AI intelligent container settlement method and device based on gravity and vision

Abstract

The invention provides an AI intelligent container settlement method and device based on gravity and vision, and the method comprises the steps: obtaining the information of commodities in sale in a database, and obtaining the standard weight corresponding to each commodity; obtaining a transaction video of a commodity purchased by a user and uploading the transaction video to a server to obtain a visual identification result of the commodity, wherein the visual identification result comprises a predicted commodity combination; recording a total weight change value in the whole transaction process, obtaining all possible commodity combinations according to the total weight change value and the standard weight corresponding to the commodity, and taking a commodity combination set formed by all possible commodity combinations as a gravity recognition result; and calculating a settlement result based on the visual identification result and the gravity identification result. According to the invention, the visual appearance and weight change of the commodity are considered during settlement, the commodity identification accuracy is improved, the automatic settlement proportion of the intelligent container is greatly improved, the settlement time is also remarkably shortened, the demand for manual intervention is reduced, and the cost is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of commodity identification, and in particular to an AI intelligent vending cabinet settlement method and device based on gravity and vision. Background Art

[0002] An AI (Artificial Intelligence) intelligent vending cabinet is an intelligent device that identifies information about purchasers and purchased items through artificial intelligence technology to achieve automated sales services, and has advantages such as automated sales, intelligent management, and optimized user experience.

[0003] Currently, most intelligent vending cabinets on the market are mainly divided into two categories: intelligent vending cabinets based on vision algorithms and weighing intelligent vending cabinets. For intelligent vending cabinets based on vision algorithms, they mainly rely on computer vision algorithms to identify purchased items, that is, by uploading videos of the purchaser's shopping process in real time, and then using computer vision algorithms to detect and identify the purchased commodities in the video. However, in the actual purchase process, due to problems such as possible commodity occlusion, camera exposure, low light, and similar commodity packaging, situations where the vision algorithm recognition results are inaccurate often occur. For weighing intelligent vending cabinets, they utilize the principle of gravity sensing. A weighing device is provided at the bottom of the shelf board to measure the weight of the commodity and monitor the number of times and pieces of the commodity picked up by the consumer to achieve correct deduction of fees. However, weighing vending cabinets also have some defects. That is, if there are commodities with similar weights but large price differences, gravity alone cannot distinguish the specific purchased commodities, which may lead to incorrect deduction of fees. Due to the above problems, the current automatic settlement ratio of intelligent vending cabinets is relatively low, and the manual participation is high, resulting in problems such as a high customer complaint ratio and poor user experience. Summary of the Invention

[0004] The present invention provides an AI intelligent vending cabinet settlement method and device based on gravity and vision to solve the defects in the prior art that intelligent vending cabinets based solely on vision algorithms or weighing have a relatively low automatic settlement ratio, high manual participation, and poor user experience, and to achieve the effect of improving the automatic settlement ratio of intelligent vending cabinets, reducing the need for manual intervention, and improving the user experience.

[0005] The present invention provides an AI intelligent vending cabinet settlement method based on gravity and vision, including:

[0006] Obtain information on commodities for sale in the database and obtain the standard weights corresponding to various commodities;

[0007] Obtain the transaction video of the user purchasing commodities and upload it to the server, and obtain the visual recognition result of the commodities from the server. The visual recognition result includes a predicted commodity combination;

[0008] Record the total weight change during the entire transaction process, calculate all possible product combinations based on the total weight change and the standard weights of various products, and use the product combination set consisting of all possible product combinations as the gravity recognition result;

[0009] A settlement result is calculated based on the visual recognition result and the gravity recognition result.

[0010] According to the AI intelligent container settlement method based on gravity and vision provided by the present invention, all possible commodity combinations are calculated based on the total weight change value and the standard weight corresponding to various commodities, including:

[0011] Determine the weight of the product with the lightest standard weight among all products;

[0012] Calculate the theoretical maximum number of items that can be taken based on the weight of the lightest item and the total weight change;

[0013] Based on the total weight change and the theoretical maximum number of commodities that can be taken, all possible commodity combinations are calculated using a dynamic programming algorithm;

[0014] Determine all possible product combinations corresponding to each product quantity.

