Food freshness detection system and method based on gas sensor array

Abstract

The invention relates to the technical field of food detection, in particular to a food freshness detection system and method based on a gas sensor array, and the system comprises the following steps: the gas sensor array is used for detecting gas information of at least one target volatile organic compound in a target sample and converting the gas information into a target voltage signal; target processing is conducted on the target voltage signal through the signal control module to obtain a first processing signal, and finally the food freshness of the target sample is analyzed through the intelligent mode recognition display system on the basis of the first processing signal and a target principal component analysis and multi-layer sensor fusion model. And obtaining a food freshness detection result of the target sample. Therefore, the problems that the food freshness detection result is inaccurate due to the cross sensitivity of sensors in related technologies, the detection range and sensitivity of a single sensor cannot meet the detection requirements in a complex environment, and the accuracy and reliability of food freshness detection are reduced are solved.

Description

Technical Field

[0001] The present application relates to the technical field of food detection, and in particular to a food freshness detection system and method based on a gas sensor array. Background Art

[0002] In recent years, with the development of detection technology, a non-invasive detection method based on gas sensors has been applied to food freshness detection, such as meat freshness detection, and has achieved good detection results. Gas sensors can detect the concentration or type of specific gases and work by converting gas concentrations into electrical signals. However, due to the inherent cross-sensitivity of gas sensors, it is difficult for a single gas sensor to guarantee accuracy when detecting characteristic gases. This limitation makes it difficult to meet practical application requirements in complex environments such as food freshness detection. Therefore, more advanced technologies and methods are needed to improve the accuracy and reliability of gas detection.

[0003] In related technologies, electronic nose systems have been used to detect the freshness of food, such as meat, fruit and vegetables. For example, by integrating multiple (Micro-Electro-Mechanical Systems) MEMS metal oxide gas sensors and combining them with a convolutional neural network model, the freshness of 20 kinds of food, such as meat, fruit and vegetables, was classified with an accuracy of 97.30%; or by using 8 metal oxide gas sensors to detect the freshness of fresh refrigerated chicken and frozen thawed chicken, combined with the K nearest neighbor algorithm for classification research, the average accuracy rates were 95.20% and 94.67% respectively.

[0004] However, due to the cross-sensitivity of sensors in related technologies, the detection results of food freshness are inaccurate, and the detection range and sensitivity of a single sensor cannot meet the detection requirements in complex environments, reducing the accuracy and reliability of food freshness detection, which needs to be solved urgently. Summary of the invention

[0005] This application is based on the following problems and understandings made by the inventor:

[0006] In modern food, meat has become an indispensable and important component due to its rich dietary protein, high bioavailability, and essential minerals and vitamins. Due to its high water content, easy corruption, and susceptibility to microbial contamination, if meat is not stored in a timely and appropriate manner, its quality will deteriorate rapidly, causing serious safety hazards and food waste. A common cause of meat loss or waste is spoilage during storage, and meat spoilage is essentially a biochemical change process in which carbohydrates, proteins, fats, etc. in meat are metabolically decomposed by microorganisms or by their own tissue enzymes. During storage, the concentration of organic compounds such as ammonia, hydrogen sulfide, alcohols, ketones, and aldehydes produced by meat is closely related to the degree of corruption, so their detection can be used to determine the freshness of meat.

[0007] In traditional food freshness testing, such as meat freshness testing, sensory evaluation, a series of physical and chemical testing methods and microbial counting methods are usually used.

[0008] Among them, sensory evaluation, as a traditional method for detecting the freshness of food, relies on the human perception system, covering vision, smell and touch to evaluate the quality of food samples. For example, in this process, sensory assessors play a key role. They need to accurately take the right amount of meat samples and place them in a clean white porcelain plate, while ensuring the cleanliness of the observation environment and the representativeness of the samples. Compared with other methods for evaluating the freshness of food, sensory evaluation shows its uniqueness in many aspects. It does not rely on complex instruments and equipment, so it is simple to operate and easy to carry, making it possible to conduct rapid on-site evaluation without complex sample processing and instrument preparation. However, since sensory evaluation relies on human perception, it requires high professional quality of assessors. Assessors need to have rich experience, keen perception and accurate judgment to ensure the accuracy and reliability of the evaluation results. At the same time, the evaluation results are easily affected by the personal status of the assessors, resulting in subjective and uncertain results. Therefore, although sensory evaluation has certain advantages in the evaluation of food freshness, it is also necessary to be aware of its limitations and potential influencing factors.

[0009] Physical and chemical testing methods use two major methods, chemical testing and physical testing, to determine whether food is spoiled. For example, chemical testing focuses on specific chemical substances produced during the spoilage process of meat, such as volatile basic nitrogen and volatile gas components, and infers the freshness of meat by measuring the changes in the content of these substances. Physical testing focuses on changes in the physical properties of meat, such as measuring parameters such as the conductivity, texture, water holding capacity and pH (pH) value of meat, to indirectly infer its freshness. Changes in these physical parameters are closely related to the meat spoilage process. However, when the meat is not obviously spoiled and the physical parameters change slightly, it is difficult to measure accurately, the accuracy of the measurement system is difficult to meet the requirements, and these parameters need to be manually tested regularly. Often when spoilage is discovered, the meat has already deteriorated and monitoring is not timely.

[0010] Microbiological detection method evaluates the safety and hygiene status and degree of spoilage of food by quantitatively and qualitatively analyzing the number, type and growth status of microorganisms in food. Common microbiological detection technologies include biochip technology, immunological detection technology, molecular biology technology and gene probe technology, which are widely used to detect the total number of bacteria and Escherichia coli in food. However, in practical applications, microbial detection has limitations. First, its selective culture process requires high operating skills and is time-consuming. It usually takes more than 24 hours of culture time, which increases the cost of detection. Secondly, the ambient temperature and humidity have a significant impact on the microbial detection experiment process, increasing the uncertainty and volatility of the experimental results. In addition, microbial detection cannot achieve non-destructive testing of food and is destructive to samples. Therefore, although microbial detection has unique advantages in food safety assessment, its operational difficulty, time-consuming, cost and destructiveness to samples limit it in certain specific scenarios.

[0011] In summary, among the traditional food freshness detection methods, sensory evaluation is intuitive, but it is labor-intensive and susceptible to subjective bias; while physical and chemical detection methods and microbial counting methods are limited by technical supervision, high cost and long detection cycle. With the growing demand for continuous monitoring of food quality, it is particularly important to develop a portable, rapid, non-invasive and simple analytical detection technology.

[0012] In recent years, with the development of detection technology, a non-invasive detection method based on gas sensors has been applied to food freshness detection and has achieved good detection results. Gas sensors can detect the concentration or type of specific gases and work by converting gas concentrations into electrical signals. However, due to the inherent cross-sensitivity of gas sensors, it is difficult for a single gas sensor to guarantee accuracy when detecting characteristic gases. This limitation makes it difficult to meet practical application needs in complex environments such as food freshness detection, and it is urgent to improve.

[0013] The present application provides a food freshness detection system and method based on a gas sensor array to solve the problems that the cross-sensitivity of sensors in the related art leads to inaccurate detection results of food freshness, and the detection range and sensitivity of a single sensor cannot meet the detection requirements in a complex environment, thereby reducing the accuracy and reliability of food freshness detection.

