Commercial kitchen design system and method based on multi-modal large model

Abstract

The application discloses a commercial kitchen design system and method based on a multi-modal large model, relates to the technical field of commercial kitchen design, and is used for solving the problem that menu content is often changed due to temporary orders and reducing the stability of kitchen operation. The application realizes real-time detection of updated names and the number of dishes with the same name in a menu database, calls corresponding dish processing time and performs marking. Within a set recognition time, the number of dishes with the same name and the processing time are used to calculate a processing frequency and perform processing sorting. The application calculates the priority of tool use according to the sorting result, and evaluates the high-frequency wear rate of tools in each category in combination with the use duration. The application screens target processing tools by comprehensively considering the priority and the wear rate, detects the position of the tool rack where the target processing tools are located and performs marking, and realizes replacement with a default commonly used tool rack through a conveyor belt, so as to improve the real-time efficiency and quality of dish making in a commercial kitchen.

Description

Technical Field

[0001] The present invention relates to the technical field of commercial kitchen design, and more particularly, to a commercial kitchen design system and method based on a multimodal large model. Background Art

[0002] As the degree of automation and intelligence in commercial kitchens continues to improve, issues such as equipment scheduling, tool selection, and job sequencing in the food processing process are becoming more complex. In existing technologies, food preparation tasks are usually driven by a preset menu list, and then the processing tools are called manually based on experience. After receiving the menu task, the operator needs to identify the names of the dishes contained therein.

[0003] The existing technology has the following deficiencies:

[0004] At present, in practical applications, menu content is often updated frequently due to factors such as temporary orders and changes in customer tastes, which leads to drastic changes in the frequency of processing tasks, chaotic tool distribution, severe wear and tear of high-frequency tools, and reduced kitchen operation stability. Therefore, a commercial kitchen design system and method based on a multimodal large model is proposed.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a commercial kitchen design system and method based on a multimodal big model, which solves the problems raised in the above-mentioned background technology by using the multimodal big model to integrate modeling and intelligent scheduling mechanism of multi-source information such as menu data, dish processing tasks, tool usage status and spatial location identification.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a commercial kitchen design method based on a multimodal large model, comprising the following steps:

[0008] Step S1: Real-time detection of menu update names and the number of dishes with the same name in the menu database, calling the corresponding stored dish processing time according to the menu update name, and marking different menu update names respectively;

[0009] Step S2: setting a recognition time, calculating the processing frequency of different marked dish names using the number of dishes with the same name within the recognition time, and sorting the different marked dish names in the menu database based on the dish processing time of the corresponding marked dish names to generate a sorting feature;

[0010] Step S3: Count the usage frequencies of different categories of processing tools based on the marked dish names, calculate the usage priority of each category of processing tools based on the sorting characteristics, collect the usage time of all processing tools, and calculate the high-frequency use wear rate of each category of processing tools based on the usage time;

[0011] Step S4: Filter out target category processing tools based on the usage priority and high-frequency wear rate of each category of processing tools, detect the tool rack where the target category processing tools are located and mark their positions, and replace the tool rack where the target category processing tools are located with the default commonly used tool rack through a conveyor belt based on the marked positions.

[0012] In a preferred embodiment, in step S1, the user terminal receives the user's selected dish name and transmits it to the menu database as the menu update name, traverses all the dish names stored in the menu database according to the menu update name and performs identification and matching, and calls the dish processing time of the dish name after identification and matching as the dish processing time of the menu update name;

[0013] Count the number of identical menu update names in the menu database in real time, and use the statistical result as the number of dishes with the same name corresponding to the updated menu name;

[0014] The menu update name, the number of dishes with the same name corresponding to the menu update name, and the dish processing time are integrated to generate menu information corresponding to the menu update name and temporarily store it in the menu database.

[0015] In a preferred embodiment, in step S1, when the user's selected dish is completed, the menu information corresponding to the updated dish name is deleted from the menu database, and different menu updated names in the menu database are marked respectively to obtain different marked dish names.