[0015] According to the AI intelligent container settlement method based on gravity and vision provided by the present invention, a commodity combination set consisting of all possible commodity combinations is used as a gravity recognition result, including:

[0016] Construct a candidate set based on all possible product combinations corresponding to each product quantity;

[0017] The product combinations in the candidate set are deduplicated to obtain a final product combination set as a gravity recognition result.

[0018] According to the present invention, an AI intelligent container settlement method based on gravity and vision is provided. Based on the total weight change value and the theoretical maximum number of commodities that can be taken, all possible commodity combinations are calculated by a dynamic programming algorithm, including:

[0019] For each possible number of items i, use the dynamic programming table dp[i][G] to represent all combinations of i items with a total weight equal to the total weight change G; where 1≤i≤n, and n is the theoretical maximum number of items that can be taken.

[0020] Initialize dp[0][0] = True; for each item, update dp[i][G] = dp[i][G]||dp[i-1][G-weight[k]], where weight[k] is the weight of the kth item;

[0021] Record all combinations that satisfy dp[i][G] is True, that is, obtain all possible product combinations.

[0022] According to the present invention, an AI intelligent container settlement method based on gravity and vision is provided, which calculates a settlement result based on the visual recognition result and the gravity recognition result, including:

[0023] Step 1: If the predicted product combination is included in the product combination set, the predicted product combination is used as the settlement result and the process proceeds to step 5; otherwise, the process proceeds to step 2;

[0024] Step 2: If the predicted product combination contains two products and only one of the two products is included in the product combination set, the product included in the product combination set is used as the settlement result and the process proceeds to step 5; otherwise, the process proceeds to step 3;

[0025] Step 3: If only one of all the product combinations in the product combination set contains the predicted product combination, the product combination containing the predicted product combination is used as the settlement result and the process proceeds to Step 5; otherwise, the process proceeds to Step 4;

[0026] Step 4: If the product combination set contains only one product combination, the product combination is used as the settlement result and the process proceeds to step 5; otherwise, the process proceeds to step 6;

[0027] Step 5: Determine the product combination as the settlement result and make the settlement;

[0028] Step 6: Determine whether settlement conditions are not met.

[0029] According to the present invention, an AI intelligent container settlement method based on gravity and vision, after determining that the settlement conditions are not met, further includes:

[0030] The transaction video of the user purchasing the goods is transmitted to the manual settlement module, the settlement result output by the manual settlement module is obtained and the settlement is performed.

[0031] According to the present invention, an AI intelligent container settlement method based on gravity and vision is provided, which records the total weight change value during the entire transaction process, including:

[0032] Record the first value when the weight value is stable after the intelligent vending cabinet is opened and the second value when the weight value is stable after the cabinet is closed, and use the difference between the first value and the second value as the total weight change value.

[0033] The present invention also provides an AI intelligent vending cabinet settlement device based on gravity plus vision, including:

[0034] A data acquisition module, configured to acquire information of commodities on sale in a database and obtain the standard weights corresponding to various commodities;

[0035] A vision recognition module, configured to acquire a transaction video of commodities purchased by a user and upload it to a server, and obtain a vision recognition result of the commodities from the server, where the vision recognition result includes a predicted commodity combination;

[0036] A gravity recognition module, configured to record the total weight change value during the entire transaction process, calculate all possible commodity combinations according to the total weight change value and the standard weights corresponding to various commodities, and use the set of commodity combinations formed by all possible commodity combinations as the gravity recognition result;

[0037] A decision-making module, configured to infer a settlement result based on the vision recognition result and the gravity recognition result.

[0038] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the processor executes the program, it implements the AI intelligent vending cabinet settlement method based on gravity plus vision as described in any one of the above.

[0039] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the AI intelligent vending cabinet settlement method based on gravity plus vision as described in any one of the above.

[0040] The AI smart vending machine settlement method and device based on gravity and vision provided by the present invention obtains the standard weight corresponding to various commodities by obtaining information on commodities on sale in the database; obtains the transaction video of the user purchasing commodities and uploads it to the server, obtains the visual recognition results of the commodities from the server, and the visual recognition results include the predicted commodity combinations; records the total weight change value of the entire transaction process, calculates all possible commodity combinations based on the total weight change value and the standard weight corresponding to various commodities, and uses the commodity combination set consisting of all possible commodity combinations as the gravity recognition result; and calculates the settlement result based on the visual recognition result and the gravity recognition result. Through the application of this method and device, the visual appearance and weight change of the commodity are taken into account at the same time during settlement, providing double protection, effectively improving the accuracy of commodity recognition, and thus significantly improving the automatic settlement efficiency of the smart vending machine. In addition, by adopting a settlement solution that integrates "gravity + vision", the automatic settlement ratio of the smart vending machine is greatly improved, and the settlement time is correspondingly significantly shortened, reducing the need for manual intervention, thereby effectively reducing the overall cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 This is a flow chart of the AI intelligent container settlement method based on gravity and vision provided by the present invention;