[0014] The first aspect of the present application provides a food freshness detection system based on a gas sensor array, comprising: a gas sensor array, the gas sensor array being used to detect gas information of at least one target volatile organic compound in a target sample, and converting the gas information into a target voltage signal; a signal control module, the signal control module being used to receive the target voltage signal sent by the gas sensor array, and performing target processing on the target voltage signal to obtain a first processed signal; an intelligent pattern recognition and display system, the intelligent pattern recognition and display system being used to receive the first processed signal sent by the signal control module, and analyzing the food freshness of the target sample based on the first processed signal and a fusion model of target principal component analysis and multilayer perceptron to obtain a food freshness detection result of the target sample.

[0015] Optionally, in one embodiment of the present application, the gas sensor array includes: a plurality of micro-electromechanical system (MEMS) gas sensors, wherein each MEMS gas sensor is used to detect the composition information and concentration information of at least one target volatile organic compound in the target sample, and convert the composition information and the concentration information into the target voltage signal.

[0016] Optionally, in one embodiment of the present application, the signal control module includes: a power supply unit, used to power the gas sensor array, the control unit and the intelligent pattern recognition and display system; a control unit, used to receive the target voltage signal, and preliminarily process the target voltage signal to obtain the first processed signal, and transmit the first processed signal to the intelligent pattern recognition and display system.

[0017] Optionally, in one embodiment of the present application, the intelligent pattern recognition and display system includes: a data communication unit, used to receive the first processed signal sent by the signal control module, and perform data preprocessing and feature extraction on the first processed signal to obtain a second processed signal of the target sample; a data analysis and display unit, used to receive the second processed signal of the target sample, and analyze the food freshness of the target sample based on the second processed signal and a fusion model of target principal component analysis and multilayer perceptron to obtain a food freshness detection result of the target sample, and display the food freshness detection result on a preset terminal.

[0018] Optionally, in one embodiment of the present application, it also includes: a test module, which is used to control the alternating collection of sample gas and baseline gas using a fan-shaped coupling structure before detecting the gas information of at least one target volatile organic compound in the target sample, so as to obtain a test environment that meets preset test conditions.

[0019] The second aspect of the present application provides a food freshness detection method based on a gas sensor array, using the food freshness detection system based on a gas sensor array as described above, wherein the method includes the following steps: detecting gas information of at least one target volatile organic compound in a target sample, and converting the gas information into a target voltage signal; performing target processing on the target voltage signal to obtain a first processed signal; and analyzing the food freshness of the target sample based on the first processed signal and a fusion model of target principal component analysis and a multi-layer perceptron to obtain a food freshness detection result of the target sample.

[0020] Optionally, in one embodiment of the present application, the food freshness of the target sample is analyzed based on the first processed signal and the fusion model of target principal component analysis and multi-layer perceptron to obtain the food freshness detection result of the target sample, including: performing data preprocessing and feature extraction on the first processed signal to obtain the second processed signal of the target sample; analyzing the food freshness of the target sample according to the second processed signal and the fusion model of target principal component analysis and multi-layer perceptron to obtain the food freshness detection result of the target sample, and displaying the food freshness detection result on a preset terminal.

[0021] Optionally, in one embodiment of the present application, before detecting the gas information of at least one target volatile organic compound in the target sample, it also includes: using a fan-shaped coupling structure to control the alternating collection of sample gas and baseline gas to obtain a test environment that meets preset test conditions.

[0022] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the food freshness detection method based on the gas sensor array as described in the above embodiment.

[0023] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned food freshness detection method based on a gas sensor array.

[0024] The fifth aspect of the present application provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned food freshness detection method based on the gas sensor array.

[0025] The embodiment of the present application can detect the gas information of at least one target volatile organic compound in the target sample through the gas sensor array, and convert the gas information into a target voltage signal, and then receive the target voltage signal sent by the gas sensor array through the signal control module, and target the target voltage signal to obtain a first processing signal, and finally analyze the food freshness of the target sample through the intelligent pattern recognition display system based on the first processing signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample, which effectively improves the accuracy and reliability of food freshness detection. Thus, the problems of inaccurate food freshness detection results due to the cross-sensitivity of sensors in the related art, and the inability of the detection range and sensitivity of a single sensor to meet the detection requirements in a complex environment, reducing the accuracy and reliability of food freshness detection, etc. are solved.

[0026] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0028] Figure 1 A schematic diagram of the structure of a food freshness detection system based on a gas sensor array provided according to an embodiment of the present application;

[0029] Figure 2 A schematic diagram of an electronic nose system according to a specific embodiment of the present application;

[0030] Figure 3 A schematic diagram of a MEMS gas sensor circuit according to a specific embodiment of the present application;

[0031] Figure 4 This is a schematic diagram of TVB-N (Total Volatile Base Nitrogen) technical test results for a specific embodiment of the present application;

[0032] Figure 5 This is a schematic diagram of GC-MS (Gas Chromatography-Mass Spectrometry) spectrum results of volatile compounds in pork under different storage conditions according to a specific embodiment of the present application;

[0033] Figure 6 This is a partial data preprocessing effect diagram of a specific embodiment of the present application;

[0034] Figure 7 This is a structural diagram of an intelligent pattern recognition display system according to a specific embodiment of the present application;

[0035] Figure 8 A schematic diagram of a grid search technology for a specific embodiment of the present application;

[0036] Fig. 9 This is a schematic diagram of PCA (Principal Component Analysis) analysis results of storage at different temperatures and time periods according to a specific embodiment of the present application;

[0037] Fig.10 This is a schematic diagram of the contribution rate of each sensor to PC1 and PC2 under storage at 25°C and 4°C in a specific embodiment of the present application;

[0038] Fig.11 This is an LDA (Linear Discriminant Analysis) diagram of storage at different temperatures and time periods in a specific embodiment of the present application;

[0039] Fig.12 This is a schematic diagram of the results of a ten-fold cross validation of a random forest model according to a specific embodiment of the present application;

[0040] Fig.13 This is a schematic diagram of the results of a ten-fold cross validation of an XGBoost model according to a specific embodiment of the present application;

[0041] Fig.14 A schematic diagram of the results of a ten-fold cross validation of an MLP (Multilayer Perceptron) model according to a specific embodiment of the present application;

[0042] Fig.15 A schematic diagram of a Scree Plot of storage at different temperatures and time periods according to a specific embodiment of the present application;

[0043] Fig.16 This is a schematic diagram of the results of a ten-fold cross validation of a fusion algorithm model according to a specific embodiment of the present application;

[0044] Fig.17 This is a test diagram of an electronic nose system according to a specific embodiment of the present application;

[0045] Fig.18 This is a radar graph of the electronic nose response stored at 25°C and 4°C for different time periods according to a specific embodiment of the present application;

[0046] Fig.19 This is a schematic diagram of the temperature and humidity test results of an electronic nose stored in different time periods according to a specific embodiment of the present application;

[0047] Fig. 20 This is a working flow chart of a food freshness detection system based on a gas sensor array according to a specific embodiment of the present application;

[0048] Fig.21 A flow chart of a food freshness detection method based on a gas sensor array provided according to an embodiment of the present application;

[0049] Fig. 22 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0050] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0051] The following describes a schematic structural diagram of a food freshness detection system based on a gas sensor array according to an embodiment of the present application with reference to the accompanying drawings. Figure 1 It is a structural schematic diagram of a food freshness detection system based on a gas sensor array according to an embodiment of the present application.

