenishment threshold of any commodity; Indicates the initial value of the replenishment threshold of the corresponding product in the current replenishment cycle; Indicates the maximum value of the replenishment threshold of the corresponding product; n is the number of times the corresponding product is picked up in the current replenishment cycle; To adjust the coefficient, the technical personnel in this field shall determine the value based on actual needs;

[0063] In the above replenishment threshold adjustment formula, is a constant greater than 0, which determines the rate at which the replenishment threshold increases. As the number of pick-ups n in the current replenishment cycle increases, the replenishment threshold of the corresponding product gets closer and closer to the maximum value. , and the value range of the replenishment threshold D is limited to the preset and between.

[0064] When any replenishment cycle begins, the initial value of the replenishment threshold based on the number of times any product is picked up To assign values, the formula is as follows:

[0065] ;

[0066] in, Indicates the replenishment threshold of the corresponding product at the end of the previous replenishment cycle; Indicates the number of times the corresponding product was picked up in the previous replenishment cycle. express The number of times the corresponding product was picked up in the previous replenishment cycle of the corresponding replenishment cycle; Represents the weight coefficient, which is assigned by technicians in this field based on actual needs;

[0067] The formula for assigning the initial value of the replenishment threshold above is to adjust the initial value of the replenishment threshold according to the change in the number of times the product is picked up in two adjacent replenishment cycles. represents the difference in the number of pick-ups between two consecutive replenishment cycles, which reflects the change in the degree of attention (pick-ups) of the product. , indicating that the product received more attention from customers in the previous replenishment cycle than in the previous cycle, and the number of pick-ups increased; on the contrary, if , it means that the product’s popularity has decreased. This part is a proportional adjustment factor. When the number of pick-ups increases, this factor is greater than 1, which will make the new replenishment threshold initial value greater than , thereby increasing the replenishment volume; when the number of pick-ups decreases, this factor is less than 1, and the replenishment volume is reduced. By considering the changes in the number of pick-ups in different replenishment cycles, the replenishment threshold can be adjusted in real time according to the popularity of the product. This can make the replenishment strategy more flexible and better adapt to the dynamic changes in market demand. For example, if a product suddenly becomes popular and the number of pick-ups increases significantly, the replenishment threshold will be increased accordingly, thereby ensuring that there is enough supply of goods on the shelves, reducing the occurrence of out-of-stock situations, and improving customer satisfaction.

[0068] S4: Update the total number of times the product is picked up and record the customer's unpurchased stay time and purchased stay time;

[0069] The method of continuously tracking and monitoring the pickup area and recording the customer's residence time is as follows: continuously capture surveillance images from surveillance video, and detect and track customers and the goods picked up by customers in each frame of surveillance images through a multi-target tracking algorithm; for example, through a multi-target tracking algorithm such as DeepSORT and FairMOT, it is possible to continuously detect and mark the location of the customer and the goods picked up by the customer in the surveillance image identified in step S3 in subsequent surveillance images. The residence time is recorded as follows: start timing from the time the goods are monitored to be picked up, and stop timing when it is detected that the goods are put back on the shelf or taken away, and the accumulated timing is the customer's residence time;

[0070] The method for determining whether the product is put back on the shelf is as follows: the range of the pick-up area is marked in the surveillance image; the pick-up area can be marked based on a preset marking frame; if the customer leaves the pick-up area and the product does not leave the pick-up area, the product is put back on the shelf; if the product is put back on the shelf, the specific time when the timer is stopped is the time when the customer leaves the pick-up area;

[0071] The method for determining whether the goods have been taken away is as follows: if both the customer and the goods leave the pick-up area, the goods are taken away; if the goods are taken away, the specific time when the timing stops is the time when both the customer and the goods leave the pick-up area;

[0072] S5: updating the first stay time of the product based on the unpurchased stay time, and updating the second stay time of the product based on the purchased stay time;

[0073] The total number of times the product is picked up is the sum of the number of times the product is picked up in all replenishment cycles;

[0074] The calculation method of the first residence time is as follows: for any kind of commodity, the customer's residence time is recorded when it is detected that the commodity is picked up; if the commodity is put back on the shelf, the recorded customer's residence time is marked as unpurchased residence time; all unpurchased residence times of the corresponding commodity are queried and the average is calculated to obtain the first residence time of the corresponding commodity.

