Household fermentation intelligent management system and method based on multi-dimensional environment perception and self-adaptive feedback

Abstract

The invention relates to a household fermentation intelligent management system and method based on multi-dimensional environment perception and self-adaptive feedback, discloses a household fermentation intelligent progress management and guidance system, and aims to solve the problems that an existing product is solidified in process, is lack of personalization, is free of closed-loop optimization, is insufficient in safety early warning and the like. And the fermentation success rate and safety are improved through a structure personalized process and self-adaptive learning. The system adapts to various family fermentation scenes such as dough, pickling and dairy products, can dynamically generate a personalized fermentation process according to geographic positions, container information and environmental data, and provides operation reminding, stage guidance and clock-in timing functions; user feedback collection and parameter self-calibration are supported, and fermentation abnormity can be intelligently diagnosed; meanwhile, nitrite content estimation and safety prompt are realized, and intake suggestions are given to high-salt products. According to the method, the traditional experience is converted into a traceable and optimizable digital process with safety early warning, the operation threshold and risk are reduced, and the consistency and safety of the fermentation finished product are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent food processing technology, specifically to an intelligent home fermentation progress management and guidance system. It is applicable to various home fermentation scenarios, including dough products (steamed buns / baozi / bread, etc.), pickled products (pickled vegetables / sauerkraut / pickled radishes, etc.), and dairy products (yogurt / cheese, etc.). It enables refined management of the fermentation process, progress control, food safety risk alerts, and fermentation result analysis and optimization, providing integrated fermentation auxiliary services for home users. Background Technology

[0002] Fermented foods, due to their unique flavor and nutritional properties, have become an important part of family diets in daily meals and handicrafts. A survey on the willingness to make fermented foods at home showed that dough-based foods, pickled foods, and dairy products were the three most popular categories, with 154 votes (29.1%), 124 votes (23.4%), and 115 votes (21.7%) respectively. Among the survey respondents, 75.3% were aged 18-25, and 62.1% were female. They were mainly distributed in third- and fourth-tier cities (31.9%) and rural / township areas (28.9%). This group has some interest in home fermentation, but currently, the control of the home fermentation process still relies on the operator's personal experience, lacking scientific, precise, and convenient auxiliary tools. Based on the survey data and the current state of technology, there are many pain points to be addressed in the home fermentation scenario, specifically as follows:

[0003] First, the low precision in controlling fermentation parameters easily leads to fermentation abnormalities. The success or failure of home fermentation is closely related to key parameters such as fermentation time, temperature and humidity, sealing timing, and venting frequency. However, the parameter requirements vary significantly among different types of fermented foods. According to the survey, "inaccurate control of fermentation time" ranked first among common user problems at 23.7%, followed by "not knowing when to seal / vent," accounting for 15.6%. These two types of problems combined account for over 39%, directly leading to insufficient or excessive fermentation, resulting in poor taste in the finished product (19.2%). For example, in the high temperatures of summer, yeast activity significantly increases during dough fermentation. If the fermentation time is not adjusted in time, problems such as over-fermentation and sour taste can easily occur. For pickled foods like kimchi, improper sealing or insufficient venting can easily lead to the growth of unwanted bacteria and spoilage. Furthermore, most respondents fermented infrequently; 24.7% had never fermented before, and 26.8% only fermented once every few months. This lack of fermentation experience further exacerbates the difficulty in controlling parameters, leading to repeated failures of the same type.

[0004] Secondly, the fermentation process lacks an effective reminder mechanism, making it easy to miss key operational points. Home fermentation is often a phased process, with corresponding steps required at each stage. Research shows that "forgetting steps" is a common user problem, accounting for 6.7%. In the current scenario, operators need to manually memorize the timing of each stage, but with the complexities of daily household tasks, forgetting steps is highly likely. Missing key operational points directly leads to fermentation failure or a decline in quality. Furthermore, the current process lacks advance warnings, preventing operators from preparing in advance and making them prone to missing crucial steps due to unforeseen circumstances. This is also a significant reason for the low frequency of fermentation and low product quality rate among users.

[0005] Third, fermentation experience cannot be iteratively optimized, and troubleshooting the causes of failure is difficult. Fermentation failures are common in home fermentation processes due to various variables, but surveys show that "lack of professional tutorials" accounts for 18.0%, ranking fourth among common user problems. When encountering fermentation failures (such as kimchi being too salty or sour, or dough not rising to the expected size), users cannot accurately pinpoint the cause and often rely on subjective judgment, making it difficult to find the root cause. Furthermore, the existing system lacks a mechanism for recording and providing feedback on the fermentation process and results. Users cannot systematically record the parameters, procedures, and results of each fermentation, nor can they obtain targeted optimization suggestions, leading to repeated occurrences of similar fermentation failures and hindering the accumulation and iteration of fermentation experience.