[0043] Figure 2 is a schematic diagram of a process for calculating all possible combinations of goods based on total weight changes in an embodiment of the present invention;

[0044] Figure 3 1 is a flow chart of determining the final settlement result based on the visually predicted commodity combination and the gravity-recognized commodity combination in an embodiment of the present invention;

[0045] Figure 4 This is a structural diagram of the AI intelligent container settlement device based on gravity and vision provided by the present invention;

[0046] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0047] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] The following will describe Figure 1-Figure 3 the AI intelligent vending cabinet settlement method based on gravity plus vision of the present invention.

[0049] As Figure 1 shown, the AI intelligent vending cabinet settlement method based on gravity plus vision provided by the present invention includes the following steps:

[0050] S1. Obtain the information of the products on sale in the database and obtain the standard weights corresponding to various products.

[0051] In an optional embodiment of the present invention, the database is located on the server side. The database stores information about various products on sale, including but not limited to the name, image, and standard weight of the products, etc. The standard weight is expressed in the unit g (gram). For example, the standard weight of a certain brand of cup yogurt on sale is 100g.

[0052] S2. Obtain the transaction video of the products purchased by the user and upload it to the server, and obtain the visual recognition result of the products from the server. The visual recognition result includes the predicted product combination.

[0053] Specifically, a camera for shooting is installed in the intelligent vending cabinet, and the camera can record the entire transaction video of the products purchased by the user. To reduce the software and hardware requirements of the intelligent vending cabinet, after obtaining the transaction video, the transaction video is uploaded to the server, and the server uses visual algorithms such as detection, tracking, and recognition to process the transaction video and obtain the final visual recognition result. The visual recognition result includes the predicted product combination. For example, after uploading a certain transaction video to the server, the server processes the video with visual algorithms and predicts that the products traded by the user are 2 bottles of a certain brand of beverage and a cup of a certain brand of yogurt, that is, the predicted product combination. In an optional embodiment of the present invention, if the software and hardware processing capabilities of the intelligent vending cabinet are sufficient, the transaction video can also be directly processed with visual algorithms locally, and the present invention does not limit this.

[0054] S3. Record the total weight change value during the entire transaction process. According to the total weight change value and the standard weights corresponding to various products, calculate all possible product combinations, and use the set of product combinations composed of all possible product combinations as the gravity recognition result.

[0055] Smart vending machines are equipped with weighing devices at the bottom of the shelves, which record the weight of items before and after they are removed. The total weight change is calculated by recording a first value when the smart vending machine's door is open and stable (items not removed), and a second value when the door is closed and stable (items removed). The difference between the first and second values is used as the total weight change. For example, if a smart vending machine records a stable weight of 25,300g after opening the door and 25,150g after closing the door, the total weight change is 25,300-25,150=150g.

[0056] In an optional embodiment of the present invention, the total weight change of the transaction process and the standard weight corresponding to each sku (commodity) can be passed to the API for calculating the sku combination, and the API calls a specified function to obtain all the sku combinations that may correspond to the current weight change. Figure 2 As shown, according to the total weight change value and the standard weight corresponding to various commodities, all possible commodity combinations are calculated, including the following steps:

[0057] S31. Determine the weight of the product with the lightest standard weight among all products.

[0058] Specifically, by traversing the standard weights of all products in the database, the minimum value is found. For example, if the product weights in the database are [200g, 300g, 150g, 250g], then the lightest product weight g = 150g. It should be understood that there are usually hundreds or thousands of product information in the database. For simplicity and ease of understanding, the examples in the following parts of the present invention are all explained using the product weights in the database of [200g, 300g, 150g, 250g].

[0059] S32. Calculate the theoretical maximum number of items that can be taken based on the weight of the lightest item and the change in total weight.

