Intelligent food freshness detection method and system in combination with environmental factors

Abstract

The invention belongs to the technical field of intelligent food freshness detection, and particularly relates to an intelligent food freshness detection method and system in combination with environmental factors. A detection unit carrying a VOCs sensing array and an environment sensing module is placed in a closed storage space of food to be detected; vOCs characteristic signals generated by food spoilage and environmental data such as temperature, humidity and oxygen concentration are synchronously collected, interference of environmental changes on detection signals is counteracted through a dynamic correction model, and a spoilage indicator signal slope and a freshness indicator signal attenuation rate are extracted through first-order difference operation; and substituting into a prediction model formed by training to calculate the dynamic freshness index and the remaining shelf life. During use, the system automatically collects data according to a preset time interval in a time sequence only by completing deployment and starting of the detection unit, outputs the freshness level and the remaining storable time length in real time through the display screen after signal correction, feature extraction and intelligent evaluation, supports wireless uploading of data to a cloud, does not need manual intervention in the whole process, and is adaptive to various food storage scenes.

Description

Technical Field

[0001] This invention belongs to the field of intelligent food freshness detection technology, specifically a method and system for intelligent food freshness detection that combines environmental factors. Background Technology

[0002] Food freshness testing is a key technology for ensuring food safety and reducing resource waste, and it is widely used in all aspects of the food industry chain, including production, storage, logistics, and sales. Existing technologies for food freshness testing mainly include sensory evaluation, physicochemical analysis, and sensor-based rapid detection technologies. Among these, sensor-based detection technology has become the mainstream development direction due to its advantages such as ease of operation, rapid response, and real-time monitoring. This type of technology captures characteristic signals such as volatile organic compounds (VOCs) and pH changes generated during food storage, and combines them with data processing algorithms to achieve qualitative or quantitative judgments of freshness. This provides efficient technical support for food quality control, helps reduce the risk of spoiled food entering the market, protects consumer health, and reduces economic losses caused by food spoilage.

[0003] In real-world applications, food storage environments (such as temperature, humidity, and oxygen concentration) change dynamically. Existing detection technologies largely fail to adequately consider the interference of environmental factors on detection signals, limiting the stability and accuracy of test results. Furthermore, current technologies primarily focus on determining the current freshness of food, making it difficult to accurately predict remaining shelf life and thus unable to provide more guiding decision-making for inventory management and sales scheduling in the food distribution process. Additionally, some detection methods suffer from complex procedures, slow response times, or high equipment costs, limiting their widespread application in small and medium-sized enterprises and end-consumer scenarios. Summary of the Invention

[0004] To address the problems mentioned in the background section, this invention provides an intelligent food freshness detection method and system that incorporates environmental factors, thereby solving the problems of low detection accuracy and shelf-life prediction accuracy under environmental interference.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligent detection of food freshness combining environmental factors, characterized by comprising the following steps: Step S1: Place the volatile organic compound sensor array and the environmental sensor module in the sealed storage space of the food to be tested, and simultaneously collect data at the initial moment. VOCs characteristic signal matrix and environmental factor data set The environmental factor data set Includes temperature T, humidity H, and oxygen concentration. ; Step S2: at time intervals Continuously collect n groups of time-series data to obtain the time-series matrix of VOCs characteristic signals and the time-series data group of environmental factors , where The value range is 1 - 6h; Step S3: Based on the time-series data group of environmental factors Construct a dynamic correction model to perform interference compensation on the time-series matrix of VOCs characteristic signals The correction formula is: ; In the formula, is The corrected VOCs characteristic signal at time , , are the correction coefficients of temperature, humidity, and oxygen concentration respectively, and the value range is 0.02 - 0.15, , , are the environmental reference values at the initial time; Step S4: Perform a first-order difference operation on the corrected time-series matrix of VOCs characteristic signals to extract the slope of the spoilage indicator signal and the decay rate of the freshness indicator signal , and form the feature vector ; Step S5: Input the feature vector F into the preset freshness prediction model to calculate the dynamic freshness index FFI and the remaining shelf life RSL. The prediction model formula is: ; ; In the formula, , are the preset feature thresholds respectively, , are the weight coefficients and , k, m are model parameters obtained by training with historical data of standard food samples; Step S6: When FFI ≤ 0.3, it is determined to be fresh; when 0.3 < FFI ≤ 0.7, it is determined to be sub-fresh; when FFI > 0.7, it is determined to be spoiled, and the remaining shelf life RSL is output synchronously.