[0006] Fourth, the lack of safety and health guidelines for fermentation poses a potential food safety hazard. Home fermentation differs fundamentally from factory fermentation. Home fermentation involves a more complex microbial community structure, significantly influenced by raw materials, seasonings, and processing methods. The microbial composition is unknown, environmental conditions are difficult to control, and food safety is challenging to test. [1] Common pathogenic microorganisms found in homemade fermented foods include bacteria such as Salmonella, Escherichia coli, and Bacillus cereus, as well as molds. These microorganisms can ferment and produce harmful toxins, which can lead to food poisoning if ingested. Mild symptoms include vomiting, diarrhea, and fever, while severe cases can be life-threatening. [2] Some fermented foods, such as pickles and fermented bean curd, produce trace amounts of harmful substances (such as nitrites) during fermentation. The content of these substances fluctuates with the fermentation time. If the content of harmful substances exceeds safe limits when consumed, it can negatively impact human health. Surveys show that 16.7% of users are concerned about the safety of fermented foods. However, in current home fermentation settings, operators cannot know the content of harmful substances in the fermented food or the safe consumption amount, and can only rely on experience to judge when and how much to consume, posing a significant safety risk. Furthermore, pickled and other fermented foods are often high in salt; excessive consumption violates dietary guidelines, but currently there is a lack of corresponding health guidelines to guide operators in scientific consumption.

[0007] Fifth, the lack of fermentation equipment and professional guidance, along with gaps in existing technology, hinders the promotion of home fermentation. For example, the limited adoption of home winemaking is related to a lack of brewing knowledge, professional brewing techniques, and wine yeast. Homemade winemaking also faces problems such as unstable raw material supply and improper ingredient addition, resulting in poor quality and taste. [3] Based on survey data, 75.7% of users indicated that solving the aforementioned fermentation-related issues would increase the frequency of fermentation. Regarding tools that can remind users of fermentation progress and provide operation guidelines, 62.2% of users (those who strongly or somewhat need them) explicitly stated their need, while only 4.7% indicated they would not use them at all. In terms of mini-program functional requirements, professional tutorial creation (17.4%), automatic matching of fermentation parameters (16.4%), timed reminders (16.4%), health and safety consumption advice (16.0%), and AI-powered failure analysis (12.2%) are the core user needs. However, currently, there are no multi-functional auxiliary tools specifically designed for home fermentation scenarios. Existing auxiliary methods are mostly single timers or fragmented tutorials, failing to provide integrated services such as fermentation parameter matching, key node reminders, failure cause investigation, health and safety guidance, and experience-based iterative optimization, thus failing to meet users' clearly defined core needs.

[0008] In summary, while users show some interest in home fermentation, the lack of scientific parameter guidance, effective reminder mechanisms, precise failure troubleshooting methods, and health and safety guidelines, coupled with the absence of corresponding integrated auxiliary tools, leads to low fermentation success rates, unstable food quality, and potential food safety hazards. Furthermore, the inability to iteratively optimize fermentation experience results in low fermentation frequency for most users, severely impacting the experience and widespread adoption of home fermented foods. Therefore, developing a home fermentation auxiliary tool that addresses these pain points, is compatible with various home fermentation categories, and provides precise parameter matching, key milestone reminders, failure cause troubleshooting, and health guidance has become an urgent need in the home fermentation field.

[0009] References:

[0010] [1] Liang Huipeng, Zhang Yahao, Chang Cong, et al. Analysis of the differences in bacterial community structure in fermented pickled vegetables from factories and homes in Sichuan Province based on PCR-DGGE and real-time PCR using 16S rRNA gene [C] / / Proceedings of the Symposium on “Technological Innovation Driving the Development of the Food Industry—From Research to Application” and the 2016 Annual Academic Conference of Sichuan Provincial Food Science and Technology Society. 2016:333-343.

[0011] [2] Xie Pan, Zhang Dandan. Microbial safety risk assessment of homemade fermented foods [J]. Modern Food, 2025, (24): 150-152. DOI: 10.16736 / j.cnki.cn41-1434 / ts.2025.24.049.

[0012] [3] Scheklin. The Influence of Maceration Method and Fermentation Tank Volume on the Flavor of Fujiminori Wine and the Design of Home Brewing Equipment [D]. Shanghai University of Applied Technology, 2016.

[0013] [4] Zhang Xueyao. Research on the design innovation of household products for making traditional food [D]. Beijing Institute of Technology, 2018. DOI:10.26948 / d.cnki.gbjlu.2018.001884. Summary of the Invention

[0014] 3.1 Purpose of the Invention

[0015] The purpose of this invention is to overcome the shortcomings of existing technologies in home fermentation, such as high operational barriers, difficulty in parameter control, lack of safety warnings, inability to iterate experience, and lack of integrated auxiliary tools. It provides an intelligent progress management and guidance system for home fermentation. This system adopts an "edge-cloud-end" collaborative architecture, combining multi-dimensional environmental perception, dynamic fermentation models, and adaptive learning mechanisms to achieve personalized customization of the home fermentation process, intelligent guidance throughout the entire process, food safety risk warnings, and continuous optimization of the fermentation model. It also covers multiple categories of home fermentation scenarios, providing users with integrated fermentation support services, improving the success rate of home fermentation, the consistency of finished products, and food safety, meeting core user needs, and promoting the popularization and promotion of home fermented foods.