[0016] In a preferred embodiment, in step S2, a period of time is selected as the recognition time, and the number of dishes with the same name as each marked dish name is extracted from the menu information within the recognition time, and the ratio of the number of dishes with the same name as each marked dish name in the menu database to the total number of dishes in the menu database is used as the frequency to be processed of the corresponding marked dish name;

[0017] A logistic regression model is constructed based on the dish processing time of each marked dish name and the frequency of pending processing of the corresponding marked dish name, and the ranking features of each marked dish name are calculated. The different marked dish names in the menu database are processed and ranked based on the ranking features.

[0018] In a preferred embodiment, in step S2, the specific steps of constructing a logistic regression model to calculate the ranking features of each marked dish name are as follows:

[0019] Data processing: The logarithm of the dish processing time of each marked dish name is taken as the dish processing index of the corresponding marked dish name;

[0020] Calculation of intermediate parameters: The ratio of the frequency of each marked dish name to be processed to the dish processing index is used as the intermediate parameter and marked as z;

[0021] Construct a logistic regression model: pass the intermediate parameters into the logistic regression formula: , where L is the logistic regression result of each marked dish name, e is the natural base, and the logistic regression result of each marked dish name is used as the ranking feature of the corresponding marked dish name;

[0022] Processing and sorting: For each marked dish name in the menu database, sort them from large to small according to the sorting characteristics.

[0023] In a preferred embodiment, in step S3, based on the marked dish name, the number of times each category of processing tools for each marked dish name is used is obtained from the dish preparation method database, the number of times each category of processing tools for the current marked dish name is accumulated to obtain the total number of times all category processing tools for the current marked dish name are used, and the ratio of the number of times each category of processing tools is used to the total number of times all category processing tools are used is calculated to obtain the usage frequency of the category processing tool corresponding to the current marked dish name;

[0024] The usage frequency of each category of processing tools for the current marked dish name is multiplied by the sorting feature of the marked dish name, and the accumulated result is calculated, and the accumulated result is used as the usage priority of each category of processing tools.

[0025] In a preferred embodiment, in step S3, the usage time of each type of processing tool is obtained by calling the dish preparation method database;

[0026] The usage time of the processing tool is the difference between the current time point and the time point recorded in the dish preparation method database when the corresponding processing tool is stored in the tool rack;

[0027] The usage time of each category of processing tools is standardized, and the standardized results are used as the high-frequency use wear rate of each category of processing tools.

[0028] In a preferred embodiment, in step S4, the use priority of each category of processing tools is standardized;

[0029] Substitute the high-frequency wear rate of each category of processing tools and the standardized usage priority of each category of processing tools into the product-amplification difference adjustment model to obtain the selection score of each category of processing tools;

[0030] The selection scores of each category processing tool are sorted in descending order, and the category processing tool corresponding to the maximum selection score of each category processing tool is used as the target category processing tool.

[0031] In a preferred embodiment, in step S4, the tool rack where the target category processing tool is located is detected and the position is marked to obtain the position mark of each target category processing tool;

[0032] The position of the tool rack corresponding to the position mark of each target category processing tool is set by two-dimensional coordinate system modeling, and the position coordinates of the corresponding position mark of each target category processing tool and the position coordinates of the default common tool rack are set. The distance length between the tool rack where the tool of each target category processing tool is located and the default common tool rack is calculated using the Euclidean distance formula;

[0033] The distance lengths between the tool rack where each target category processing tool is located and the default commonly used tool rack are sorted in order from small to large, and the tool rack where the target category processing tool is located corresponding to the minimum distance length between the tool rack where the target category processing tool is located and the default commonly used tool rack is selected and replaced with the default commonly used tool rack through a conveyor belt.