[0060] S33. Based on the total weight change and the theoretical maximum number of goods that can be taken, all possible combinations of goods are calculated using a dynamic programming algorithm.

[0061] S34. Determine all possible product combinations corresponding to each product quantity.

[0062] Specifically, based on the change in total weight G and the weight of the lightest item g, we calculate the theoretical maximum number of items that can be taken in the most extreme case (i.e., taking the lightest item at a time). We calculate n = G / g, where n is the theoretical maximum number of items that can be taken. For example, if G = 600g and g = 150g, then n = 600 / 150 = 4. This means that a maximum of 4 items can be taken (each of the lightest items). For each possible number of items i from 1 to n, we calculate all possible combinations of i items whose total weight equals G.

[0063] For each possible number of items i, use the dynamic programming table dp[i][G] to represent all combinations of i items with a total weight equal to the total weight change value G; where 1≤i≤n, n is the theoretical maximum number of items that can be taken.

[0064] Initialize dp[0][0] = True, that is, select 0 items, and it is feasible to have a total weight of 0. For each item, update dp[i][G] = dp[i][G]||dp[i-1][G-weight[k]], where weight[k] is the weight of the kth item.

[0065] Record all combinations that satisfy dp[i][G] is True, that is, obtain all possible product combinations.

[0066] To illustrate, let's use an example. Suppose the weights of the items are [150, 200, 250, 300], the total weight variation G = 600, the weight of the lightest item g = 150, and the theoretical maximum number of items that can be taken is n = 4. The process of calculating all possible combinations of items using the dynamic programming algorithm is as follows:

[0067] (1) When i=1, check whether there is a single product with a weight of 600. Result: No.

[0068] (2) When i=2, check the combination of two products:

[0069] 150+450 (there is no product with a weight of 450), result: none.

[0070] 200+400 (there is no product with a weight of 400), result: none.

[0071] 250+350 (there is no product with a weight of 350), result: none.

[0072] 300+300=600, result: [300,300] is a combination.

[0073] (3) When i=3, check the combination of three products:

[0074] 150 + 200 + 250 = 600, Result: [150, 200, 250] is a combination.

[0075] 150 + 150 + 300 = 600, Result: [150, 150, 300] is a combination.

[0076] 200 + 200 + 200 = 600, Result: [200, 200, 200] is a combination.

[0077] (4) When i = 4, check the combinations of four items:

[0078] 150 × 4 = 600, Result: [150, 150, 150, 150] is a combination. Other combinations are invalid.

[0079] Therefore, in the above examples, all possible item combinations calculated by the dynamic programming algorithm are:

[0080] Two items: [300, 300].

[0081] Three items: [150, 200, 250], [150, 150, 300], [200, 200, 200].

[0082] Four items: [150, 150, 150, 150].

[0083] Taking the set of item combinations composed of all possible item combinations as the gravity recognition result, it includes the following steps:

[0084] S35. Construct a candidate set based on all possible item combinations corresponding to each item quantity.

[0085] Taking the above item combinations as an example, the constructed set of item combinations is {[300, 300], [150, 200, 250], [150, 150, 300], [200, 200, 200], [150, 150, 150, 150]}.

[0086] S36. Remove duplicates from the item combinations in the candidate set to obtain the final set of item combinations as the gravity recognition result.

[0087] Since the order of products in a product combination is not fixed, for example, in a product combination of [150, 200, 250], there are permutations such as [150, 200, 250], [150, 250, 200], [200, 150, 250], [200, 250, 150], [250, 200, 150], [250, 150, 200], etc. Therefore, it is necessary to remove duplicates from the product combinations in the set, and only retain one permutation for each product combination.

[0088] S4. Based on the visual recognition result and the gravity recognition result, calculate the settlement result.

[0089] Specifically, the predicted product combination of the visual recognition result is denoted as vsku, the set of product combinations of the gravity recognition result is denoted as wsku_set, and the product combinations in wsku_set are denoted as wsku. As Figure 3 shown, step S4 includes the following sub-steps:

[0090] S41. When the product combination set contains the predicted product combination, use the predicted product combination as the settlement result and go to step S45; otherwise, go to step S42.

[0091] This step is to determine whether the sku (product) combination vsku predicted by vision is in the set of sku combinations wsku_set recognized by gravity. If it is, it means that this sku combination has passed the double verification of vision and gravity. Therefore, it can directly enter step S45 for algorithm settlement according to this vsku; otherwise, go to the next step.