[0016] 3.2 Technical Solution

[0017] To achieve the aforementioned objectives, Home Fermentation Expert provides an intelligent home fermentation progress management and guidance system. This system employs a "terminal-cloud-edge" collaborative architecture, comprising three layers: a user terminal layer, a network transmission layer, and a cloud processing layer. These layers work together to achieve full functionality, including data collection, model calculation, process push, feedback optimization, and safety alerts. The system also integrates five core functional modules: a parameter acquisition module, a dynamic fermentation module, a food safety risk control module, an intelligent diagnostic module, and a feedback correction module. This forms a closed-loop management system of "parameter acquisition - process generation - operation guidance - feedback analysis - model optimization." The specific technical solution is as follows:

[0018] 3.2.1 System Overall Architecture

[0019] User terminal layer: includes smart mobile terminal and optional environmental sensing unit; the smart mobile terminal runs the "Fermentation Butler" mini-program, responsible for user interaction, image acquisition, and local notification push, supports quick login via WeChat / Alipay, and is compatible with smartphones, tablets, smart kitchen appliance control screens, and other devices with data input, storage, calculation, and information push functions; the environmental sensing unit can connect to temperature and humidity sensors and smart bottle caps (including air pressure / gas sensors) to collect micro-environmental data around the fermentation container. If no hardware is available, it obtains geographic information through GPS positioning and calculates environmental data in conjunction with third-party meteorological APIs.

[0020] Network transmission layer: Enables bidirectional data interaction between the user terminal layer and the cloud processing layer through 4G / 5G / Wi-Fi networks, ensuring the real-time transmission of parameters, feedback information, and fermentation schemes, and ensuring the stability and timeliness of data interaction.

[0021] The cloud processing layer is the core computing and storage module of the system, including business servers, a computing engine center, and a data storage center. The business servers handle user requests, account management, and community data. The computing engine center is the core module, with built-in fermentation kinetics models, image recognition inference engines, and food safety assessment models, responsible for core algorithm calculations. The data storage center stores fermentation knowledge graphs (standard parameters for various foods), user historical data, correction preference models, and standard fermentation parameter databases for various categories. It supports cloud updates and can supplement and update standard process parameters and correction coefficients according to different regions, seasons, and the characteristics of fermented ingredients to ensure the accuracy of the baseline fermentation model.

[0022] 3.2.2 Core Functional Modules

[0023] Parameter acquisition module: Used to acquire basic fermentation parameters, including fermentation category information, user geographical location information, fermentation environment temperature and humidity, fermentation container type and volume, and fermentation start time; it calls a third-party meteorological API through GPS coordinates to obtain local temperature, humidity, and air pressure, and calculates the actual fermentation temperature and humidity by combining indoor environment compensation algorithms; it queries a preset heat transfer coefficient table based on the container material (glass / ceramic / plastic) and volume (L) to generate the container's "temperature lag factor" to solve fermentation deviations caused by slow cooling and heating inside large jars, ensuring the accuracy of fermentation parameters.

[0024] The dynamic fermentation module is responsible for generating dynamic fermentation sequences, rather than fixed timelines. Based on the data from the parameter acquisition module, it constructs a time-series fermentation task according to a pre-defined fermentation parameter model. It uses an accumulated temperature algorithm to calculate fermentation maturity, using the following formula: When the ambient temperature T(t) changes (such as a sudden drop in temperature), the system automatically recalculates the remaining time required and dynamically adjusts the time nodes for "sealing", "venting" or "terminating". At the same time, the fermentation process is defined as a multi-stage state machine consisting of aerobic proliferation period, anaerobic fermentation period and post-ripening period. Different environmental parameter thresholds are matched for each stage to achieve refined control of the fermentation process.

[0025] The food safety risk control module enables food safety risk assessment and health guidance, calculates the content of harmful substances, and provides food safety and dietary health advice. For pickled foods, it has a built-in three-dimensional curved surface nitrite prediction model based on "time-temperature-salinity" to determine in real time whether the fermentation process is in the peak nitrite formation zone or the safe degradation zone. If the user attempts to open the jar in the peak zone (usually on the 3rd to 7th day), the system will forcefully pop up a red alert. At the same time, it has a built-in healthy diet calculator to calculate the sodium content per unit weight of the finished product based on the amount of raw materials input, compare it with the data in the "Chinese Dietary Guidelines", and output personalized consumption suggestions to guide users to consume scientifically.

[0026] The intelligent diagnostic module employs a hybrid architecture of "edge-side data acquisition + cloud-based inference + RAG retrieval-enhanced generation." Utilizing computer vision and AI technologies, it analyzes user feedback after fermentation, providing summary analyses and suggestions. High-resolution photos of the fermented material taken by users, along with contextual information about the fermentation process (such as fermentation type, duration, temperature and humidity, and user sensory evaluations), are packaged into multimodal data and sent to a cloud-deployed Visual Language Model (VLM). The cloud server, through role setting, knowledge injection (RAG), and task instructions, constructs system prompts incorporating domain knowledge. The VLM performs feature extraction and semantic alignment to accurately determine the causes of fermentation failure and push targeted rectification suggestions. The image recognition model and AI fault analysis module are trained using massive amounts of fermentation data to continuously improve the accuracy of fermentation status recognition and fault analysis.

[0027] Feedback and Correction Module: Enables closed-loop iteration of fermentation experience, solves the problem of "pleasing everyone," achieves personalized experience, and better suits user tastes; transforms users' sensory evaluations (acidity, texture, flavor, etc.) into standard values ​​through structured questionnaires, establishes sensory evaluation vectors, and quantifies subjective feelings; based on food science principles, maps sensory dimensions to specific process parameters (such as fermentation time, brine concentration, and ambient temperature), uses an iterative algorithm with damping coefficients to correct the fermentation model, saves user preference offsets, and allows subsequent fermentation tasks to directly call the personalized model, achieving precise adaptation of "one model per person."