[0034] The commercial kitchen design system based on the multimodal large model includes a menu update module, a dish sorting module, a tool screening module, and a position exchange module;

[0035] The menu update module is used to detect the menu update name and the number of dishes with the same name in the menu database in real time, call the corresponding stored dish processing time through the menu update name and integrate the menu update name and the number of dishes with the same name to generate menu information, mark different menu update names respectively to obtain different marked dish names, and pass each marked dish name and the menu information corresponding to the marked dish name to the dish sorting module;

[0036] The menu sorting module is used to receive menu information and set a recognition time. During the recognition time, it processes and sorts the names of different marked dishes according to the menu information and generates sorting features. The sorting features of the names of different marked dishes are passed to the tool screening module.

[0037] The tool screening module counts the usage frequency of different categories of processing tools based on the marked dish names, calculates the usage priority of each category of processing tools based on the sorting characteristics, collects the usage time of all processing tools and calculates the high-frequency use wear rate of each category of processing tools, and comprehensively considers the usage priority and high-frequency use wear rate to screen out the target category of processing tools from all categories of processing tools and input them into the position exchange module;

[0038] The position exchange module detects the tool rack where the target category processing tool is located and marks the position, and replaces the tool rack where the target category processing tool is located with the default common tool rack through a conveyor belt based on the marked position.

[0039] The technical effects and advantages of the present invention are as follows:

[0040] The present invention detects the updated menu names and the number of dishes with the same name in a menu database in real time, calls the corresponding stored dish processing time according to the updated menu name, marks dishes with different names respectively, sets an identification time, calculates the to-be-processed frequency of dishes with different labeled names by using the number of dishes with the same name within the identification time, and sorts the dishes with different labeled names in the updated menu name in combination with the dish processing time, calculates the usage priority of each category of processing tools based on the usage frequency of labeled name dishes, collects the usage time of all processing tools to calculate the high-frequency usage wear rate of each category of processing tools, screens out target category processing tools based on the usage priority and high-frequency usage wear rate of the processing tools, detects the tool rack where the target category processing tool is located and marks its position, and replaces the tool rack where the target category processing tool is located with the default commonly used tool rack through a conveyor belt based on the marked position, thereby achieving real-time production efficiency and quality of dishes in commercial kitchens. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a flow chart of the commercial kitchen design method based on a multimodal large model of the present invention.

[0042] Figure 2 This is a module diagram of the commercial kitchen design system based on the multimodal large model of the present invention. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0044] The present invention detects the updated menu names and the number of dishes with the same name in a menu database in real time, calls the corresponding stored dish processing time according to the updated menu name, marks dishes with different names respectively, sets an identification time, calculates the to-be-processed frequency of dishes with different labeled names by using the number of dishes with the same name within the identification time, and sorts the dishes with different labeled names in the updated menu name in combination with the dish processing time, calculates the usage priority of each category of processing tools based on the usage frequency of labeled name dishes, collects the usage time of all processing tools to calculate the high-frequency usage wear rate of each category of processing tools, screens out target category processing tools based on the usage priority and high-frequency usage wear rate of the processing tools, detects the tool rack where the target category processing tool is located and marks its position, and replaces the tool rack where the target category processing tool is located with the default commonly used tool rack through a conveyor belt based on the marked position, thereby achieving real-time production efficiency and quality of dishes in commercial kitchens.

[0045] Example 1, a commercial kitchen design method based on a multimodal large model, such as Figure 1 As shown, the following steps are included:

[0046] Step S1: Real-time detection of menu update names and the number of dishes with the same name in the menu database, calling the corresponding stored dish processing time according to the menu update name, and marking different menu update names respectively;

[0047] Step S2: setting a recognition time, calculating the processing frequency of different marked dish names using the number of dishes with the same name within the recognition time, and sorting the different marked dish names in the menu database based on the dish processing time of the corresponding marked dish names to generate a sorting feature;

[0048] Step S3: Count the usage frequencies of different categories of processing tools based on the marked dish names, calculate the usage priority of each category of processing tools based on the sorting characteristics, collect the usage time of all processing tools, and calculate the high-frequency use wear rate of each category of processing tools based on the usage time;

[0049] Step S4: Filter out target category processing tools based on the usage priority and high-frequency wear rate of each category of processing tools, detect the tool rack where the target category processing tools are located and mark their positions, and replace the tool rack where the target category processing tools are located with the default commonly used tool rack through a conveyor belt based on the marked positions.