[0052] like Figure 1 As shown, the food freshness detection system 10 based on the gas sensor array includes: a gas sensor array 100 , a signal control module 200 and an intelligent pattern recognition and display system 300 .

[0053] Specifically, the gas sensor array 100 is used to detect gas information of at least one target volatile organic compound in a target sample and convert the gas information into a target voltage signal.

[0054] In the embodiment of the present application, the target sample is the food to be tested, which may be meat, vegetables, fruits, etc. The embodiment of the present application will be described in detail using meat as an example.

[0055] In the actual implementation process, the embodiment of the present application can set the gas sensor array 100 in the following steps. For example, the gas sensor array 100 combines multiple types of sensors, each sensor has different response characteristics to different VOCs (Volatile Organic Compounds). Therefore, the embodiment of the present application can detect the gas information of at least one target volatile organic compound in the target sample through the gas sensor array 100, and convert the gas information into a target voltage signal; in addition, the embodiment of the present application can also detect the gas information of multiple target volatile organic compounds in the target sample through the gas sensor array 100, effectively eliminating the inherent cross-sensitivity of the gas sensor, and improving the applicability and accuracy of food detection.

[0056] It should be noted that as meat spoils, specific VOCs will be produced. Volatile organic compounds refer to any carbon compounds that participate in atmospheric photochemical reactions, except carbon monoxide, carbon dioxide, carbonic acid, metal carbides, metal carbonates and ammonium carbonate, such as ammonia, hydrogen sulfide, alcohols, ketones and aldehydes. These carbon compounds are key indicators for assessing the freshness of meat.

[0057] Optionally, in one embodiment of the present application, the gas sensor array 100 includes:

[0058] A plurality of micro-electromechanical system (MEMS) gas sensors, wherein each MEMS gas sensor is used to detect the composition information and concentration information of at least one target volatile organic compound in a target sample, and convert the composition information and concentration information into a target voltage signal.

[0059] As a possible implementation method, the gas sensor array 100 can be composed of multiple MEMS gas sensors, wherein the MEMS gas sensor has the characteristics of high detection accuracy and low interference, and each MEMS gas sensor has high sensitivity and selectivity for specific volatile organic compounds to reduce cross-sensitivity problems. Each MEMS gas sensor plays a specific role in the array, and together they realize comprehensive monitoring of meat freshness and can fully capture a variety of volatile organic compounds generated during meat spoilage; for example, Figure 2 As shown, the embodiment of the present application can detect gas information of specific volatile organic compounds in meat, such as ammonia concentration, hydrogen sulfide concentration, alcohol concentration, ketone concentration or aldehyde concentration, through different MEMS gas sensors in the gas sensor array 100, and convert the component information and concentration information into a target voltage signal, thereby effectively improving the feasibility of food freshness detection.

[0060] like Figure 3As shown in the figure, the circuit design of a gas sensor array module is shown, including key circuit components such as the sensor's output voltage (Vout), load resistance (RL), analog signal output (A), ground (GND), heating voltage (Vh) and test voltage (Vc). These components together constitute the working circuit of the gas sensor, which can be used to detect and respond to specific gases.

[0061] It should be noted that in order to adapt to different application scenarios, especially for the detection of volatile organic compounds in different foods, such as meat, fruits, vegetables, etc., the system design allows flexible replacement of sensors in the gas sensor array. Therefore, those skilled in the art can selectively replace certain gas sensors in the array according to the specific food to be measured, thereby achieving efficient and accurate monitoring of volatile organic compounds in different meats.

[0062] For example, in the meat freshness assessment, TVB-N is a key chemical indicator that is proportional to the degree of meat spoilage. The food freshness detection system, also known as the electronic nose system, can simulate the human sense of smell and use a gas sensor array to analyze meat volatile gases to achieve rapid and non-destructive freshness detection. The accurate determination of TVB-N is crucial for calibrating electronic nose sensors, providing a reference standard for electronic noses to ensure the accuracy and consistency of freshness determination, which is of great value to the basic application of electronic noses in meat quality analysis.

[0063] First, in order to verify the accuracy of the data, the TVB-N technical test can be performed twice on meat samples under the same storage conditions, and the average value is used as the final result. In the test, the automatic Kjeldahl method is used to determine the TVB-N content in the meat samples. In a sterile environment, 10±0.1 grams of uniformly processed pork samples were accurately weighed and placed in a distillation device. Then, 75 ml of pure water was added to the device, and the sample was soaked for 30 minutes, and shaken every 10 minutes during the soaking process to ensure that the sample and water are fully mixed. Before the formal test, the instrument was cleaned and pre-run according to the operating manual of the automatic Kjeldahl nitrogen analyzer to ensure the normal working condition of the equipment. First, a reagent blank test was performed to obtain the background value, and then 1 gram of magnesium oxide was added to the distillation device containing the sample, and it was quickly connected to the distiller, and the distillation process was started according to the preset program. The content of TVB-N in the sample is calculated according to the following formula:

[0064]

[0065] Among them, V 1 is the volume of hydrochloric acid standard titration solution consumed by the target test solution (ml), V 2is the volume of standard hydrochloric acid titration solution consumed by the blank reagent (ml), c is the concentration of standard hydrochloric acid titration solution (mol / L), and m is the mass of the sample (g).

[0066] Secondly, China's national food safety standards clearly define the physical and chemical requirements for fresh meat, including the upper limit of volatile basic nitrogen content of 15mg / 100g. This standard provides a quantitative basis for judging whether meat is fresh. Once the TVB-N value in a meat sample exceeds the limit specified in the above standard, it usually means that the meat has deteriorated and its safety for consumption is no longer guaranteed.

[0067] like Figure 4 As shown, the test results show that the freshness of meat samples at different storage temperatures changes over time as follows:

[0068] 1. Storage conditions at 25℃:

[0069] 0 to 6 hours: The volatile basic nitrogen content increases slowly and the meat remains fresh.

[0070] 6 to 12 hours: The content increases, the freshness begins to decrease, and reaches a sub-fresh state.

[0071] After 12 hours: the content rises sharply, and after 18 hours it exceeds the spoilage limit and the meat is spoiled.

[0072] 2. 4℃ Storage Conditions:

[0073] 0 to 1 day: The volatile basic nitrogen content increases significantly and the meat remains fresh.

[0074] 1 to 3 days: The growth rate slows down, but the freshness begins to decrease, reaching a semi-fresh state.

[0075] After 3 days: the content increases sharply, and after 4 days it exceeds the spoilage limit and the meat becomes spoiled.

[0076] Among them, Table 1 is the sample TVB-N technical test result table. By further analyzing the data in Table 1 below, the change in the freshness of meat under different storage temperatures and times can be clearly seen. The specific Table 1 is as follows:

[0077] Table 1

[0078]

[0079] The TVB-N technical test results show that storage temperature and time have a significant impact on the freshness of meat. At 25°C, meat spoils quickly; at 4°C, although the spoilage rate is relatively slow, long-term storage will still cause meat spoilage. Therefore, during the storage and transportation of meat, temperature and time should be strictly controlled to ensure the quality and safety of meat.