[0075] The first dwell time is the average hesitation time of customers after picking up the product before finally giving up the purchase; if the first dwell time of a product is long, it may indicate that the customer has a high interest in the product, but may be hesitant due to factors such as price and brand awareness. If the first dwell time of a product is short, it may mean that the customer is not very interested in the product, or the product information is not attractive enough. Therefore, the longer the first dwell time, the higher the sales potential of the product.

[0076] The calculation method of the second stay time is as follows: for any kind of commodity, start recording the customer's stay time when it is detected that the commodity is picked up; if the commodity is taken away, mark the recorded customer's stay time as the purchased stay time; query all purchased stay times of the corresponding commodity and calculate the average to obtain the second stay time of the corresponding commodity.

[0077] The second dwell time is the average hesitation time of customers after picking up the product and finally deciding to buy it; if the second dwell time of a product is long, it may indicate that the purchase decision of the product is difficult, such as the high price; or the customer is not very loyal to the brand of the product and may compare between multiple brands. Or the purchase frequency of the product is low, and the customer may do more research and comparison before buying. If the second dwell time of a product is short, it may indicate that the customer is more loyal to the brand of the product and habitually buys products of the brand of the product, or the purchase frequency of the product is high, and the customer has formed a fixed purchase habit when buying. Therefore, the shorter the second dwell time of a product, the greater its sales potential.

[0078] S6: Calculate the purchase conversion rate of the product, and update the stocking threshold of the product based on the purchase conversion rate and the first residence time and the second residence time; if the inventory of any product is less than the corresponding stocking threshold, send a stocking reminder to the management personnel.

[0079] The calculation formula of the purchase conversion rate is as follows:

[0080] ;

[0081] Among them, C represents the purchase conversion rate of any product; Indicates the sales volume of the corresponding product; Indicates the total number of times the corresponding product is picked up.

[0082] The purchase conversion rate reflects the ratio of the number of items sold to the total number of times the item is picked up, and reflects the attractiveness of the item to customers. If the purchase conversion rate of a certain item is high, it indicates that consumers are highly satisfied with the item and have a strong willingness to buy it. You can continue to increase the purchase of such items to ensure sufficient inventory. If the purchase conversion rate of a certain item is low, it may mean that there are some problems with the item, such as overpriced, poor quality, incomplete functions, etc. You can consider reducing inventory appropriately.

[0083] The formula for updating the stock threshold of the product is as follows:

[0084] ;

[0085] Among them, B represents the updated stocking threshold of the product; Indicates the first dwell time of the corresponding product; represents the second dwelling time of the corresponding product; C represents the purchase conversion rate of the corresponding product; represents the adjustment coefficient of the first residence time, k represents the adjustment coefficient of the second residence time, represents the adjustment coefficient of purchase conversion rate, represents the reference value of the first dwell time, represents the reference value of the second dwell time, Indicates the benchmark value of purchase conversion rate. The reference stocking thresholds for the corresponding commodities are set by technicians in this field based on actual needs.

[0086] First, set the initial stocking threshold based on expert experience , in Settings When updating the stocking threshold, you can refer to the sales speed, shelf life and other data of the product; then adjust the stocking threshold based on the first dwell time, the second dwell time and the purchase conversion rate. Based on the update formula of the above stocking threshold, the stocking threshold can be dynamically adjusted according to the sales situation of the product; for example: for products with a short second dwell time and a high purchase conversion rate, based on the adjustment of the above formula, the stocking threshold can be appropriately increased to ensure that the inventory of the product is sufficient to meet the purchase needs of customers; for products with a long first dwell time and a low purchase conversion rate, based on the above formula, the stocking threshold can be appropriately lowered to avoid inventory backlogs.

[0087] Example 2

[0088] This embodiment is the second embodiment of the present invention; it is based on the same inventive concept as embodiment 1, Figure 2 This embodiment introduces an automatic replenishment smart shelf based on the number of times a product is picked up, including a shelf body, a camera module, a database, a processing module, and a communication module; wherein:

[0089] The shelf body includes M grids, where M is a positive integer; each grid is a shelf area, and each shelf area is used to place a specified type of goods;

[0090] The camera module includes surveillance cameras installed above and on the sides of each shelf area, which are used to capture surveillance images of the corresponding shelf area; the monitoring range of the camera module covers the entire shelf, can record activities in the shelf area, and provide image data for subsequent target detection.