[0050] The specific implementation is as follows:

[0051] In step S1, the user receives the user's selected dish name through the user terminal, and transfers the dish selection name as the menu update name into the menu database. According to the menu update name, all the dish names stored in the menu database are traversed and identified and matched, and the dish processing time of the dish name after identification and matching is called as the dish processing time of the menu update name;

[0052] Count the number of identical menu update names in the menu database in real time, and use the statistical result as the number of dishes with the same name corresponding to the updated menu name;

[0053] The menu update name, the number of dishes with the same name corresponding to the menu update name, and the dish processing time are integrated into the menu information corresponding to the menu update name and temporarily stored in the menu database. When the user's selected dish is completed, the menu information corresponding to the dish update name is deleted from the menu database.

[0054] It should be explained that the menu database is a temporary database for storing and calling menu information. In this example, it is used to call or delete menu information with updated dish names.

[0055] Different menu update names in the menu database are marked respectively to obtain different marked dish names.

[0056] It should be noted that in the process of identifying and matching the updated menu name with all the dish names stored in the menu database, the menu database stores the names of all the dishes that the kitchen can make and the characters of each dish name. Identification and matching are performed by comparing the characters of the updated menu name with the characters of each dish name one by one. If the characters of the updated menu name are the same as the characters of each dish name, the identification and matching is successful.

[0057] In step S2, a period of time is selected as the recognition time, and within the recognition time, the menu database is accessed to call the menu information of each marked dish name, and the number of dishes with the same name as each marked dish name is extracted from the menu information. The ratio of the number of dishes with the same name as each marked dish name in the menu database to the total number of dishes in the menu database is used as the frequency to be processed of the corresponding marked dish name;

[0058] The more dishes with the same name as the marked dish name, the greater the frequency of dishes with the corresponding marked names to be processed, and the more necessary it is to prepare the marked dishes in advance.

[0059] It needs to be explained that since more than one dish can be prepared in one pot during dish preparation, dishes with the same mark can be prepared at the same time. Therefore, the more dishes with the same name are marked, the more it is necessary to prepare the marked dishes in advance, thereby improving the efficiency of dish preparation.

[0060] Extract the dish processing time from the menu information of each marked dish name, and build a logistic regression model based on the pending frequency of the corresponding marked dish name to process and sort the different marked dish names in the menu database. The specific steps are as follows:

[0061] Data processing: The logarithm of the dish processing time of each marked dish name is taken as the dish processing index of the corresponding marked dish name;

[0062] Calculation of intermediate parameters: The ratio of the frequency of each marked dish name to be processed to the dish processing index is used as the intermediate parameter and marked as z;

[0063] Construct a logistic regression model: pass the intermediate parameters into the logistic regression formula: , where L is the logistic regression result of each marked dish name, e is the natural base, and the logistic regression result of each marked dish name is used as the ranking feature of the corresponding marked dish name;

[0064] Processing and sorting: For each marked dish name in the menu database, sort them from large to small according to the sorting characteristics.

[0065] It should be noted that the dish processing time is the time required for dish preparation. The shorter the dish processing time of the marked dish name, the smaller the dish processing index, and the faster the dish processing. Preparing the corresponding dishes in advance can meet the dining needs of some users in advance. The greater the pending frequency of the marked dish name or the smaller the dish processing index, the greater the sorting feature of the marked dish name, and the more necessary it is to prepare the marked dish in advance.