[0028] 3.2.3 System Function Implementation Process

[0029] Fermentation task creation: Users select the fermentation category and input basic parameters through the terminal applet. The parameter acquisition module completes data collection and verification, supporting both manual input and selection modes. Fermentation category and container type are drop-down selection options, while temperature and container size are manual input options. The location can be automatically obtained through the terminal's location function, improving the user's ease of operation.

[0030] Fermentation model generation: Based on the fermentation category, the system matches a basic fermentation model in the database, loads the user preference offset parameter, and combines it with the basic parameters input by the user to generate an adapted initial fermentation model.

[0031] Dynamic task construction: The dynamic fermentation module generates a fermentation process flow and operation nodes that include time nodes and corresponding operation requirements based on the fermentation model and initial parameters; then, it periodically detects changes in environmental temperature and humidity through the parameter acquisition module and dynamically adjusts the operation nodes to ensure the adaptability of the fermentation process.

[0032] Reminders and Execution Guidance: The system presets time points before operation nodes (adjusted according to user operating habits), and sends operation reminders to users in the form of terminal pop-ups and message pushes to guide users to complete the corresponding fermentation operations. At the same time, it pushes exclusive fermentation tutorials strongly related to the fermentation category (including video and text tutorials on ingredient processing, operation steps, etc.) and precautions to reduce the probability of operation errors from the source; it supports user check-in operations. After completing the corresponding operation, users can click the "Complete Check-in" button on the system interface to provide operation feedback.

[0033] Anomaly Detection and Risk Assessment: If a user fails to complete the check-in within the specified time, the system will determine it as an anomaly, call the food safety risk control module to calculate the content of toxic substances, and output risk warning information; different levels of anomaly reminders will be triggered according to the overdue time to ensure that users are aware of the risks in a timely manner and can take countermeasures.

[0034] Fermentation Completion and Feedback Collection: After fermentation is complete, the system guides users to fill in a description of the fermentation results and a structured feedback questionnaire, and they can also choose to upload fermentation images. The system stores all feedback information in a structured manner, links it to the ID of this fermentation task, and forms a unique fermentation feedback data archive, providing data support for subsequent model optimization.

[0035] AI Summary, Analysis, and Suggestions: Based on the description of fermentation results in user feedback and the uploaded fermentation images, the system calls the intelligent diagnostic module and uses the Visual Language Model (VLM) to analyze the reasons for failure, providing an evaluation of the fermentation effect, the reasons for failure, and suggestions for subsequent improvements.

[0036] Model correction and learning: Based on the collected structured feedback questionnaires and historical records, the system calls the feedback correction module to correct the user's personalized fermentation model, stores the user preference offset for subsequent fermentation tasks, and realizes continuous iterative optimization of the model.

[0037] 3.3 Beneficial Effects

[0038] The intelligent progress management and guidance system for home fermentation of the present invention solves many technical pain points in the existing home fermentation field through structured processes and adaptive learning mechanisms, and has the following significant beneficial effects:

[0039] Precise parameter personalization: The system can dynamically generate fermentation process based on user's geographical location, container information, and environmental data, automatically match key parameters such as sealing timing and venting frequency, and combine accumulated temperature algorithm and container thermal inertia calculation to effectively solve the problem of inaccurate parameter control, reduce the probability of under-fermentation or over-fermentation, and improve the quality of finished product; at the same time, it accumulates user preferences through feedback correction module to achieve "one model per person" and adapt to the operating habits and taste preferences of different users.

[0040] Intelligent operation guidance throughout the entire process: Covering all scenarios including pre-fermentation preparation, key operation nodes, and timeout warnings, it sets up a dual mechanism of advance reminders and timeout reminders to avoid missing operation nodes due to busy schedules; at the same time, it provides step-by-step production instructions and precautions, supports check-in and timekeeping, ensures that the fermentation process is continuous and controllable, lowers the threshold for home fermentation, adapts to the usage habits of users of different ages, and enhances users' enthusiasm for operation.

[0041] Closed-loop experience iteration and optimization: By collecting user feedback on finished products through structured questionnaires and combining it with AI intelligent diagnosis, the causes of fermentation failure can be accurately identified. The fermentation model is continuously revised based on user feedback, and personalized preference parameters are saved to achieve digital accumulation and iteration of fermentation experience. This solves the pain point of recurring similar failures and helps users gradually improve their fermentation level.

[0042] Dual protection for food safety and health: Built-in predictive models for harmful substances such as nitrites estimate their content in real time and provide safe consumption tips. For high-salt fermented products, personalized consumption suggestions are provided based on the "Dietary Guidelines for Chinese Residents". Food safety risks are avoided in all dimensions from the fermentation process to the consumption stage, ensuring the health of users' diets. At the same time, tiered warnings are triggered for abnormal fermentation conditions, further improving fermentation safety.

[0043] Integrated service for all categories: Adaptable to common household fermentation scenarios such as dough, pickled products, and dairy products, integrating multiple functions such as parameter matching, operation guidance, safety warnings, and AI analysis. Unlike single timers or tutorial tools on the market, it can provide customized support for different categories, meet the personalized needs of multi-category fermentation in household scenarios, and enhance the system's practicality and promotional value.