[0080] Furthermore, the embodiment of the present application can further track the freshness and deterioration process of food, and understand the degree of food corruption through HS-SPME GC-MS technology, wherein HS-SPME GC-MS technology is a highly sensitive analytical method that combines HS-SPME (Solid-Phase Microextraction) and GC-MS (GasChromatography-Mass Spectrometry). This method is capable of detecting extremely low concentrations of volatile organic compounds in food. HS-SPME adsorbs volatile components in samples through fused silica fibers, and then these components are separated and analyzed in a GC-MS system. This technology can sample in real time, quickly capture changes in volatile components of food, and accurately identify and quantitatively analyze different volatile components in samples through a GC-MS system. By comparing mass spectrometry information, the chemical structure can be determined, and the peak area or height is used for quantitative analysis.

[0081] First, HS-SPME GC-MS was used to analyze the volatile components in the samples. The specific steps are as follows:

[0082] (1) Sample preparation: Place the sample in a specially designed sealed air bag to ensure that the headspace gas can be effectively extracted.

[0083] (2) Headspace extraction: Under controlled conditions, a sealed air bag containing food is placed in a constant temperature incubator and the temperature is set at 40°C. Extraction is performed under this condition for 30 minutes so that the volatile components can reach equilibrium in the headspace of the air bag.

[0084] (3) SPME pretreatment: Before using the solid phase microextraction fiber, it was placed in an environment of 250°C for 30 min to improve the extraction efficiency and remove possible contaminants.

[0085] (4) Gas chromatography-mass spectrometry analysis: The chromatographic column model used was DB-5. The temperature program for the chromatographic analysis started at 40°C and was maintained for 3 min, then increased to 230°C at a rate of 5°C per minute and maintained at this temperature for 5 min. The entire analysis process lasted 43 min. The temperature of the injection port was maintained at 250°C, and the flow rate of the carrier gas (usually helium) was set to 1 mL / min.

[0086] (5) Mass spectrometry detection conditions: The ion source temperature was maintained at 250°C, and the interface temperature was also 250°C. The ionization method was electron impact (EI), the electron energy was set to 70 eV, and automatic tuning was performed. A full scan was performed in SCAN mode to obtain a complete mass spectrum.

[0087] (6) SPME injection and cleaning: The SPME fiber was used to analyze volatile components in the injection port for 1 min. To ensure that there was no cross-contamination between consecutive measurements, the extraction fiber was cleaned in the injection port at 250 °C for 10 min after each analysis to ensure complete desorption of the residues from the previous sample.

[0088] By using the above method, volatile components can be effectively extracted from food samples and accurately identified and quantitatively analyzed by GC-MS.

[0089] Next, the volatile gases of pork under different storage conditions can be analyzed by HS-SPME GC-MS technology. The GC-MS spectrum results are as follows: Figure 5 As shown. There is a clear separation between the chromatographic peaks representing volatile compounds in the sample. The positions of the peaks vary under different storage conditions and time periods. After qualitative and comparative analysis of the HS-GCIMS database, 52 compounds were identified, including alkanes (18), esters (9), benzenes (6), alcohols (5), alkenes (5), ketones (3), acids (3), aromatic hydrocarbons (1), sulfurs (1) and aldehydes (1).

[0090] Among them, microorganisms can decompose fatty acids during the spoilage process to produce a series of volatile compounds, including saturated alkanes, such as n-undecane, n-dodecane, etc. These alkane gases are usually the products of fatty acid decomposition; esterification is the process of producing odor through the esterification reaction of alcohol and carboxylic acid. Esters with short-chain fatty acids may have a pleasant odor to a certain extent, while esters with long chains will produce fat odor; during the spoilage process of meat, its proteins and amino acids are decomposed by microorganisms to produce benzene and acid substances; ketones may be related to microbial spoilage caused by oil or protein oxidation; aldehydes are produced by fat decomposition, secondary alcohol dehydrogenation or microbial alkane degradation, which will produce strong acidic and fatty taste in meat; alcohols are produced by microbial metabolism of amino acid decarboxylation and dehydrogenation.

[0091] The signal control module 200 is used to receive the target voltage signal sent by the gas sensor array 100, and perform target processing on the target voltage signal to obtain a first processed signal.

[0092] In some embodiments, the embodiments of the present application may set up a signal control module 200, which can receive the target voltage signal sent by the gas sensor array 100 and perform target processing on the target voltage signal, for example, performing signal denoising processing and signal normalization processing on the target voltage signal to obtain a first processed signal, thereby effectively improving the accuracy of food freshness detection.

[0093] Optionally, in one embodiment of the present application, the signal control module 200 includes:

[0094] A power supply unit is used to supply power to the gas sensor array 100 , the control unit and the intelligent pattern recognition and display system 300 .

[0095] In some embodiments, the power supply unit in the signal control module 200 can provide power 300 for the gas sensor array 100, the control unit and the intelligent pattern recognition and display system. That is, it can provide a stable power supply for each component of the entire food freshness detection system, thereby effectively improving the stability of the food freshness detection system.

[0096] The control unit is used to receive the target voltage signal, preliminarily process the target voltage signal to obtain a first processed signal, and transmit the first processed signal to the intelligent pattern recognition display system 300 .

[0097] In actual implementation, the control unit in the signal control module 200 may receive the electrical signal and perform preliminary processing on the electrical signal, for example, Figure 2 As shown, the STM32 single-chip microcomputer control unit can receive electrical signals, and the single-chip microcomputer can amplify and filter the electrical signals, and use an A / D converter to convert the analog signal into a digital signal to obtain a first processed signal, and send the first processed signal to the intelligent pattern recognition display system 300 through a USB interface, thereby ensuring the accuracy of data transmission.

[0098] The intelligent pattern recognition display system 300 is used to receive the first processed signal sent by the signal control module 200, and analyze the food freshness of the target sample based on the first processed signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample.

[0099] As a possible implementation method, the embodiment of the present application can set up an intelligent pattern recognition and display system 300, which can receive the first processed signal sent by the signal control module 200, and based on a multi-pattern recognition algorithm, such as a PCA-MLP fusion model, use the first processed signal to determine the food freshness of the target sample. For example, the food freshness can be divided into fresh, that is, the food is in the best condition and has no obvious signs of corruption; semi-fresh, that is, the food begins to show slight signs of corruption but is still edible; not fresh, that is, the food is obviously corrupt and not suitable for consumption, etc., effectively improving the classification and recognition accuracy of food freshness.

[0100] Optionally, in one embodiment of the present application, the intelligent pattern recognition display system 300 includes:

[0101] The data communication unit is used to receive the first processed signal sent by the signal control module 200, and perform data preprocessing and feature extraction on the first processed signal to obtain a second processed signal of the target sample.