[0006] Optionally, the volatile organic compound sensing array includes at least 4 gas-sensitive elements with different sensitive materials, which specifically respond to trimethylamine, ethanol, hydrogen sulfide, and acetaldehyde respectively, and the sensitive material is a metal oxide semiconductor doped with graphene.

[0007] Optionally, the correction coefficient in step S3 , , The method for determining the coefficient is as follows: In a constant temperature and humidity test chamber, the rate of change of VOCs signals under different environmental parameters is tested by the single-factor variable method, and the fitting coefficient is obtained by linear regression analysis.

[0008] Optionally, the training process of the freshness prediction model in step S5 includes: selecting more than three standard food samples, conducting accelerated spoilage experiments under five different environmental conditions, collecting more than 1000 sets of time-series data, and optimizing the model parameters using the gradient descent algorithm. , , k, m, so that the mean square error between the predicted value and the measured value is ≤0.02.

[0009] A smart food freshness detection system incorporating environmental factors, characterized in that it includes: Data acquisition module: includes a volatile organic compound (VOC) sensor array and an environmental sensor module. The VOC sensor array is used to collect VOC characteristic signals, and the environmental sensor module is used to simultaneously collect temperature, humidity, and oxygen concentration data. Signal correction module: connected to the data acquisition module, configured with the dynamic correction model described in step S3 of claim 1, used to compensate for environmental interference of VOCs characteristic signals; Feature extraction module: Connected to the signal correction module, used to perform first-order difference operations and extract the slope of the spoilage indicator signal and the attenuation rate of the freshness indicator signal; Intelligent evaluation module: pre-stores the freshness prediction model described in step S5 of claim 1, receives the feature vector output by the feature extraction module, and calculates the dynamic freshness index and remaining shelf life; Display output module: Connects to the intelligent assessment module and is used to visually display freshness level and remaining shelf life data.

[0010] Optionally, the sampling frequency of the data acquisition module is 0.1-1Hz, the detection limit of VOCs characteristic signals is ≤0.1ppm, the temperature measurement accuracy of the environmental sensing module is ±0.2℃, and the humidity measurement accuracy is ±2%RH.

[0011] Optionally, it also includes a wireless transmission module, which adopts the LoRa communication protocol to upload the collected time-series data to the cloud server in real time, so as to realize multi-node data synchronization management.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention simultaneously acquires VOCs characteristic signals and environmental factor data during food storage, and uses a dynamic correction model to make targeted adjustments to the detection signals, so that the VOCs signals can truly reflect the spoilage process of the food itself, rather than distorted signals interfered with by environmental fluctuations. This improves the stability and reliability of freshness detection results and ensures accurate detection feedback under different storage conditions.

[0013] Based on time-series acquired corrected VOCs signals, the slope of spoilage indicators and the attenuation rate of freshness indicators are extracted through first-order differential operations. This allows for the precise capture of the dynamic changes in food from fresh to spoiled. Combined with a pre-defined freshness prediction model, these characteristic parameters are transformed into intuitive freshness indices and remaining shelf-life data. This process-characteristic-based quantitative assessment method not only clarifies the current freshness status of food but also provides users with practically meaningful storage duration references, facilitating informed decision-making in food distribution, warehousing, and consumption. Attached Figure Description