[0066] In step S3, based on the marked dish name, the number of times each category of processing tools is used for each marked dish name is obtained through the dish preparation method database, the number of times each category of processing tools is used for the current marked dish name is accumulated to obtain the total number of times all category processing tools for the current marked dish name are used, and the ratio of the number of times each category of processing tools is used to the total number of times all category processing tools are used is calculated to obtain the usage frequency of the category processing tool corresponding to the current marked dish name;

[0067] The frequency of use of each category of processing tools currently marking the dish name is multiplied by the sorting feature of the marked dish name, and the accumulated result is calculated and used as the use priority of each category of processing tools;

[0068] For example, there are three dish names, A, B, and C, and three types of processing tools, a, b, and c. The corresponding processing tools used in dish name A are category a and category b. Among them, the frequency of use of category a is 80%, and the frequency of use of category b is 20%.

[0069] The corresponding processing tools used in the names of dishes marked with B are category b and category c, among which category b is used 40% of the time and category c is used 60% of the time;

[0070] The corresponding processing tools used in the names of dishes marked with C are category a and category c, among which category a is used at a frequency of 30% and category c is used at a frequency of 70%;

[0071] Then, according to the content in step S2, the sorting features of the dish names can be obtained as follows: the sorting feature value of the dish name marked by A is 5, the sorting feature value of the dish name marked by B is 4, and the sorting feature value of the dish name marked by C is 3. The calculation formula for the use priority of each processing tool is:

[0072] ay=5×80%+3×30%=4.9;

[0073] In the formula, ay is the priority of the a-category processing tool, 80% is the frequency of the a-category processing tool in the A-marked dish name, 5 is the ranking feature value of the A-marked dish name, 3 is the ranking feature value of the C-marked dish name, and 30% is the frequency of the a-category processing tool in the C-marked dish name;

[0074] by=5×20%+4×40%=2.6;

[0075] In the formula, by is the priority of the b-category processing tool, 20% is the frequency of the b-category processing tool in the A-marked dish name, 5 is the ranking feature value of the A-marked dish name, 4 is the ranking feature value of the B-marked dish name, and 40% is the frequency of the b-category processing tool in the B-marked dish name;

[0076] cy=4×60%+3×70%=4.5;

[0077] Where cy is the priority of the c-category processing tool, 60% is the frequency of the c-category processing tool in the B-marked dish name, 4 is the ranking feature value of the B-marked dish name, 3 is the ranking feature value of the C-marked dish name, and 70% is the frequency of the c-category processing tool in the C-marked dish name;

[0078] The usage time of each category of processing tools is obtained by calling the dish preparation method database;

[0079] The usage time of the processing tool is the difference between the current time point and the time point recorded in the dish preparation method database when the corresponding processing tool is stored in the tool rack;

[0080] It is understood that the same category of processing tools includes multiple individual processing tools. When calculating the usage time of each category of processing tools, the average usage time of all individual processing tools in the same category of processing tools is taken;

[0081] It should be noted that the dish preparation method database refers to a database used to record and store the standard preparation steps and process flows corresponding to each marked dish name; the database covers process parameter information including but not limited to the types of processing tools required for each marked dish name in the preparation process, the order of use of each processing tool, the number of times it is used, the current time, and the time point recorded in the dish preparation method database when the corresponding processing tool is stored in the tool rack; its core structure is a multi-dimensional correspondence table based on dish identification and production process, which supports calling and matching by dish name or marked dish name, and outputs the number of times each category of processing tools for each marked dish name is used, as well as the current time point and the time point recorded in the dish preparation method database when the corresponding processing tool is stored in the tool rack;

[0082] The usage time of each category of processing tools is standardized, and the standardized results are used as the high-frequency use wear rate of each category of processing tools;

[0083] It should be noted that the standardization methods include but are not limited to standard linear transformation based on interval scaling, Z-Score standardization method based on statistics, or normalization method based on nonlinear mapping function. The application methods of standardization are not described in detail here.