[0102] For example, Figure 2 As shown, the data communication unit in the intelligent pattern recognition display system 300, namely Figure 1 The lower computer in the first processing signal can perform data preprocessing and feature extraction on the first processing signal. Since the original time-sampled sensor response signal usually has high-dimensional characteristics, this not only increases the computational complexity, but also easily introduces noise interference, thereby affecting the accuracy and reliability of the data. Data preprocessing can reduce noise and smooth data, which can effectively reduce the negative impact of missing values, outliers or inconsistencies on model training and prediction. Feature extraction can extract representative information from the original signal, reduce data dimensions, and provide a more effective and stable data basis for subsequent pattern recognition and classification. Therefore, preprocessing and feature extraction of sensor response signals are key steps to improve the performance of electronic nose systems.

[0103] like Figure 6 As shown in the figure, the Savitzky-Golay technique is used to preprocess the data. The signal is smoothed and noise is removed based on sliding window and polynomial fitting. The basic idea of ​​this method is to apply a moving window to the data sequence, and then use polynomial fitting on each window to approximate the data in the window. In this way, a smoothed estimate of the data in the window can be obtained, and the degree of smoothing can be controlled by the order of the fitting polynomial; feature extraction is the process of extracting representative attributes or features that can describe the characteristics of the data from the original data. In the experiment, the difference between the sensor response peak and the stable baseline is selected as the feature value, and the sensor data is feature extracted, where the baseline reference gas is ambient air. The data set is the response signal of 13 gas sensors, so that the second processed signal of the target sample can be obtained, which improves the accuracy and reliability of food freshness detection.

[0104] The data analysis and display unit is used to receive the second processed signal of the target sample, and analyze the food freshness of the target sample according to the second processed signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample, and display the food freshness detection result on a preset terminal.

[0105] In some embodiments, Figure 2 As shown, the data analysis and display unit in the intelligent pattern recognition display system 300, namely Figure 2The host computer in the method can analyze the food freshness of the target sample according to the fusion model of the second processed signal and the target principal component analysis and the multi-layer perceptron. For example, the food freshness can be divided into fresh, semi-fresh, not fresh and other test results, and the corresponding test results of the food can be displayed on a mobile phone or tablet computer, thereby effectively improving the accuracy and convenience of food freshness detection.

[0106] For example, Figure 7 As shown, it is a flow chart of the intelligent pattern recognition display system, which shows the data flow and processing process of the lower computer software and the upper computer software, and reflects the system's efficient collaborative working mechanism in data acquisition, processing, storage and analysis.

[0107] For example, traditional machine learning methods are mainly based on statistical and mathematical models to predict the labels or values ​​of unknown data, and are widely used to process large data sets and perform tasks such as information extraction, feature selection, and classification. Principal component analysis reduces dimensions through orthogonal transformation and retains the main changes. Linear discriminant analysis performs classification by maximizing inter-class distances and minimizing intra-class distances through projection. Random forests and extreme gradient boosting trees solve complex problems through ensemble learning. This experiment compares and analyzes the discrimination of sample freshness by these models.

[0108] Since traditional methods require manual feature selection and parameter adjustment to train models, and parameter selection will directly affect the training effect of the model, this experiment uses grid search technology to find the best parameter combination. Through grid search to traverse the hyperparameter space, combined with ten-fold cross validation to reduce the impact of data set division, optimize hyperparameter selection, and improve model training effect, the basic process is as follows: Figure 8 shown.

[0109] Among them, PCA (Principal Component Analysis) is a classic multivariate statistical method that aims to map high-dimensional data sets to low-dimensional space through linear transformation to capture the main variance in the data. This method achieves data dimensionality reduction by discovering the principal components (eigenvectors) in the data, which are linear combinations with the largest variance in the original data. The core idea of ​​PCA is to map the original data to a new coordinate system by finding the direction that best characterizes the variability of the data.

[0110] Under the storage conditions of 25℃ and 4℃, the PCA analysis results of the samples are as follows Fig. 9 As shown:

[0111] (1) At 25°C, the first principal component (PC1) explained 74.10% of the total variance, and the second principal component (PC2) explained 13.10%, which together explained 87.20% of the information. Meat stored at different times was difficult to separate on the PCA graph, especially the data at 0 h, 6 h, 12 h, and 18 h had a high degree of overlap.

[0112] (2) At 4°C, PC1 explained 74.70% of the total variance, PC2 explained 12.30%, and a total of 87.00% of the information was explained. The data of 0 days, 1 day, and 3 days overlapped highly with the data of 4 days.

[0113] This is due to the linear transformation characteristics of PCA itself, which may not be able to effectively handle the complex relationship between nonlinear categories, resulting in a large category overlap in low-dimensional space. The difficulty in classification may be due to the complexity of the volatile compounds formed by meat spoilage. However, as the storage space increases, the data points move to the right, and this relative change shows that PCA can identify volatiles formed due to spoilage, although it is difficult to observe a clear separation.

[0114] Secondly, PCA analysis can also quantitatively evaluate the contribution of each sensor to the principal component, which helps to understand the relationship between sensor data and their importance in the data dimensionality reduction process. Fig.10 As shown in the bar graph, the higher bars in the bar graph indicate that the corresponding sensor has a greater contribution to the principal component, indicating that the signal of this sensor is of high importance for the interpretation and classification of the data and contains important information about the sample. Fig.10 It can be seen that:

[0115] (1) At 25°C, the S6, S2, S1, and S7 sensors contribute more to the first principal component (PC1), while the S5 and S3 sensors contribute more to the second principal component (PC2).

[0116] (2) Under 4°C conditions, S6, S7, and S3 sensors contribute most to PC1, and S5-2 contributes most to PC2.

[0117] The experimental results show that the content of alcohols, alkanes and ammonia compounds is relatively high during meat spoilage. Further analysis found that compared with 4°C, the contribution rate of each sensor to the main components of 25°C samples was more even, which reflects that the volatile compound composition of samples under 25°C conditions is more complex and the cross-sensitivity of the sensors is more significant, which may lead to relatively poor classification results.

[0118] In addition, LDA (Linear Discriminant Analysis) is a classic pattern recognition and statistical method that aims to reduce the dimensionality of multi-category data by maximizing inter-class differences and minimizing intra-class differences. This method maps high-dimensional data to low-dimensional space through linear combination to maximize the preservation of category information in the new coordinate system. Its core goal is to find the projection direction so that samples of different categories have the greatest separation in this direction to optimize classification performance.

[0119] In the experiment, LDA was used to classify the freshness of pork stored at two different storage conditions: 25°C and 4°C. Fig.11 As shown. The results show that the accuracy of LDA classification of pork freshness under 25℃ storage conditions is 85.00%. In contrast, under 4℃ storage conditions, the accuracy of LDA classification of pork freshness reached 93.30%. It can be clearly seen that LDA has a high degree of differentiation of the freshness of samples under refrigerated storage conditions, but has a poor effect on the freshness of samples under normal temperature conditions. This shows that under low-temperature storage conditions, the LDA model has a more accurate ability to identify the freshness of pork, providing an important quantitative evaluation of the impact of pork storage conditions.

[0120] RF (Random Forest) consists of multiple decision trees, each of which is trained on a different subset of data and trained using a different subset of features. Random Forest is highly robust and accurate, and performs well in processing large-scale data and high-dimensional features.

[0121] The experiment used a random forest classifier to classify the data and used cross validation to evaluate the performance of the model. The purpose of the whole process is to evaluate the performance of the random forest classifier on a given dataset and to use a test set to verify the prediction accuracy of the model.