[0014] Figure 1 This is a flowchart of the overall method in this invention; Figure 2 This is a flowchart of the overall system modules in this invention; Detailed Implementation The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] like Figures 1 to 2 As shown, this invention provides an intelligent detection method for food freshness that incorporates environmental factors, characterized by the following steps: Step S1: Place the volatile organic compound sensor array and the environmental sensor module in the sealed storage space of the food to be tested, and simultaneously collect data at the initial moment. VOCs characteristic signal matrix and environmental factor data set The environmental factor data set Includes temperature T, humidity H, and oxygen concentration. ; Step S2: at time intervals By continuously acquiring n sets of time-series data, the time-series matrix of VOCs characteristic signals can be obtained. and environmental factor time series data set ,in The value ranges from 1 to 6h; Step S3: Based on the environmental factor time series data set Construct a dynamic correction model to perform interference compensation on the VOCs characteristic signal time series matrix The correction formula is as follows: ; In the formula, is the corrected VOCs characteristic signal at time , , are the correction coefficients of temperature, humidity, and oxygen concentration respectively, and the value range is 0.02 - 0.15, , , are the environmental reference values at the initial time; Step S4: Perform a first-order difference operation on the corrected VOCs characteristic signal time series matrix to extract the slope of the spoilage indicator signal and the decay rate of the freshness indicator signal, and form the feature vector ; Step S5: Input the feature vector F into the preset freshness prediction model to calculate the dynamic freshness index FFI and the remaining shelf life RSL. The prediction model formula is as follows: ; ; In the formula, , are the preset feature thresholds respectively, , are the weight coefficients and , k, m are model parameters obtained by training with historical data of standard food samples; Step S6: When FFI ≤ 0.3, it is determined to be fresh; when 0.3 < FFI ≤ 0.7, it is determined to be sub-fresh; when FFI > 0.7, it is determined to be spoiled, and the remaining shelf life RSL is output synchronously.

[0016] The volatile organic compound sensing array includes at least 4 gas-sensitive elements made of different sensitive materials, which specifically respond to trimethylamine, ethanol, hydrogen sulfide, and acetaldehyde respectively. The sensitive material is a metal oxide semiconductor doped with graphene.

[0017] The determination method of the correction coefficients , , in Step S3 is as follows: In a constant temperature and humidity test chamber, the change rate of VOCs signals under different environmental parameters is tested by the single-factor variable method, and the fitting coefficient is obtained through linear regression analysis.

[0018] The training process of the freshness prediction model in step S5 includes: selecting more than three standard food samples, conducting accelerated spoilage experiments under five different environmental conditions, collecting more than 1000 sets of time-series data, and optimizing the model parameters using the gradient descent algorithm. , , k, m, so that the mean square error between the predicted value and the measured value is ≤0.02.

[0019] A smart food freshness detection system incorporating environmental factors, characterized in that it includes: Data acquisition module: includes a volatile organic compound (VOC) sensor array and an environmental sensor module. The VOC sensor array is used to collect VOC characteristic signals, and the environmental sensor module is used to simultaneously collect temperature, humidity, and oxygen concentration data. Signal correction module: connected to the data acquisition module, configured with the dynamic correction model described in step S3 of claim 1, used to compensate for environmental interference of VOCs characteristic signals; Feature extraction module: Connected to the signal correction module, used to perform first-order difference operations and extract the slope of the spoilage indicator signal and the attenuation rate of the freshness indicator signal; Intelligent evaluation module: pre-stores the freshness prediction model described in step S5 of claim 1, receives the feature vector output by the feature extraction module, and calculates the dynamic freshness index and remaining shelf life; Display output module: Connects to the intelligent assessment module and is used to visually display freshness level and remaining shelf life data.

[0020] The data acquisition module has a sampling frequency of 0.1-1Hz, a detection limit of ≤0.1ppm for VOCs characteristic signals, and a temperature measurement accuracy of ±0.2℃ and a humidity measurement accuracy of ±2%RH for the environmental sensing module.

[0021] It also includes a wireless transmission module, which uses the LoRa communication protocol to upload the collected time-series data to the cloud server in real time, enabling multi-node data synchronization management.

[0022] This embodiment uses fresh pork as the food to be tested. Under two typical storage scenarios, namely a home refrigerator (dynamic temperature range of 0-8℃) and a warehouse (10-15℃, humidity 60%-80%), the detection method and system of this invention are used to detect freshness and predict remaining shelf life, verifying the practicality and accuracy of the technical solution.

[0023] S1. Deployment of the detection unit: The VOCs sensor array and environmental sensor module are fixed inside a sealed pork preservation container. The preservation container has micro-ventilation holes with a diameter ≤0.5mm. The preservation container is placed in a household refrigerator compartment (initial temperature 4℃, humidity 70%RH, oxygen concentration 21%vol) and a warehouse (initial temperature 12℃, humidity 75%RH, oxygen concentration 21%vol). The initial VOCs characteristic signal matrix is ​​collected in the first minute after the system is started. =[2.1V,1.8V,0.3V,0.5V] and corresponding environmental factor data sets .