[0084] In step S4, the use priority of each category of processing tools is standardized so that the use priority of each category of processing tools and the high-frequency use wear rate of each category of processing tools are kept in the same dimension;

[0085] Substitute the high-frequency wear rate of each category of processing tools and the standardized usage priority of each category of processing tools into the product-amplification difference adjustment model to obtain the selection score of each category of processing tools. The specific formula is expressed as follows:

[0086] ;

[0087] Where, Rate the selection of processing tools in each category, To standardize the use priority of each category of processing tools, The wear rate of high-frequency use of each type of processing tools, is the nonlinear response adjustment parameter;

[0088] The selection scores of each category processing tool are sorted from largest to smallest, and the category processing tool corresponding to the maximum selection score of each category processing tool is used as the target category processing tool;

[0089] It should be noted that when the high-frequency wear rate of each category of processing tools is greater, the use priority of each category of processing tools after standardization is lower, and the selection score of each category of processing tools is lower, it means that the category of tools is not suitable for priority selection in the current environment. Conversely, when the high-frequency wear rate of each category of processing tools is lower, the use priority of each category of processing tools after standardization is higher, and the selection score of each category of processing tools is higher, it means that the category of processing tools has high priority and low wear risk, and is suitable for current selection.

[0090] Among them, the product-amplified difference adjustment model is a nonlinear product-type evaluation function, which is used to perform differential amplification processing on the target object under multiple weight indicators;

[0091] Detecting the tool rack where the target category processing tool is located and marking its position to obtain the position mark of each target category processing tool;

[0092] It is understandable that processing tools of the same category may be placed on different tool racks, so that each tool rack can complete the preparation of any marked dish. Furthermore, there is no limitation on the placement of processing tools of the same category and no further description is given here.

[0093] Furthermore, the position of the tool rack corresponding to the position mark of each target category processing tool is set by two-dimensional coordinate system modeling, and the position coordinates of the corresponding position mark of each target category processing tool and the position coordinates of the default common tool rack are set. The distance length between the tool rack where the tool of each target category processing tool is located and the default common tool rack is calculated by the Euclidean distance formula;

[0094] Specifically, let the position coordinates of the tool rack where each target category processing tool is located be , the default location coordinates of the commonly used tool shelf are , i is the tool rack where the processing tool of the i-th target category is located;

[0095] The Euclidean distance formula is used to calculate the distance between the tool shelf where each target category processing tool is located and the default common tool shelf, which can be expressed as follows:

[0096] ;

[0097] Where, The distance between the tool rack where the processing tool of the i-th target category is located and the default common tool rack;

[0098] It should be noted that two-dimensional coordinate system modeling refers to a spatial modeling method that maps each tool rack or transfer point into a two-dimensional spatial coordinate point based on the plane layout of the production area. It will not be described in detail here;

[0099] The distance lengths between the tool rack where each target category processing tool is located and the default commonly used tool rack are sorted in order from small to large, and the tool rack where the target category processing tool is located corresponding to the minimum distance length between the tool rack where the target category processing tool is located and the default commonly used tool rack is selected and replaced with the default commonly used tool rack through a conveyor belt.

[0100] Example 2, a commercial kitchen design system based on a multimodal large model, such as Figure 2 As shown, a commercial kitchen design method based on a multimodal large model is implemented, including a menu update module, a dish sorting module, a tool screening module, and a position exchange module;

[0101] The menu update module is used to detect the menu update name and the number of dishes with the same name in the menu database in real time, call the corresponding stored dish processing time through the menu update name and integrate the menu update name and the number of dishes with the same name to generate menu information, mark different menu update names respectively to obtain different marked dish names, and pass each marked dish name and the menu information corresponding to the marked dish name to the dish sorting module;

[0102] The menu sorting module is used to receive menu information and set a recognition time. During the recognition time, it processes and sorts the names of different marked dishes according to the menu information and generates sorting features. The sorting features of the names of different marked dishes are passed to the tool screening module.