[0044] The architecture is flexible, easy to operate, and highly scalable: It adopts an edge-cloud-edge collaborative architecture, which supports hardware sensing unit access or pure software operation without the need for additional dedicated equipment; user terminals can log in quickly via WeChat / Alipay, and the operation process is simple; the cloud database supports real-time updates and can supplement and optimize parameters according to regional and seasonal changes; image recognition and AI diagnostic models can continuously improve accuracy through data training, and have strong scalability and practicality.

[0045] This invention transforms traditional fermentation experience into a traceable, optimizable, and safety-alert-enabled digital process, enabling intelligent and standardized management of home fermentation. It effectively improves the success rate and safety of home fermented food production, lowers the operational threshold and food safety risks, and enhances the consistency of fermented products. It meets the needs of 75.7% of users who want to increase fermentation frequency and also satisfies the usage needs of 62.2% of users for related tools, demonstrating broad application scenarios and practical value. Detailed Implementation

[0046] The present invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention. The scope of protection of the present invention is determined by the claims.

[0047] The system of this invention is compatible with all types of home fermented products. The implementation logic of each type is consistent. Only the corresponding benchmark fermentation model and process parameters need to be matched. The following uses homemade kimchi (pickled food) as an example to describe the specific implementation process of the system in detail. At the same time, the system's anomaly handling and model optimization mechanism, as well as the adaptation method for other fermented product types, are explained.

[0048] 4.1 Example 1: Fermentation Management Implementation Process of Pickled Vegetables - Homemade Kimchi

[0049] 4.1.1 System Hardware Basics

[0050] The intelligent home fermentation progress management and guidance system of the present invention runs on a terminal device. The terminal device has a built-in processor and memory. The memory pre-stores a fermentation parameter database, a benchmark fermentation model algorithm, an image recognition model, an AI fault analysis module, and a dietary recommendation database. The processor executes the program instructions in the memory to realize all functions such as fermentation parameter acquisition, model matching, process generation, reminder push, feedback recording, and model optimization. At the same time, the terminal device supports user interaction functions such as image shooting / uploading, text input, and check-in operation, providing hardware support for fermentation progress management.

[0051] 4.1.2 Complete Implementation Process

[0052] In this embodiment, the user is a family residing in Chengdu, Sichuan Province. The fermented product is pickled vegetables (main ingredients are cabbage, radish, and chili peppers). The fermentation container is a 5L glass airtight jar, and the fermentation environment temperature is 22℃ (room temperature). Based on the above basic parameters, the home fermentation progress management method of the present invention is executed. The specific steps are as follows:

[0053] Step 1: Obtain the basic fermentation parameters input by the user.

[0054] Users input basic fermentation parameters through the fermentation management system interface on their terminal devices. The system verifies and stores these parameters. In this embodiment, the parameters obtained are specifically as follows:

[0055] - Fermented products: Pickled products - Kimchi (main ingredients are cabbage, radish, and chili peppers);

[0056] - Fermentation environment parameters: Location: Chengdu, Sichuan Province; Fermentation temperature: 22℃ (room temperature);

[0057] - Fermentation container parameters: Container type is glass airtight jar, container size is 5L.

[0058] The system supports both manual input and selection modes. Fermentation type and container type are drop-down selection options, while temperature and container size are manual input options. The location can be automatically obtained through the terminal's location function, improving the ease of use for users.

[0059] Step 2: Match the baseline fermentation model, generate the fermentation process flow, and push out tutorials / precautions.

[0060] Based on the acquired basic fermentation parameters, the system matches the corresponding kimchi benchmark fermentation model from a pre-set fermentation parameter database.

[0061] 1. The fermentation parameter database stores the standard process parameters for kimchi: the standard filling amount for a 5L container is 80% of the container volume (4L), the standard salinity is 3%, the initial sealed fermentation time is 3 days, the first turning time is on the 2nd day of fermentation, the maturity assessment time is on the 5th day of fermentation, the optimal consumption period is on the 5th-10th day of fermentation, and the fermentation shelf life is 30 days under sealed refrigeration; it also stores environmental parameter correction coefficients and container parameter correction coefficients, among which the sealing correction coefficient for glass containers is 1.0 (no air permeability loss), the temperature correction coefficient for a normal temperature environment of 20-25℃ is 1.0 (standard fermentation temperature range for kimchi), and the regional parameters such as humidity and air pressure in Chengdu have no significant impact on kimchi fermentation, so the regional correction coefficient is 1.0;

[0062] 2. The system combines standard process parameters with various correction coefficients (all 1.0) to generate a kimchi baseline fermentation model adapted to this embodiment. Based on this model, a fermentation process flow including time nodes and corresponding operational requirements is generated. The specific process flow in this embodiment is as follows:

[0063] Table 1. Fermentation Process Flowchart for Homemade Pickles

[0064] Fermentation time nodes Core operational requirements Operating Instructions Day 0 of fermentation (initial stage) Canning + Salting + Sealing Prepare ingredients for a 4L container, add 3% salt, stir well, pour into a glass jar, seal the jar, and check the seal. Fermentation Day 2, 10:00 AM Open the can, stir, and release the gas. Open the jar and gently stir the kimchi ingredients to release fermentation gases. Then reseal the jar. The entire process should take no more than one minute. Day 3 of fermentation, 18:00 Initial status check Observe the clarity of the kimchi broth, taste the acidity / saltiness, and confirm the effectiveness of the primary fermentation. Fermentation Day 5, 10:00 AM Mature tasting Test the fermentation flavor of kimchi to determine if it has reached its optimal consumption period. Fermentation Day 10, 20:00 Shelf life reminder This indicates that the kimchi is nearing the end of its optimal consumption period; it is recommended to consume it as soon as possible or store it in a sealed container in the refrigerator.