[0091] The database is used to store the commodity information of each commodity, including the inventory quantity, sales quantity, number of pick-ups, total number of pick-ups, unpurchased stay time, and purchased stay time of each commodity;

[0092] The processing module is configured with a target detection algorithm and a multi-target tracking algorithm, which are used to detect and track customers and the goods picked up by customers from surveillance images; the processing module is also configured with a replenishment threshold for each product, and automatically replenishes the products based on the number of times each product is picked up and the corresponding replenishment threshold in each replenishment cycle to generate a replenishment list; the replenishment list includes a list of products that need to be replenished from the warehouse to the smart shelf and the replenishment quantity of each product, wherein the replenishment quantity plus the redundancy of the product is not higher than the maximum capacity of the smart shelf for the product.

[0093] The processing module is also configured with a stocking threshold for each product and an update formula for the stocking threshold. The stocking threshold for each product is updated by calculating the purchase conversion rate, the first dwell time, and the second dwell time. If the inventory of any product is less than the corresponding stocking threshold, the processing module generates a stocking reminder.

[0094] The communication module is used to send the replenishment list and stocking reminder to the management personnel. The communication module realizes the data transmission between the smart shelf and the external system, ensuring the real-time transmission of information.

[0095] The specific functional implementation of each of the above modules refers to the relevant content of the automatic replenishment method based on the number of times a product is picked up described in Example 1, and will not be repeated here.

[0096] Example 3

[0097] Based on the same inventive concept as other embodiments, refer to Figure 3 ,This embodiment introduces an application scenario of an automatic replenishment smart shelf based on the number of ,items picked up.

[0098] The smart shelf is a comprehensive system that integrates monitoring equipment and data processing and analysis modules. It aims to achieve efficient inventory management and replenishment decisions, improve sales efficiency and customer satisfaction through real-time collection and intelligent analysis of product-related information.

[0099] The main body of the shelf is a multi-layer structure, and each layer is used as an independent shelf area to place specific types of goods, such as shampoo. The grid design facilitates the classification and management of different goods, and is also conducive to the precise positioning and data collection of monitoring equipment. The camera module includes monitoring cameras installed above and above each shelf area. The monitoring range of the camera covers the entire shelf area, and can clearly capture the process of picking up goods and the behavior of customers. When the goods are detected to be picked up, the processing module will retrieve the monitoring video taken by the corresponding camera. The image data captured by the camera is used for target detection and analysis, such as identifying customers and the types and quantities of shampoos picked up by customers. The database is the data storage center of the smart shelf system. It stores various data related to goods, including the inventory, sales, number of pick-ups, total number of pick-ups, unpurchased stay time, purchased stay time and other information of each product. These data are the basis for the system to analyze and make decisions. The database needs to have efficient data storage and retrieval capabilities to meet the system's processing needs for large amounts of data. The processing module is the core component of the smart shelf system, which undertakes the key tasks of data processing and analysis. It is equipped with target detection algorithm and multi-target tracking algorithm to detect and track customers and the shampoo picked up by customers from the surveillance images captured by the camera. The processing module is also responsible for calculating various indicators, such as purchase conversion rate, replenishment threshold, stocking threshold, etc. It conducts comprehensive analysis and decision-making based on the data provided by the camera module and the historical data stored in the database. For example, it determines whether replenishment is needed based on the number of times shampoo is picked up and the redundancy, and adjusts the stocking threshold based on the dwell time and purchase conversion rate. The communication module is responsible for realizing data transmission between the smart shelf system and the external system. It sends the replenishment list and stocking reminder generated by the processing module to the management personnel so that the management personnel can understand the inventory situation in time and make corresponding decisions. At the same time, the communication module can also receive information from external systems, such as supplier supply information, etc., to provide a more comprehensive reference for the system's decision-making.