[0103] The tool screening module counts the usage frequency of different categories of processing tools based on the marked dish names, calculates the usage priority of each category of processing tools based on the sorting characteristics, collects the usage time of all processing tools and calculates the high-frequency use wear rate of each category of processing tools, and comprehensively considers the usage priority and high-frequency use wear rate to screen out the target category of processing tools from all categories of processing tools and input them into the position exchange module;

[0104] The position exchange module detects the tool rack where the target category processing tool is located and marks the position, and replaces the tool rack where the target category processing tool is located with the default common tool rack through a conveyor belt based on the marked position.

[0105] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0106] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0107] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0108] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0109] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A commercial kitchen design method based on a multimodal large model, characterized by: The following steps are involved: Step S1: Real-time detection of menu update names and the number of dishes with the same name in the menu database, calling the corresponding stored dish processing time according to the menu update name, and marking different menu update names respectively; Step S2: setting a recognition time, calculating the processing frequency of different marked dish names using the number of dishes with the same name within the recognition time, and sorting the different marked dish names in the menu database based on the dish processing time of the corresponding marked dish names to generate a sorting feature; Step S3: Count the usage frequencies of different categories of processing tools based on the marked dish names, calculate the usage priority of each category of processing tools based on the sorting characteristics, collect the usage time of all processing tools, and calculate the high-frequency use wear rate of each category of processing tools based on the usage time; Step S4: Filter out target category processing tools based on the usage priority and high-frequency wear rate of each category of processing tools, detect the tool rack where the target category processing tools are located and mark their positions, and replace the tool rack where the target category processing tools are located with the default commonly used tool rack through a conveyor belt based on the marked positions.

2. The commercial kitchen design method based on a multimodal large model according to claim 1, characterized in that: In step S1, the user terminal receives the user's selected dish name and transfers it into the menu database as the menu update name. According to the menu update name, all the dish names stored in the menu database are traversed and identified and matched. The dish processing time of the dish name after identification and matching is called as the dish processing time of the menu update name. Count the number of identical menu update names in the menu database in real time, and use the statistical result as the number of dishes with the same name corresponding to the updated menu name; The menu update name, the number of dishes with the same name corresponding to the menu update name, and the dish processing time are integrated to generate menu information corresponding to the menu update name and temporarily store it in the menu database.

3. The commercial kitchen design method based on a multimodal large model according to claim 2, characterized in that: In step S1, when the user's selected dish is completed, the menu information corresponding to the updated dish name is deleted from the menu database, and different menu updated names in the menu database are marked respectively to obtain different marked dish names.

4. The commercial kitchen design method based on a multimodal large model according to claim 2, characterized in that: In step S2, a period of time is selected as the recognition time, and the number of dishes with the same name as each marked dish name is extracted from the menu information within the recognition time, and the ratio of the number of dishes with the same name as each marked dish name in the menu database to the total number of dishes in the menu database is used as the frequency to be processed of the corresponding marked dish name; A logistic regression model is constructed based on the dish processing time of each marked dish name and the frequency of pending processing of the corresponding marked dish name, and the ranking features of each marked dish name are calculated. The different marked dish names in the menu database are processed and ranked based on the ranking features.

5. The commercial kitchen design method based on a multimodal large model according to claim 4 is characterized by: In step S2, a logistic regression model is constructed to calculate the ranking features of each marked dish name. The specific steps are as follows: Data processing: The logarithm of the dish processing time of each marked dish name is taken as the dish processing index of the corresponding marked dish name; Calculation of intermediate parameters: The ratio of the frequency of each marked dish name to be processed to the dish processing index is used as the intermediate parameter and marked as z; Construct a logistic regression model: pass the intermediate parameters into the logistic regression formula: , where L is the logistic regression result of each marked dish name, e is the natural base, and the logistic regression result of each marked dish name is used as the ranking feature of the corresponding marked dish name; Processing and sorting: For each marked dish name in the menu database, sort them from large to small according to the sorting characteristics.