[0092] S42. When the predicted product combination contains two products and only one of the two products is included in the product combination set, use the product included in the product combination set as the settlement result and go to step S45; otherwise, go to step S43.

[0093] This step is to handle the situation of multiple detections in the visual result (in addition to the purchased products, there are additional detections). If vsku only contains two skus and only one sku of the two skus is included in wsku_set, we consider the other sku to be a product detected additionally. At this time, enter step S45 and perform algorithm settlement according to the visual sku included in wsku_set. Otherwise, go to step S43.

[0094] S43. When only one product combination in all product combinations of the product combination set contains the predicted product combination, use the product combination containing the predicted product combination as the settlement result and go to step S45; otherwise, go to step S44.

[0095] This step is to handle the situation of missed detection in the visual results (some of the purchased goods are not detected). If there is only one combined wsku in the wsku_set that contains the vsku, then enter step S45 and perform algorithmic settlement according to this wsku; otherwise, enter step S44.

[0096] S44. In the case where the set of product combinations contains only one product combination, use this product combination as the settlement result and enter step S45; otherwise, enter step S46.

[0097] If the wsku_set contains only one combined wsku, it means that only one product combination meets the current weight change. Therefore, directly enter step S45 and perform algorithmic settlement according to this product combination.

[0098] S45. Determine the product combination as the settlement result and perform settlement.

[0099] This step corresponds to the situation of settlement based on visual results and gravity results, without manual intervention.

[0100] S46. Determine that the settlement conditions are not met.

[0101] This step corresponds to the situation where the above conditions are not met. At this time, automatic settlement based on visual results and gravity results cannot be achieved.

[0102] In an alternative embodiment of the present invention, after it is determined in step S46 that the settlement conditions are not met, it further includes: transmitting the transaction video of the goods purchased by the user to the manual settlement module, obtaining the settlement result output by the manual settlement module and performing settlement. The staff can obtain the transaction video of the goods purchased by the user through the manual settlement module. After manually judging the types and quantities of the purchased goods, the judgment result is returned to the intelligent vending cabinet through the manual settlement module for the intelligent vending cabinet to perform settlement.

[0103] The following describes in detail the AI intelligent vending cabinet settlement method based on gravity plus vision provided by the present invention with a specific embodiment:

[0104] Assume that the weights of the goods in the database are [200g, 300g, 150g, 250g], that is, the lightest good weighs 150g.

[0105] The intelligent vending cabinet obtains the transaction video of the goods purchased by the user and uploads it to the server. The predicted product combination obtained by the server through visual recognition processing is soda of brand A (250g) and yogurt of brand B (200g), that is, [250, 200].

[0106] The total weight change value during the transaction detected by the intelligent vending cabinet is 450 grams. It can be known that the maximum number of items that can be taken theoretically is 450 / 150 = 3, which means that at most 3 items can be taken. For each possible number of items from 1 to 3, on the premise that the total weight meets 450g, the following item combination plans are obtained through calculation:

[0107] Two items: [200, 250], [300, 150].

[0108] Three items: [150, 150, 150].

[0109] Therefore, the set of item combinations obtained by gravity recognition is {[200, 250], [300, 150], [150, 150, 150]}.

[0110] Based on the calculation method in step S4 above, it can be known that the item combination [250, 200] obtained by visual recognition is in the set of item combinations, indicating that this item combination has passed the double inspection of vision and gravity. Therefore, directly determine this item combination (soda of brand A and yogurt of brand B) as the settlement result and conduct the settlement without manual intervention.

[0111] In summary, the AI intelligent vending cabinet settlement method based on gravity plus vision provided by the present invention obtains the information of the items on sale in the database and the standard weights corresponding to various items; obtains the transaction video of the items purchased by the user and uploads it to the server, and obtains the visual recognition result of the items from the server. The visual recognition result includes the predicted item combination; records the total weight change value during the entire transaction process, and calculates all possible item combinations based on the total weight change value and the standard weights corresponding to various items, and forms the set of item combinations composed of all possible item combinations as the gravity recognition result; based on the visual recognition result and the gravity recognition result, calculates the settlement result. Through the application of this method, both the visual appearance and weight change of the items are considered during settlement, providing double guarantees, effectively improving the item recognition accuracy, and thus significantly improving the automatic settlement efficiency of the intelligent vending cabinet. In addition, by adopting the settlement scheme that integrates "gravity + vision", the automatic settlement ratio of the intelligent vending cabinet is greatly increased, and the settlement time is correspondingly significantly shortened, reducing the need for manual intervention, and thus effectively reducing the overall cost.