[0065] 3. The system also pushes a kimchi-specific fermentation tutorial (including video and text tutorials on ingredient processing, salinity adjustment, and sealing operations) and precautions (such as avoiding oil contamination in the container, avoiding direct sunlight during room temperature fermentation, and preventing contamination by miscellaneous bacteria when turning the container) to the user's terminal. The tutorials and precautions are strongly related to the fermentation category, reducing the probability of operational errors from the source.

[0066] Step 3: Generate and push time-series task reminders, and users provide check-in feedback.

[0067] Based on the aforementioned fermentation process, the system generates time-series task reminders and pushes them to user terminals according to preset rules. Reminder types include advance reminders and immediate reminders. Overdue reminders are triggered for cases of incomplete or missed actions. The reminder push rules and check-in feedback requirements in this embodiment are as follows:

[0068] 1. Turning over the tank on the second day of fermentation: A reminder will be sent 1 hour in advance (9:00 on the second day of fermentation), and an immediate reminder will be sent when the time point arrives (10:00 on the second day of fermentation); if the user has not performed the operation by 10:30, a reminder for overdue operation will be sent every 30 minutes until the user completes the check-in;

[0069] 2. Initial status check on day 3 of fermentation, mature tasting on day 5 of fermentation, etc.: Send a reminder 30 minutes in advance, and send an instant reminder when the milestone is reached;

[0070] 3. All reminders are presented in the form of terminal pop-up windows and message push. After the user completes the corresponding operation, they can click the "Complete Check-in" button on the system interface to provide operation feedback. The system records the check-in time and check-in status (completed / incomplete).

[0071] In this embodiment, the user completes the turning operation and checks in at 10:05 on the second day of fermentation, and checks in and provides feedback at 18:00 on the third day of fermentation and 10:00 on the fifth day of fermentation.

[0072] Step 4: Record the actual operation feedback information submitted by the user.

[0073] While the user is checking in, the system guides the user to submit feedback information on the actual operation. In this embodiment, the feedback information submitted by the user includes text feedback and fermentation status images, specifically:

[0074] 1. Text feedback: The acidity is slightly acidic, the saltiness is moderate, there is no off-odor, the volume does not expand significantly, the flavor intensity is light and fragrant, the user satisfaction score is 9 out of 10, and there is no description of any abnormal fermentation effect.

[0075] 2. Fermentation status images: Users take and upload front and side images of the kimchi glass jar, including the state of the ingredients and the broth layer;

[0076] 3. Check-in feedback information: Check-in time for turning the tank on the second day of fermentation is 10:05 (5 minutes overdue), check-in time for the initial status check is 18:00, and check-in time for mature tasting is 10:00.

[0077] The system stores all the above feedback information in a structured manner, links it to the task ID of this kimchi fermentation, and forms a unique fermentation feedback data archive to provide data support for subsequent model optimization.

[0078] Step 5: Based on feedback from actual operations, optimize and adjust the baseline fermentation model to generate a personalized fermentation model.

[0079] 1. Image recognition and analysis: The system calls the image recognition model to analyze the fermentation state image of kimchi uploaded by the user and extracts the fermentation state characteristics: the kimchi broth is clear and there is no turbidity or sediment, there is no distribution of bacteria or microbial film on the surface of the ingredients, and the pore structure is uniform (fermentation gas is produced normally). The fermentation state is determined to be consistent with the preset characteristics of the benchmark fermentation model.

[0080] 2. Feedback Information Comparison: The system compares the text feedback indicators (acidity, saltiness, flavor intensity, satisfaction, etc.) submitted by users with the preset indicators of the kimchi benchmark fermentation model. In this embodiment, the difference between each indicator and the preset indicator is within a reasonable range of ±5%, with only a deviation in the check-in feedback due to the overdue 5-minute turnover operation.

[0081] 3. Parameter Optimization and Adjustment: Based on the comparison results, the system optimizes the parameters of the benchmark fermentation model according to preset correction rules, generating a personalized kimchi fermentation model adapted to the user. Considering the slight delay of 5 minutes in user operation, the advance reminder time for subsequent fermentation tasks is adjusted from 1 hour to 1 hour and 10 minutes, while retaining the original process parameters such as temperature, container, and salinity. The parameter combination for the user's location in Chengdu, 22℃ room temperature, and 5L glass sealed jar is marked as "high-quality fermentation parameter combination" to improve the subsequent matching priority.

[0082] Step 6: Continuous Iterative Optimization of Personalized Fermentation Model

[0083] Based on the optimized personalized kimchi fermentation model, the system automatically updates the time nodes and operation requirements of subsequent fermentation tasks. In this embodiment, the reminder for the shelf life on the 10th day of fermentation is extended from 30 minutes to 40 minutes in advance to suit the user's operating habits. If the user ferments kimchi again, the system will directly call the personalized fermentation model instead of the initial baseline fermentation model. If the user submits new operation feedback information, the system will continue to iterate and optimize the personalized model based on the new data to achieve personalized fermentation management with "one model per person".