[0024] S2. Time-series data acquisition: Set the time interval Δt=3h, and continuously acquire data for 48 hours to obtain the VOCs characteristic signal time-series matrix X(t) and the environmental factor time-series data set E(t). For example, 9 hours after the refrigerator scenario starts, the environmental data is 5℃, 72%RH, and 20.8%vol, and the corresponding VOCs signal matrix is ​​[2.3V, 1.9V, 0.4V, 0.6V]; 15 hours after the warehouse scenario starts, the environmental data is 13℃, 78%RH, and 20.5%vol, and the corresponding VOCs signal matrix is ​​[2.8V, 2.2V, 0.7V, 0.9V].

[0025] S3. Environmental Interference Correction: The correction coefficients were calibrated using the single-factor variable method in the constant temperature and humidity test chamber. The core parameters are shown in the table below: Based on the above coefficients, interference compensation is performed on the VOCs signal. Taking the data from 9 hours after the refrigerator scenario is started as an example, the correction formula is substituted as follows: = ×[1+0.08×(5-4) / 4+0.05×(72-70) / 70+0.03×(20.8-21) / 21]; The calculated corrected VOCs signal matrix is ​​[2.35V, 1.94V, 0.41V, 0.61V], eliminating environmental fluctuation interference.

[0026] S4. Freshness Feature Extraction: Perform first-order difference operation (Δ) on the corrected VOCs time-series signal. = - ), Take the average slope of the trimethylamine and hydrogen sulfide signals. The average attenuation rate of ethanol and acetaldehyde signals is taken. For example, 24 hours after the refrigerator scene is started, the corrected VOCs signal is [2.7V, 2.1V, 0.6V, 0.8V], and the signal at the previous moment is [2.6V, 2.0V, 0.55V, 0.75V]. The calculated values ​​are... =0.025V / h, =0.3, forming the eigenvector F=[0.025,0.3].

[0027] S5. Dynamic Freshness Assessment: The freshness prediction model is trained and optimized using standard samples. The core parameters and calculation examples are shown in the table below: S6. Result Output: Determine the grade based on the FFI value range (0.3-0.7 is near fresh), and simultaneously display "Near Fresh", "FFI=0.40", and "Remaining Shelf Life: 2.6 days". The data is uploaded to the cloud server via the LoRa module.

[0028] In a refrigerator setting, the actual spoilage time for pork is 7 days, and the predicted deviations for each stage of the remaining shelf life are ≤0.3 days. In a warehouse setting, the actual spoilage time is 3 days, with a prediction deviation of ≤0.2 days, demonstrating good prediction accuracy. After environmental interference correction, the FFI fluctuation amplitude is reduced by more than 60% compared to the uncorrected method, ensuring the stability of the detection results under dynamic conditions.

[0029] The working principle and usage process of this invention are as follows: A detection unit equipped with a VOCs sensor array and an environmental sensor module is placed in a sealed storage space for the food to be tested. Simultaneously, it collects VOCs characteristic signals generated by food spoilage, along with environmental data such as temperature, humidity, and oxygen concentration. A dynamic correction model is used to offset the interference of environmental changes on the detection signals. Then, first-order differential calculations are used to extract the slope of the spoilage indicator signal and the attenuation rate of the freshness indicator signal. These are then substituted into a trained prediction model to calculate the dynamic freshness index and remaining shelf life. In use, only the deployment and startup of the detection unit are required. The system automatically collects data sequentially at preset time intervals. After signal correction, feature extraction, and intelligent evaluation, the freshness level and remaining storage time are displayed in real time on a screen. Data can be wirelessly uploaded to the cloud. The entire process requires no manual intervention and is suitable for various food storage scenarios.