[0122] Table 2 is the RF model data classification accuracy table, and the specific table 2 is as follows:

[0123] Table 2

[0124] Sample storage temperature Accuracy Standard Deviation 25℃ 90.17％ 0.1252 4℃ 92.17％ 0.1233

[0125] Combination Fig.12 As shown in Table 2, the results of the random forest model after ten-fold cross validation. Among them, the RF model has an accuracy of 90.17% for the freshness classification of samples stored at 25°C, with a standard deviation of 0.1252; the accuracy of 92.17% for the freshness classification of samples stored at 4°C, with a standard deviation of 0.1233. The results show that the RF model has a good recognition effect on the freshness of meat, and has a better classification effect on samples stored under refrigerated conditions.

[0126] XGBoost (Extreme Gradient Boosting Tree) is an efficient machine learning algorithm based on gradient boosting for classification and regression. It improves performance through optimization methods and feature engineering. It consists of multiple decision trees, each of which is built based on the residual of the previous tree and is gradually optimized to improve prediction accuracy.

[0127] The experiment used the XGBoost classifier to perform data classification prediction, which also included the process of cross-validation and model evaluation. This process is similar to the previous random forest classifier. The purpose is to evaluate the performance of the XGBoost classifier on a given data set and use the test set to verify the prediction accuracy of the model. By adjusting the algorithm parameters (the number of trees, etc.), 10-fold cross-validation achieved the highest accuracy. The relevant results are detailed in Table 3, which is the data classification accuracy table of the XGBoost model. The specific Table 3 is as follows:

[0128] Table 3

[0129] Sample storage temperature Accuracy Standard Deviation 25℃ 82.67％ 0.1629 4℃ 92.17％ 0.1233

[0130] Combination Fig.13 Table 3 shows the results of the XGBoost model after ten-fold cross validation. Among them, the XGBoost model has an accuracy of 82.67% for the freshness classification of samples stored at 25°C, with a standard deviation of 0.1629; the accuracy of 92.17% for the freshness classification of samples stored at 4°C, with a standard deviation of 0.1233. The results show that the XGBoost model has a good recognition effect on the freshness of refrigerated samples, but the classification accuracy of samples stored at room temperature is about 10% lower than that under refrigerated conditions, and the recognition effect is poor.

[0131] Furthermore, the experiment used MLP to perform data classification prediction, which also included the process of cross-validation and model evaluation. The purpose of this process is to evaluate the performance of the multi-layer perceptron model on a given data set. Cross-validation can more accurately evaluate the generalization ability of the model. By adjusting the activation function, cross-validation obtains the highest accuracy. Table 4 is the MLP model data classification accuracy table. The specific table 4 is as follows:

[0132] Table 4

[0133] Sample storage temperature MLP Accuracy Standard Deviation 25℃ 94.00％ 0.1280 4℃ 96.00％ 0.0799

[0134] Combination Fig.14 Table 4 shows the results of the MLP model after ten-fold cross validation. Among them, the MLP model has an accuracy of 94.00% for the freshness classification of samples stored at 25°C, with a standard deviation of 0.1280; the accuracy of the MLP model for the freshness classification of samples stored at 4°C is 96.00%, with a standard deviation of 0.0799. The results show that the MLP model can clearly identify the freshness of samples, and has good recognition of the freshness of samples stored at both room temperature and refrigerated temperature conditions.

[0135] In some embodiments, there are significant performance differences in the test results for 13 sensors. According to the response values ​​of each sensor, it is found that different sensors react significantly differently to samples of different freshness. Some sensors respond significantly, while others are relatively weak. In addition, the original data set has a high dimension and contains many interference factors. In order to improve the performance of the model, the data needs to be deeply analyzed and optimized. By selecting the number of principal components to be retained, a controlled dimensionality reduction of the data dimension is achieved while maximizing the total variance of the original data. This feature enables PCA to effectively extract the most important information when processing high-dimensional data, thereby reducing redundancy and simplifying the data structure, providing more explanatory features for further analysis and model building. In the experiment, such as Fig.15 As shown in the figure, the most appropriate dimension for data dimensionality reduction is determined by drawing a scree plot.

[0136] The PCA Scree Plot is as follows Fig.15 (a) and Fig.15 (b) In-depth analysis shows that the explained variance ratio drops sharply before the number of principal components is 2, and stabilizes after the number of principal components is 5. 5 is the "elbow point" of the Scree Plot, indicating that after this point, the additional principal components contribute less to the increase in the explained variance ratio. At the same time, since the total contribution rate of the first five principal components at two different temperatures has exceeded 97%, it was decided to select these five principal components as the new feature space in the experiment to retain most of the important information in the data, further simplify the model, and improve the interpretability and efficiency of the model.

[0137] By combining PCA dimensionality reduction with RF, XGBoost and MLP models, the classification effect of the model is further optimized. The results of the 10-fold cross validation of the fusion algorithm model are as follows Fig.16 As shown, the average accuracy is shown in Table 5, which is the classification accuracy table of PCA, RF, XGBoost and MLP fusion models. The specific Table 5 is as follows:

[0138] Table 5

[0139]

[0140] This application compares the performance of three fusion models, PCA-RF, PCA-XGBoost and PCA-MLP, under different storage conditions.

[0141] Under 25℃ storage conditions, the classification accuracy of the PCA-RF model was 90.20%, which was only 0.03% higher than the RF model, and the effect was not significant. Its standard deviation was 0.1699, which was higher than the RF model, indicating that the model stability decreased. For the PCA-XGBoost model, its classification accuracy was 83.34%. Although it was 0.67% higher than the XGBoost model, the improvement was still not obvious, and the standard deviation was 0.1825, and the model stability was also reduced. In contrast, the classification accuracy of the PCA-MLP model at 25℃ reached 94.67%, which was 0.67% higher than the MLP model, and the effect was not significant, but the standard deviation was reduced to 0.1189, and the model stability was enhanced.

[0142] Under 4°C storage conditions, the performance of each model varies. The classification accuracy of the PCA-RF model is 93.50%, which is 1.33% higher than the RF model, a slight improvement, and its standard deviation is 0.1001, which is relatively small, indicating that the model is relatively stable under this condition. The classification accuracy of the PCA-XGBoost model is 93.00%, which is 0.83% higher than the XGBoost model, and also shows a certain improvement effect, but its standard deviation is 0.1552, and the stability is slightly reduced. The most outstanding is the PCA-MLP model, which has a classification accuracy of 97.83% under 4°C storage conditions, an increase of 1.83% over the MLP model, showing a significant improvement effect, and its standard deviation is only 0.0333, and the model stability is extremely high.

[0143] PCA dimensionality reduction technology effectively improves classification accuracy by simplifying the data model and focusing on key features. Under 4°C storage conditions, the classification effects of RF, XGBoost and MLP models are better than those under 25°C conditions, which is consistent with the analysis of the sensor's contribution to the principal component, reflecting the differences in meat quality changes at different temperatures and their impact on model performance. At lower temperatures, meat quality and volatile compounds change steadily, which helps the model identify and classify more accurately.