[0112] Based on the same inventive concept, the present invention also provides an AI intelligent vending cabinet settlement device based on gravity plus vision. The following describes the AI intelligent vending cabinet settlement device based on gravity plus vision provided by the present invention. The AI intelligent vending cabinet settlement device based on gravity plus vision described below can be mutually referred to the AI intelligent vending cabinet settlement method described above.

[0113] As Figure 4 As shown in the figure, the AI intelligent vending cabinet settlement device based on gravity and vision provided by the present invention includes a data acquisition module 41, a vision recognition module 42, a gravity recognition module 43, and a decision module 44.

[0114] The data acquisition module 41 is used to acquire information on the goods on sale in the database and obtain the standard weights corresponding to various goods.

[0115] The vision recognition module 42 is used to acquire the transaction video of the goods purchased by the user and upload it to the server, and obtain the vision recognition result of the goods from the server. The vision recognition result includes the predicted product combination.

[0116] The gravity recognition module 43 is used to record the total weight change value during the entire transaction process, calculate all possible product combinations based on the total weight change value and the standard weights corresponding to various goods, and use the set of product combinations composed of all possible product combinations as the gravity recognition result.

[0117] The decision module 44 is used to infer the settlement result based on the vision recognition result and the gravity recognition result.

[0118] Figure 5 An example of the physical structure diagram of an electronic device is shown in Figure 5 As shown in the figure, the electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540. Among them, the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call the logical instructions in the memory 530 to execute the AI intelligent vending cabinet settlement method based on gravity and vision provided by the above-mentioned various methods. The method includes:

[0119] Acquire information on the goods on sale in the database and obtain the standard weights corresponding to various goods.

[0120] Acquire the transaction video of the goods purchased by the user and upload it to the server, and obtain the vision recognition result of the goods from the server. The vision recognition result includes the predicted product combination.

[0121] Record the total weight change value during the entire transaction process, calculate all possible product combinations based on the total weight change value and the standard weights corresponding to various goods, and use the set of product combinations composed of all possible product combinations as the gravity recognition result.

[0122] Infer the settlement result based on the vision recognition result and the gravity recognition result.

[0123] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0124] In another aspect, the present invention further provides a computer program product, comprising a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the AI intelligent container settlement method based on gravity and vision provided by the above methods, which includes:

[0125] Get information about the products on sale in the database and obtain the standard weights corresponding to various products;

[0126] Obtaining a transaction video of a user purchasing a product and uploading it to a server, and obtaining visual recognition results of the products from the server, wherein the visual recognition results include a predicted product combination;

[0127] Record the total weight change during the entire transaction process, calculate all possible product combinations based on the total weight change and the standard weights of various products, and use the product combination set consisting of all possible product combinations as the gravity recognition result;

[0128] A settlement result is calculated based on the visual recognition result and the gravity recognition result.

[0129] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the computer program is implemented to perform the AI intelligent container settlement method based on gravity and vision provided by the above methods, the method comprising:

[0130] Get information about the products on sale in the database and obtain the standard weights corresponding to various products;

[0131] Obtaining a transaction video of a user purchasing a product and uploading it to a server, and obtaining visual recognition results of the products from the server, wherein the visual recognition results include a predicted product combination;

[0132] Record the total weight change during the entire transaction process, calculate all possible product combinations based on the total weight change and the standard weights of various products, and use the product combination set consisting of all possible product combinations as the gravity recognition result;

[0133] A settlement result is calculated based on the visual recognition result and the gravity recognition result.

[0134] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0135] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. An AI intelligent container settlement method based on gravity and vision, characterized in that: Including: Obtain information on the goods for sale in the database and obtain the standard weights corresponding to various goods; Obtain the transaction video of the goods purchased by the user and upload it to the server, and obtain the visual recognition result of the goods from the server. The visual recognition result includes a predicted goods combination; Record the total weight change value during the entire transaction process. Based on the total weight change value and the standard weights corresponding to various goods, calculate all possible goods combinations, and use the set of goods combinations composed of all possible goods combinations as the gravity recognition result; Based on the visual recognition result and the gravity recognition result, deduce the settlement result.