[0100] When a customer picks up shampoo, the surveillance image captured by the camera module will also be retrieved by the processing module. The processing module determines the specific behavior of the customer picking up shampoo and related information, such as the time of picking up and the number of times the shampoo is picked up, through comprehensive analysis of these data, and stores this information in the database. The processing module performs analysis based on the historical data stored in the database and the new data currently collected. For example, it will compare the number of times the shampoo is picked up in the current replenishment cycle with the number of times the shampoo is picked up in the previous replenishment cycle, and adjust the replenishment threshold according to the change in the number of times the shampoo is picked up. At the same time, it will calculate the redundancy of shampoo to determine whether replenishment is needed. In terms of adjusting the stocking threshold, the processing module will conduct a comprehensive analysis based on the unpurchased residence time, purchased residence time and purchase conversion rate of shampoo to determine whether the stocking threshold needs to be adjusted. After making a replenishment and stocking decision, the processing module will send the replenishment list and stocking reminder to the management personnel through the communication module. The management personnel arrange the replenishment and stocking work based on this information, and can also feedback the relevant decisions to the processing module so that the system can further optimize the decision-making process.

[0101] During the whole process, the data in the database will be continuously updated. For example, when a replenishment or stocking operation is completed, the inventory, sales volume and other data in the database will be updated accordingly. These updated data will become the basis for the next analysis and decision-making, forming a dynamic cycle. At the same time, the data between different components is shared. For example, the processing module can obtain the data in the database for analysis at any time, and the data provided by the camera module will also be stored in the database in a timely manner for use by the processing module.

[0102] The smart shelf automatically replenishes shampoo in the following way: First, the replenishment threshold is adjusted dynamically based on the number of times shampoo was picked up in the previous replenishment cycle and the number of times it was picked up in the previous replenishment cycle. For example, if the number of pick-ups in the previous cycle was 80 and the number of pick-ups in the previous cycle was 50, the replenishment threshold will be increased accordingly. Then the redundancy of shampoo is calculated, and if the redundancy is less than the replenishment threshold, replenishment is performed. If the replenishment threshold is 10 bottles and the redundancy is detected to be 6 bottles, replenishment is required.

[0103] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the purpose and scope of protection of the present invention, which are all within the protection of the present invention.

Claims

1. An automatic replenishment method based on the number of times a product is picked up, characterized in that: The following steps are involved: S1: Monitor the shelves and goods through surveillance video, and when it is detected that the goods are picked up, capture the surveillance image of the picking area; the picking area is the shelf area corresponding to the picked goods; S2: performing target detection on the surveillance image of the pickup area to identify the customers in the surveillance image and the types and quantities of the corresponding commodities; S3: Continuously track and monitor the pickup area, record the customer's stay time and the number of times each product is picked up in the current replenishment cycle; automatically replenish the products based on the number of times each product is picked up and the corresponding replenishment threshold; The replenishment threshold for any product is adjusted based on the number of times the corresponding product is picked up during the current replenishment cycle. The formula is as follows: ； Where D represents the adjusted replenishment threshold of any commodity; Indicates the initial value of the replenishment threshold of the corresponding product; Indicates the maximum value of the replenishment threshold of the corresponding product; n is the number of times the corresponding product is picked up in the current replenishment cycle; is the adjustment factor; The automatic replenishment of the goods specifically includes: automatically replenishing the goods based on the cumulative replenishment quantity, cumulative sales volume and replenishment threshold of the corresponding goods in all replenishment cycles of any goods; S4: Update the total number of times the product is picked up and record the customer's unpurchased stay time and purchased stay time; S5: updating a first stay time of the product based on the unpurchased stay time, and updating a second stay time of the product based on the purchased stay time; the first stay time is the average of the unpurchased stay time of the product; and the second stay time is the average of the purchased stay time of the product; S6: Calculate the purchase conversion rate of the product, and update the stocking threshold of the product based on the purchase conversion rate, the first dwell time, and the second dwell time; the purchase conversion rate of any product is the ratio of the sales volume of the product to the total number of pick-ups; if the inventory volume of any product is less than the corresponding stocking threshold, send a stocking reminder to the management personnel; The method for updating the stocking threshold of a product is as follows: setting a benchmark value of the first residence time, a benchmark value of the second residence time, a benchmark value of the purchase conversion rate, and a reference stocking threshold for each product, and calculating and updating the stocking threshold for each product in combination with the first residence time, the second residence time, and the purchase conversion rate of each product.

2. The automatic replenishment method based on the number of times a product is picked up as claimed in claim 1, characterized in that: The method for continuously tracking and monitoring the pickup area and recording the customer's stay time is as follows: continuously capture surveillance images from surveillance video, and detect and track the customer and the goods picked up by the customer in each frame of the surveillance image through a multi-target tracking algorithm; the stay time is recorded as follows: start timing from the time the goods are monitored to be picked up, and stop timing when it is detected that the goods are put back on the shelf or taken away, and the accumulated timing time is the customer's stay time.