[0144] In order to effectively identify the freshness of meat, the experiment compared the impact of different pattern recognition algorithms on the performance of the electronic nose recognition model. Specifically, multiple algorithms such as PCA, LDA, RF, XGBoost and MLP were used to classify the test data of the electronic nose system. After comparative analysis, it was found that the MLP algorithm performed best in classification accuracy. In addition, by analyzing the contribution rate of each sensor to the principal component through PCA, it was found that the contribution of some sensors in meat freshness detection was relatively low. This finding shows that there is room for optimization of the response data of the original sensor array. Further, using the PCA ScreenPlot diagram for analysis, it was found that after reducing the dimension of the data set to 5 dimensions, the key information of the data can still be effectively retained, and it is expected to further optimize the algorithm performance. Finally, this application constructed a PCA-MLP model and confirmed that its classification accuracy was as high as 97.83%, providing strong technical support for the application of electronic noses in the field of meat freshness detection.

[0145] Optionally, in one embodiment of the present application, the system 10 of the embodiment of the present application further includes:

[0146] The test module is used to control the alternating collection of sample gas and baseline gas using a fan-shaped coupling structure before detecting the gas information of at least one target volatile organic compound in the target sample, so as to obtain a test environment that meets preset test conditions.

[0147] As a possible implementation method, a test module can be set up in the embodiment of the present application to simulate the actual test environment. The fan-shaped coupling structure can be used to control the alternating collection of sample gas and baseline gas (air gas) to obtain a test environment that meets certain test conditions, thereby effectively improving the accuracy and reliability of food freshness detection.

[0148] For example, the electronic nose system test can be carried out in advance in the embodiment of the present application, and the electronic nose system test platform simulates the actual environment to evaluate its application in meat freshness detection. During the test, meat samples of different freshness are placed in a static gas collection device, and the sensor array of the electronic nose senses volatile gases and generates response signals. After data collection and processing, the electronic nose makes a preliminary judgment on the freshness of the meat. This process helps to identify key sensors and optimize the sensor array, which is crucial to the development and optimization of pattern recognition algorithms, thereby improving the accuracy and reliability of the electronic nose system in meat freshness assessment.

[0149] First, build the test environment. According to the software and hardware design of the acquisition system, combined with the system test requirements and laboratory conditions, complete the test environment construction of the acquisition system. Fig.17As shown. The test method of the electronic nose system follows the following steps: First, place the pretreated sample in the sample chamber in the gas collection device and let it stand for 30 minutes. During this period, keep the components of the gas volatilization control device in a closed state to ensure that the volatile gases in the sample are fully released into the sample chamber. At the same time, preheat the specific sensor array and monitor the baseline state of the multi-sensor array in real time on the host computer. After the standing period is over and the baseline is stable, rotate the shaft on the gas volatilization control device to open the fan-shaped coupling structure so that the sensor array is in full contact with the sample volatile organic compounds. Subsequently, monitor the response of the sensor array in real time. In the sensor response stage, the system setting time is 300s. It has been verified in the early stage that this parameter can ensure that all sensors respond fully.

[0150] Secondly, the gas sensor response results are analyzed, such as Fig.18 As shown in the figure, each ray in the radar chart represents sample data at different time points, and the length of the ray reflects the size of the sensor response characteristic value, which is quantified by the voltage value. The experiment draws the radar chart with the average value of the sensor response value in 8 tests to make the test results more reliable. Fig.18 (a) shows the response results of the electronic nose to meat at different time periods under storage conditions of 25°C and (b) 4°C. In the experiment, the S1, S2, S3, S6 and S7 sensors all showed high response values ​​at different storage temperatures, indicating that volatile organic compounds such as ammonia, ketones and alcohols are mainly produced during the spoilage process of pork, which are effectively captured by the sensor and converted into quantifiable voltage signals.

[0151] The following conclusions were drawn from the experiment:

[0152] At 25°C, the pork was fresh at 0 hours and 6 hours, and the sensor response value was low; it became semi-fresh at 12 hours, and the response value increased; it was completely corrupted at 18 hours and 24 hours, and the response value increased significantly.

[0153] At 4°C, pork was fresh on day 0 and day 1, with low response values; it became semi-fresh on day 2 and day 3, with increased response values; and it was completely corrupted on day 4, with a significantly increased response value.

[0154] This discovery reveals the significant impact of storage time and temperature on the freshness and degree of spoilage of pork. The changes in the sensor's response are consistent with changes in pork quality, providing an important reference for food preservation and quality monitoring.

[0155] According to the experimental results, as the storage time of meat increases, the response value of the gas sensor array generally shows an upward trend, and the response changes of the gas sensor array to samples at different storage temperatures are not the same, which shows that the composition and content of volatile substances produced by meat at different storage temperatures during the corruption process are different. At the same time, the difference between the response value in the sub-fresh state and the response value in the fresh state is small, and the response value of a single sensor is difficult to distinguish. Storage time and temperature have a significant impact on the freshness and degree of corruption of pork. The response changes of the sensor are consistent with the changes in pork quality, which provides an important reference for food preservation and quality monitoring.

[0156] Secondly, the response results of the temperature and humidity sensors can also be analyzed. The electronic nose system detects the temperature and humidity of samples stored at different temperatures and at different times. Fig.19 As shown. The environment temperature of the temperature and humidity sensor detection specific sensor array in this experiment is 32℃ and the environment humidity is 25.61%RH. Fig.19 It can be seen that the temperature has basically no significant change during the sample spoilage process, and the humidity is about twice as high as the ambient humidity. Under normal temperature storage conditions, the humidity of the sample shows a trend of first rising and then falling. In the process of the sample changing from fresh to semi-fresh, its humidity increases. This may be because as the freshness of the sample decreases, the moisture in it begins to be released, causing the humidity to rise. After the sample begins to spoil, its humidity decreases. This may be due to the consumption or dissipation of moisture caused by certain chemical reactions during the spoilage process. The samples under refrigerated conditions have the same change pattern, but their humidity is generally lower than under normal temperature conditions.

[0157] like Fig. 20 As shown, the detection process of the food freshness detection system based on the gas sensor array is described below with a specific embodiment.

[0158] First, the electronic nose system, i.e., the food freshness detection system, is activated and the gas sensor array begins to detect volatile gases in the meat samples.

[0159] Among them, input data: none; output data: system startup signal.

[0160] Secondly, the signal control module receives the signal from the gas sensor array and performs preliminary processing.

[0161] Among them, input data: sensor signal; output data: processed signal.

[0162] Again, the intelligent pattern recognition display system receives the processed signal and performs data acquisition and analysis.

[0163] Among them, input data: processed signal; output data: collected data and analysis results.

[0164] Finally, the host computer software displays real-time data and provides a user interaction interface.

[0165] Among them, input data: user operation instructions; output data: displayed real-time data and analysis results.

[0166] In summary, the electronic nose system in the embodiment of the present application can realize rapid and accurate detection of food freshness, and at the same time has high stability and reliability, portability and practicality, and provides an effective technical means for food safety monitoring.