[0084] Step 7: Exception Handling Process (This is a backup process in this embodiment and is not actually triggered)

[0085] The method of this invention includes an exception handling step. If it is detected that the user has not checked in to respond to the task reminder within a preset time, the exception handling process will be triggered. In this embodiment, the user only completed the can-turning operation 5 minutes late, and the system did not trigger a serious exception process. Only a minor non-time operation prompt was displayed on the check-in interface. If the user has not completed the can-turning operation for more than 2 hours, the system will trigger the following in sequence: ① a conspicuous red non-operation prompt; ② an overdue fermentation risk prompt (prompting that gas accumulation in the can may lead to excessive pressure in the sealed can and a sour taste in the kimchi); ③ a food safety warning (if no operation is performed for more than 12 hours, it indicates a risk of contamination by miscellaneous bacteria, and it is recommended to check the fermentation status immediately).

[0086] If a user submits images and descriptions of abnormalities related to fermentation failure (such as moldy kimchi, cloudy broth, or bitter taste), the system will call the AI ​​fault analysis module to determine the cause of fermentation failure (such as low salinity, oil stains on the container, or poor sealing) by combining the image recognition results and the description of abnormalities, and push the cause and solution to the user's terminal.

[0087] Step 8: Sending Metabolic Products and Dietary Recommendations

[0088] In this embodiment, the fermented product is kimchi, which is a high-salt type of fermented product containing specific metabolites. The system executes the steps for metabolites and dietary recommendations:

[0089] 1. Metabolite Concentration Estimation and Intake Recommendations: Based on the fermentation progress of kimchi, the system uses a nitrite prediction model to estimate the concentration of nitrite produced during fermentation in real time and pushes intake recommendations to users: Nitrite concentration is in the rising phase during the first 3 days of fermentation, so it is recommended not to eat it; on the 5th day of fermentation, the nitrite concentration drops below the national standard, so it can be eaten normally; after the 10th day of fermentation, the nitrite concentration continues to decrease, and the taste is better.

[0090] 2. Low-sodium diet recommendations: Kimchi is a high-sodium fermented product. In accordance with the "Dietary Guidelines for Chinese Residents", the system pushes low-sodium intake recommendations to users: It is recommended that each serving not exceed 200g, be paired with light staple foods / vegetables, and reduce the intake of other high-sodium foods on the day. Patients with hypertension and kidney disease are advised to eat it in small amounts or not at all.

[0091] 4.2 Example 2: Implementation and Adaptation Instructions for Other Fermented Products

[0092] The intelligent home fermentation progress management and guidance system of this invention is not only applicable to kimchi. For fermented products such as dough (e.g., 500g of steamed bun dough) and dairy products (e.g., 1L of homemade yogurt), the implementation process is completely consistent with the above-mentioned kimchi example. The core difference lies only in the matching of the fermentation parameter database and the different process parameters of the benchmark fermentation model.

[0093] 1. For steamed bun dough fermentation (dough type, 30℃, 2L ceramic basin): The system's matching baseline fermentation model will include time nodes such as kneading and proofing, primary fermentation, secondary fermentation, and steaming preparation. The operation requirements are that the dough volume expands to 1.5 times and springs back when pressed with a finger. The image recognition features are that the dough has uniform air holes and a smooth surface. The reminder mechanism is adapted to the key nodes of dough fermentation (such as fermentation time adjustment and kneading nodes). The feedback correction module optimizes parameters such as fermentation time and yeast addition amount for sensory evaluations such as taste and fluffiness. There is no nitrite prediction step, and only basic operation safety reminders are provided.

[0094] 2. For homemade yogurt (dairy product, 42℃, 1L glass container): The system's matching baseline fermentation model will include time points such as inoculation, constant temperature fermentation, and refrigeration passivation. The operation requirements are that the yogurt should solidify into a paste without whey separation, and the image recognition features should be that the yogurt has a uniform texture and no spots of miscellaneous bacteria. The metabolic product recommendations are probiotic intake recommendations, and the dietary recommendations are low-fat intake recommendations (if it is whole milk fermentation). The abnormal handling focuses on issues such as contamination by miscellaneous bacteria and insufficient bacterial activity.

[0095] The logic for exception handling, model optimization, and reminder push for each fermentation category is consistent with the kimchi example. The system only needs to call the corresponding preset parameters and function modules according to the fermentation category to achieve standardized and personalized fermentation progress management.

[0096] 4.3 Example 3: Overall System Operation Instructions

[0097] The intelligent home fermentation progress management and guidance system of the present invention is the hardware carrier of the above method. The memory stores all the program instructions and preset databases that implement the above steps. The processor completes the fully automated operation of the process of "data acquisition → model matching → process generation → reminder push → feedback recording → model optimization → suggestion push" by executing the program instructions. At the same time, it supports manual intervention by users (such as modifying the fermentation temperature and adjusting the operation time node), taking into account the needs of automated management and personalized operation.