2. The AI intelligent vending cabinet settlement method based on gravity plus vision according to claim 1, wherein Based on the total weight change value and the standard weights corresponding to various goods, calculate all possible goods combinations, including: Determine the weight of the goods with the lightest standard weight among all goods; Based on the weight of the lightest goods and the total weight change value, calculate the theoretical maximum number of goods that can be taken; Based on the total weight change value and the theoretical maximum number of goods that can be taken, calculate all possible goods combinations through a dynamic programming algorithm; Respectively determine all possible goods combinations corresponding to each quantity of goods.

3. The AI intelligent container settlement method based on gravity and vision according to claim 2 is characterized in that: Use the set of goods combinations composed of all possible goods combinations as the gravity recognition result, including: Construct a candidate set based on all possible goods combinations corresponding to each quantity of goods; Remove duplicates from the goods combinations in the candidate set to obtain the final set of goods combinations as the gravity recognition result.

4. The AI intelligent vending cabinet settlement method based on gravity plus vision according to claim 3, characterized in that, Based on the total weight change value and the theoretical maximum number of goods that can be taken, calculate all possible goods combinations through a dynamic programming algorithm, including: For each possible quantity i of goods that can be taken, use the dynamic programming table dp[i][G] to represent all combinations of taking i goods with the total weight exactly equal to the total weight change value G; where 1 ≤ i ≤ n, and n is the theoretical maximum number of goods that can be taken; Initialize dp[0][0] = True; for each good, update dp[i][G] = dp[i][G] || dp[i - 1][G - weight[k]], where weight[k] is the weight of the kth good; Record all combinations that satisfy dp[i][G] being True, that is, obtain all possible goods combinations.

5. The AI intelligent container settlement method based on gravity and vision according to claim 1 is characterized in that: Based on the visual recognition result and the gravity recognition result, deduce the settlement result, including: Step 1: When the predicted goods combination is included in the set of goods combinations, use the predicted goods combination as the settlement result and go to Step 5; otherwise, go to Step 2; Step 2: When there are two goods in the predicted goods combination and only one of the two goods is included in the set of goods combinations, use the good included in the set of goods combinations as the settlement result and go to Step 5; otherwise, go to Step 3; Step 3: When only one goods combination in all goods combinations of the set of goods combinations includes the predicted goods combination, use the goods combination including the predicted goods combination as the settlement result and go to Step 5; otherwise, go to Step 4; Step 4: If the product combination set contains only one product combination, the product combination is used as the settlement result and the process proceeds to step 5; otherwise, the process proceeds to step 6; Step 5: Determine the product combination as the settlement result and make the settlement; Step 6: Determine whether settlement conditions are not met.

6. The AI intelligent vending cabinet settlement method based on gravity plus vision according to claim 5, wherein After determining that the settlement conditions are not met, it also includes: The transaction video of the user purchasing the goods is transmitted to the manual settlement module, the settlement result output by the manual settlement module is obtained and the settlement is performed.

7. The AI intelligent container settlement method based on gravity and vision according to any one of claims 1 to 6, characterized in that: Record the total weight change during the entire transaction, including: The first value of the weight value of the smart container when the door is opened and the second value of the weight value when the door is closed are recorded, and the difference between the first value and the second value is used as the total weight change value.

8. An AI intelligent container settlement device based on gravity and vision, characterized in that: include: The data acquisition module is used to obtain information about the goods on sale in the database and obtain the standard weight corresponding to various goods; A visual recognition module is used to obtain a transaction video of a user purchasing a product and upload it to a server, and obtain visual recognition results of the products from the server, wherein the visual recognition results include a predicted product combination; A gravity recognition module is used to record the total weight change value during the entire transaction process, calculate all possible product combinations based on the total weight change value and the standard weight corresponding to various products, and use the product combination set consisting of all possible product combinations as the gravity recognition result; A decision module is used to calculate a settlement result based on the visual recognition result and the gravity recognition result.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, the AI intelligent container settlement method based on gravity and vision is implemented as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the AI intelligent container settlement method based on gravity and vision is implemented as described in any one of claims 1 to 7.

Images