[0167] According to the food freshness detection system based on the gas sensor array proposed in the embodiment of the present application, the gas information of at least one target volatile organic compound in the target sample can be detected by the gas sensor array, and the gas information can be converted into a target voltage signal, and then the target voltage signal sent by the gas sensor array can be received by the signal control module, and the target voltage signal can be processed to obtain a first processing signal, and finally the food freshness of the target sample can be analyzed by the intelligent pattern recognition display system based on the first processing signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample, which effectively improves the accuracy and reliability of food freshness detection. Thus, the problems of inaccurate food freshness detection results due to the cross-sensitivity of sensors in the related art, and the inability of the detection range and sensitivity of a single sensor to meet the detection requirements in a complex environment, reducing the accuracy and reliability of food freshness detection, etc. are solved.

[0168] in, Fig.21 A schematic flow chart of a method for detecting food freshness based on a gas sensor array provided in an embodiment of the present application.

[0169] like Fig.21 As shown, the food freshness detection method based on the gas sensor array includes the following steps:

[0170] In step S2101, gas information of at least one target volatile organic compound in a target sample is detected, and the gas information is converted into a target voltage signal.

[0171] In step S2102, the target voltage signal is subjected to target processing to obtain a first processed signal.

[0172] In step S2103, based on the first processed signal and the fusion model of the target principal component analysis and the multi-layer perceptron, the food freshness of the target sample is analyzed to obtain the food freshness detection result of the target sample.

[0173] Among them, in one embodiment of the present application, based on the first processed signal and the fusion model of the target principal component analysis and the multi-layer perceptron, the food freshness of the target sample is analyzed to obtain the food freshness detection result of the target sample, including: performing data preprocessing and feature extraction on the first processed signal to obtain the second processed signal of the target sample; analyzing the food freshness of the target sample according to the second processed signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample, and displaying the food freshness detection result on a preset terminal.

[0174] Optionally, in one embodiment of the present application, before detecting the gas information of at least one target volatile organic compound in the target sample, it also includes: using a fan-shaped coupling structure to control the alternating collection of sample gas and baseline gas to obtain a test environment that meets preset test conditions.

[0175] It should be noted that the aforementioned explanation of the embodiment of the food freshness detection system based on the gas sensor array is also applicable to the food freshness detection method based on the gas sensor array of this embodiment, which will not be repeated here.

[0176] According to the food freshness detection method based on the gas sensor array proposed in the embodiment of the present application, the gas information of at least one target volatile organic compound in the target sample can be detected by the gas sensor array, and the gas information can be converted into a target voltage signal, and then the target voltage signal sent by the gas sensor array is received by the signal control module, and the target voltage signal is subjected to target processing to obtain a first processing signal, and finally the food freshness of the target sample is analyzed by the intelligent pattern recognition display system based on the first processing signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample, which effectively improves the accuracy and reliability of food freshness detection. Thus, the problems that the detection results of food freshness are inaccurate due to the cross-sensitivity of sensors in the related art, and the detection range and sensitivity of a single sensor cannot meet the detection requirements in a complex environment, and the accuracy and reliability of food freshness detection are reduced are solved.

[0177] Fig. 22 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0178] A memory 2201 , a processor 2202 , and a computer program stored in the memory 2201 and executable on the processor 2202 .

[0179] When the processor 2202 executes the program, the food freshness detection method based on the gas sensor array provided in the above embodiment is implemented.

[0180] Furthermore, the electronic device further comprises:

[0181] The communication interface 2203 is used for communication between the memory 2201 and the processor 2202 .

[0182] The memory 2201 is used to store computer programs that can be executed on the processor 2202 .

[0183] The memory 2201 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0184] If the memory 2201, the processor 2202 and the communication interface 2203 are implemented independently, the communication interface 2203, the memory 2201 and the processor 2202 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig. 22 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0185] Optionally, in a specific implementation, if the memory 2201, the processor 2202 and the communication interface 2203 are integrated on a chip, the memory 2201, the processor 2202 and the communication interface 2203 can communicate with each other through an internal interface.

[0186] The processor 2202 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0187] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned food freshness detection method based on the gas sensor array is implemented.

[0188] This embodiment also provides a computer program product, including a computer program. When the computer program is executed, it is used to implement the above-mentioned food freshness detection method based on the gas sensor array.

[0189] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0190] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0191] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0192] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0193] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0194] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0195] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0196] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A food freshness detection system based on a gas sensor array, characterized in that: include: A gas sensor array, the gas sensor array is used to detect gas information of at least one target volatile organic compound in a target sample and convert the gas information into a target voltage signal; A signal control module, the signal control module is used to receive the target voltage signal sent by the gas sensor array, and perform target processing on the target voltage signal to obtain a first processed signal; An intelligent pattern recognition and display system, wherein the intelligent pattern recognition and display system is used to receive the first processed signal sent by the signal control module, and analyze the food freshness of the target sample based on the first processed signal and a fusion model of target principal component analysis and multi-layer perceptron to obtain a food freshness detection result of the target sample.

2. The system according to claim 1, characterized in that The gas sensor array comprises: A plurality of micro-electromechanical system (MEMS) gas sensors, wherein each MEMS gas sensor is used to detect the composition information and concentration information of at least one target volatile organic compound in the target sample, and convert the composition information and the concentration information into the target voltage signal.

3. The system according to claim 1, characterized in that The signal control module comprises: A power supply unit, used to supply power to the gas sensor array, the control unit and the intelligent pattern recognition display system; A control unit is used to receive the target voltage signal, preliminarily process the target voltage signal to obtain the first processed signal, and transmit the first processed signal to the intelligent pattern recognition display system.

4. The system according to claim 1, characterized in that The intelligent pattern recognition display system comprises: a data communication unit, configured to receive the first processed signal sent by the signal control module, and perform data preprocessing and feature extraction on the first processed signal to obtain a second processed signal of the target sample; A data analysis and display unit is used to receive the second processed signal of the target sample, and analyze the food freshness of the target sample according to the second processed signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample, and display the food freshness detection result on a preset terminal.

5. The system according to claim 1, characterized in that Also includes: A test module is used to control the alternating collection of sample gas and baseline gas using a fan-shaped coupling structure before detecting gas information of at least one target volatile organic compound in a target sample, so as to obtain a test environment that meets preset test conditions.

6. A method for detecting food freshness based on a gas sensor array, characterized in that: A food freshness detection system based on a gas sensor array as described in any one of claims 1 to 5 is used, wherein the method comprises the following steps: detecting gas information of at least one target volatile organic compound in a target sample, and converting the gas information into a target voltage signal; Performing target processing on the target voltage signal to obtain a first processed signal; Based on the first processed signal and a fusion model of target principal component analysis and multi-layer perceptron, the food freshness of the target sample is analyzed to obtain a food freshness detection result of the target sample.

7. The method according to claim 6, characterized in that The method of analyzing the food freshness of the target sample based on the first processed signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample includes: Performing data preprocessing and feature extraction on the first processed signal to obtain a second processed signal of the target sample; The food freshness of the target sample is analyzed according to the second processed signal and the fusion model of the target principal component analysis and the multi-layer perceptron to obtain the food freshness detection result of the target sample, and the food freshness detection result is displayed on a preset terminal.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the food freshness detection method based on the gas sensor array as described in any one of claims 6 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the food freshness detection method based on a gas sensor array as described in any one of claims 6 to 7.

10. A computer program product, comprising a computer program, characterized in that The computer program is executed by a processor to implement the food freshness detection method based on a gas sensor array as described in any one of claims 6 to 7.

Images