[0098] The system's fermentation parameter database supports cloud updates, allowing for the addition and updating of standard process parameters and correction coefficients based on different regions, seasons, and the characteristics of fermented ingredients, ensuring the accuracy of the baseline fermentation model. The image recognition model and AI fault analysis module are trained using massive amounts of fermentation data to continuously improve the accuracy of fermentation status recognition and fault analysis. Attached Figure Description

[0099] Figure 1 This is a diagram of the overall system architecture described in this invention;

[0100] Figure 2 This is a flowchart illustrating the collaborative process of the various modules in the system of this invention.

[0101] Figure 3 This is a control logic diagram for the entire fermentation process.

Claims

1. A method for managing the fermentation progress at home, characterized in that, Includes the following steps: Obtain the basic fermentation parameters input by the user, which include at least the fermentation category, fermentation environment parameters, and fermentation container parameters; Based on the aforementioned basic fermentation parameters, the corresponding benchmark fermentation model is matched from the preset fermentation parameter database, and a fermentation process flow containing multiple time nodes and corresponding operational requirements is generated. At the same time, fermentation tutorials and precautions corresponding to the fermentation category are pushed out. According to the fermentation process, a time-series task reminder is generated and pushed to the user terminal. Users who have completed the fermentation operation can check in and provide feedback at the corresponding time node. Record the actual operation feedback information submitted by the user at the stated time point; Based on the actual operation feedback information, the parameters of the benchmark fermentation model are optimized and adjusted to generate a personalized fermentation model adapted to the user.

2. The method for managing home fermentation progress according to claim 1, characterized in that, The fermented products include at least one of pickled products, dough products, and dairy products; the pickled products include at least one of pickled vegetables, pickled cowpeas, pickled radishes, and pickled peppers; the dough products include at least one of steamed bun dough, fermented dumpling dough, and European bread sourdough dough; and the dairy products include at least one of yogurt and cheese.

3. The method for managing home fermentation progress according to claim 1, characterized in that, The fermentation environment parameters include user location information and fermentation temperature; the fermentation container parameters include the type and size of the fermentation container; the fermentation parameter database stores standard process parameters, environmental parameter correction coefficients, and container parameter correction coefficients corresponding to different fermentation categories; the benchmark fermentation model is generated based on the standard process parameters combined with the environmental parameter correction coefficients and container parameter correction coefficients.

4. The method for managing home fermentation progress according to claim 1, characterized in that, The task reminder includes at least one of advance reminder, instant reminder, and overdue non-operation reminder. The advance reminder is a reminder with a preset time before the fermentation operation. The instant reminder is a reminder when the fermentation operation time node arrives. The overdue non-operation reminder is a reminder after the fermentation operation time node has expired.

5. The method for managing home fermentation progress according to claim 1, characterized in that, It also includes an exception handling step: when it is detected that the user has not checked in and responded to the task reminder within a preset time, the exception handling process is triggered; the exception handling process includes at least one of the following: no operation prompt, expired fermentation risk prompt, and food safety warning.

6. The method for managing home fermentation progress according to claim 1, characterized in that, The actual operation feedback information includes text feedback submitted by the user, fermentation status images, and check-in feedback information; the text feedback includes at least acidity, saltiness, volume expansion degree, flavor intensity, and user satisfaction indicators, as well as descriptions of abnormal fermentation effects, including at least one of excessive saltiness, excessive acidity, insufficient fermentation time, dough expansion not meeting standards, and insufficient yeast flavor.

7. The method for managing home fermentation progress according to claim 6, characterized in that, Based on the fermentation state image, the fermentation state characteristics are analyzed using image recognition technology. The state characteristics include at least one of pore structure, surface microbial film distribution, and liquid clarity. When a user submits a fermentation state image and abnormal description related to fermentation failure, the cause of the fermentation failure is determined through AI analysis and pushed to the user terminal.

8. The method for managing home fermentation progress according to claim 1, characterized in that, Based on the optimized and adjusted personalized fermentation model parameters, the time nodes, operation requirements and process flow of subsequent fermentation tasks are automatically updated and adjusted to achieve continuous iterative optimization of the fermentation model.

9. The method for managing home fermentation progress according to claim 1, characterized in that, It also includes steps related to metabolites and dietary recommendations: when the fermented product contains a specific type of metabolite, the concentration of the metabolite is estimated in real time based on the fermentation progress, and relevant intake recommendations for the metabolite are provided to the user; when the fermented product is high in salt, dietary recommendations related to low salt intake are provided to the user in accordance with the Dietary Guidelines for Chinese Residents.

10. The method for managing the progress of home fermentation according to claim 9, characterized in that, The fermented products containing specific types of metabolites include kimchi, and the specific metabolites are trace amounts of harmful substances produced during the fermentation process.

11. The method for managing the progress of home fermentation according to any one of claims 1-10, characterized in that, The optimization and adjustment of the parameters of the benchmark fermentation model specifically involves comparing the fermentation effect evaluation indicators in the actual operation feedback information with the preset indicators of the benchmark fermentation model, and adjusting the time nodes, fermentation environment parameter thresholds, and operation requirements of the benchmark fermentation model based on the comparison difference and preset correction rules.

12. A home fermentation progress management system, characterized in that, The method includes a processor and a memory, the memory storing program instructions; the processor executes the program instructions to implement the home fermentation progress management method as described in any one of claims 1 to 11.

13. A computer-readable storage medium having program instructions stored thereon, characterized in that, When the program instructions are executed by the processor, the home fermentation progress management method as described in any one of claims 1 to 11 is implemented.

Images