[0030] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0031] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent detection of food freshness combining environmental factors, characterized in that, It includes the following steps: Step S1: Place the volatile organic compound sensor array and the environmental sensor module in the sealed storage space of the food to be tested, and simultaneously collect data at the initial moment. VOCs characteristic signal matrix and environmental factor data set The environmental factor data set Includes temperature T, humidity H, and oxygen concentration. ; Step S2: at time intervals By continuously acquiring n sets of time-series data, the time-series matrix of VOCs characteristic signals can be obtained. and environmental factor time series data set ,in The value ranges from 1 to 6h; Step S3: Based on environmental factor time series data sets Construct a dynamic correction model for the time series matrix of VOCs characteristic signals. Interference compensation is performed using the following correction formula: ; In the formula, for Time-corrected VOCs characteristic signal , , These are correction factors for temperature, humidity, and oxygen concentration, respectively, with values ​​ranging from 0.02 to 0.

15. , , The initial environmental baseline value; Step S4: Perform time-series analysis on the corrected VOCs characteristic signal matrix. Perform first-order difference operations to extract the slope of the putrefaction indicator signal. and the attenuation rate of freshness indicator signals , constitute the feature vector ; Step S5: Input the feature vector F into a preset freshness prediction model to calculate the dynamic freshness index FFI and the remaining shelf life RSL. The formula of the prediction model is: ； ; In the formula, , These are preset feature thresholds, , The weighting coefficients and k and m are model parameters, which are obtained by training with historical data of standard food samples; Step S6: When FFI ≤ 0.3, it is determined to be fresh; when 0.3 < FFI ≤ 0.7, it is determined to be sub-fresh; when FFI > 0.7, it is determined to be spoiled, and the remaining shelf life RSL is output synchronously.

2. The intelligent food freshness detection method combining environmental factors according to claim 1, characterized in that, The volatile organic compound sensing array includes at least 4 gas sensing elements made of different sensitive materials, which specifically respond to trimethylamine, ethanol, hydrogen sulfide, and acetaldehyde respectively. The sensitive material is a metal oxide semiconductor doped with graphene.

3. The intelligent food freshness detection method combining environmental factors according to claim 1, characterized in that, The correction coefficient in step S3 , , The method for determining the coefficient is as follows: In a constant temperature and humidity test chamber, the rate of change of VOCs signals under different environmental parameters is tested by the single-factor variable method, and the fitting coefficient is obtained by linear regression analysis.

4. The intelligent food freshness detection method combining environmental factors according to claim 1, characterized in that, The training process of the freshness prediction model in step S5 includes: selecting more than three standard food samples, conducting accelerated spoilage experiments under five different environmental conditions, collecting more than 1000 sets of time-series data, and optimizing the model parameters using the gradient descent algorithm. , , k, m, so that the mean square error between the predicted value and the measured value is ≤0.

02.

5. A smart food freshness detection system that incorporates environmental factors, characterized in that, It includes: Data acquisition module: It includes a volatile organic compound sensing array and an environmental sensing module. The volatile organic compound sensing array is used to collect VOCs characteristic signals, and the environmental sensing module is used to synchronously collect temperature, humidity, and oxygen concentration data; Signal correction module: Connected to the data acquisition module, configured with the dynamic correction model described in step S3 of claim 1, and used to compensate for environmental interference of the VOCs characteristic signals; Feature extraction module: Connected to the signal correction module, and used to perform first-order difference operations and extract the slope of the spoilage indicator signal and the attenuation rate of the freshness indicator signal; Intelligent evaluation module: Preset the freshness prediction model described in step S5 of claim 1, receive the feature vector output by the feature extraction module, and calculate the dynamic freshness index and the remaining shelf life; Display output module: Connected to the intelligent evaluation module, and used to visually display the freshness level and the remaining shelf life data.

6. The intelligent detection method and system for food freshness based on environmental factors according to claim 5, characterized in that, The sampling frequency of the data acquisition module is 0.1 - 1 Hz, the detection limit of the VOCs characteristic signal is ≤ 0.1 ppm, the temperature measurement accuracy of the environmental sensing module is ±0.2 °C, and the humidity measurement accuracy is ±2%RH.

7. The intelligent food freshness detection method and system combining environmental factors according to claim 5, characterized in that, It further includes a wireless transmission module. The wireless transmission module uses the LoRa communication protocol to upload the collected timing data to the cloud server in real time to achieve multi-node data synchronization management.

Images