6. The commercial kitchen design method based on a multimodal large model according to claim 3 is characterized by: In step S3, based on the marked dish name, the number of times each category of processing tools is used for each marked dish name is obtained through the dish preparation method database, the number of times each category of processing tools is used for the current marked dish name is accumulated to obtain the total number of times all category processing tools for the current marked dish name are used, and the ratio of the number of times each category of processing tools is used to the total number of times all category processing tools are used is calculated to obtain the usage frequency of the category processing tool corresponding to the current marked dish name; The usage frequency of each category of processing tools for the current marked dish name is multiplied by the sorting feature of the marked dish name, and the accumulated result is calculated, and the accumulated result is used as the usage priority of each category of processing tools.

7. The commercial kitchen design method based on a multimodal large model according to claim 6, characterized in that: In step S3, the usage time of each type of processing tool is obtained by calling the dish preparation method database; The usage time of the processing tool is the difference between the current time point and the time point recorded in the dish preparation method database when the corresponding processing tool is stored in the tool rack; The usage time of each category of processing tools is standardized, and the standardized results are used as the high-frequency use wear rate of each category of processing tools.

8. The commercial kitchen design method based on a multimodal large model according to claim 7 is characterized by: In step S4, the use priority of each category of processing tools is standardized; Substitute the high-frequency wear rate of each category of processing tools and the standardized usage priority of each category of processing tools into the product-amplification difference adjustment model to obtain the selection score of each category of processing tools; The selection scores of each category processing tool are sorted in descending order, and the category processing tool corresponding to the maximum selection score of each category processing tool is used as the target category processing tool.

9. The commercial kitchen design method based on a multimodal large model according to claim 8, characterized in that: In step S4, the tool rack where the target category processing tool is located is detected and the position is marked to obtain the position mark of each target category processing tool; The position of the tool rack corresponding to the position mark of each target category processing tool is set by two-dimensional coordinate system modeling, and the position coordinates of the corresponding position mark of each target category processing tool and the position coordinates of the default common tool rack are set. The distance length between the tool rack where the tool of each target category processing tool is located and the default common tool rack is calculated using the Euclidean distance formula; The distance lengths between the tool rack where each target category processing tool is located and the default commonly used tool rack are sorted in order from small to large, and the tool rack where the target category processing tool corresponding to the minimum distance length is located is selected and replaced with the default commonly used tool rack through a conveyor belt.

10. A commercial kitchen design system based on a multimodal large model, used to implement the commercial kitchen design method based on a multimodal large model according to any one of claims 1 to 9, characterized in that: Including menu update module, dish sorting module, tool screening module and position exchange module; The menu update module is used to detect the menu update name and the number of dishes with the same name in the menu database in real time, call the corresponding stored dish processing time through the menu update name and integrate the menu update name and the number of dishes with the same name to generate menu information, mark different menu update names respectively to obtain different marked dish names, and pass each marked dish name and the menu information corresponding to the marked dish name to the dish sorting module; The menu sorting module is used to receive menu information and set a recognition time. During the recognition time, it processes and sorts the names of different marked dishes according to the menu information and generates sorting features. The sorting features of the names of different marked dishes are passed to the tool screening module. The tool screening module counts the usage frequency of different categories of processing tools based on the marked dish names, calculates the usage priority of each category of processing tools based on the sorting characteristics, collects the usage time of all processing tools and calculates the high-frequency use wear rate of each category of processing tools, and comprehensively considers the usage priority and high-frequency use wear rate to screen out the target category of processing tools from all categories of processing tools and input them into the position exchange module; The position exchange module detects the tool rack where the target category processing tool is located and marks the position, and replaces the tool rack where the target category processing tool is located with the default common tool rack through a conveyor belt based on the marked position.

Images