3. The automatic replenishment method based on the number of times a product is picked up as claimed in claim 2, characterized in that: The method for determining whether the product is put back on the shelf is as follows: the range of the pick-up area is marked in the surveillance image; if the customer leaves the pick-up area but the product does not leave the pick-up area, the product is put back on the shelf; if the product is put back on the shelf, the specific time when the timer is stopped is the time when the customer leaves the pick-up area; The method for determining whether the goods are taken away is as follows: if both the customer and the goods leave the pickup area, the goods are taken away; if the goods are taken away, the specific time to stop timing is the time when both the customer and the goods leave the pickup area.

4. The automatic replenishment method based on the number of times a product is picked up as claimed in claim 3, characterized in that: A corresponding replenishment threshold is preset for each commodity; the method for automatically replenishing commodities based on the number of times each commodity is picked up and the corresponding replenishment threshold is as follows: if the redundancy of any commodity in the current replenishment cycle is less than the corresponding replenishment threshold, the commodity is replenished; the redundancy is the cumulative replenishment quantity of the commodity in all replenishment cycles minus the cumulative sales of the corresponding commodity in all replenishment cycles; in the current replenishment cycle, if the number of times any commodity is picked up is not less than the preset pick-up number threshold, the replenishment threshold is adjusted.

5. The automatic replenishment method based on the number of times a product is picked up as claimed in claim 4, characterized in that: When any replenishment cycle begins, the initial value of the replenishment threshold based on the number of times any product is picked up To assign values, the formula is as follows: ； in, Indicates the replenishment threshold of the corresponding product at the end of the previous replenishment cycle; Indicates the number of times the corresponding product was picked up in the previous replenishment cycle. express The number of times the corresponding product was picked up in the previous replenishment cycle of the corresponding replenishment cycle; Represents the weight coefficient.

6. The automatic replenishment method based on the number of times a product is picked up as claimed in claim 5, characterized in that: The total number of times the product is picked up is the sum of the number of times the product is picked up in all replenishment cycles; The calculation method of the first residence time is as follows: for any kind of commodity, the customer's residence time is recorded when it is detected that the commodity is picked up; if the commodity is put back on the shelf, the recorded customer's residence time is marked as unpurchased residence time; all unpurchased residence times of the corresponding commodity are queried and the average is calculated to obtain the first residence time of the corresponding commodity.

7. The automatic replenishment method based on the number of times a product is picked up as claimed in claim 6, characterized in that: The calculation method of the second stay time is as follows: for any kind of commodity, start recording the customer's stay time when it is detected that the commodity is picked up; if the commodity is taken away, mark the recorded customer's stay time as the purchased stay time; query all purchased stay times of the corresponding commodity and calculate the average to obtain the second stay time of the corresponding commodity.

8. An automatic replenishment smart shelf based on the number of times a product is picked up, used to implement the automatic replenishment method based on the number of times a product is picked up as described in any one of claims 1 to 7, characterized in that: It includes a shelf body, a camera module, a database, a processing module, and a communication module; among which: The shelf body includes M grids, where M is a positive integer; each grid is a shelf area, and each shelf area is used to place a specified type of goods; The camera module includes surveillance cameras installed above and above the side of each shelf area, and is used to capture surveillance images of the corresponding shelf area; The database is used to store the commodity information of each commodity, including the inventory quantity, sales quantity, number of pick-ups, total number of pick-ups, unpurchased stay time, and purchased stay time of each commodity; The processing module is configured with a target detection algorithm and a multi-target tracking algorithm, which are used to detect and track customers and the goods picked up by customers from the surveillance images; the processing module is also configured with a replenishment threshold for each product, and automatically replenishes the products based on the number of times each product is picked up and the corresponding replenishment threshold in each replenishment cycle, and generates a replenishment list; The processing module is also configured with a stocking threshold for each commodity and an updating formula for the stocking threshold. The stocking threshold for each commodity is updated by calculating the purchase conversion rate, the first dwell time, and the second dwell time. If the inventory of any commodity is less than the corresponding stocking threshold, the processing module generates a stocking reminder. The communication module is used to send the replenishment list and the stocking reminder to the management personnel.

Images