Method, system and equipment for monitoring microbial content of dairy product in real time based on impedance method

Abstract

The invention discloses a dairy product microorganism content real-time monitoring method, system and equipment based on an impedance method, and relates to the technical field of microorganism monitoring. The method is realized by virtue of a specific biological sensitive membrane modified micro-fluidic chip sensor, a broadband impedance analysis module and an embedded machine learning analysis unit, wherein an automatic sampler extracts a dairy product sample and injects the dairy product sample into the sensor, and target microorganisms are captured and enriched through specific bacteriophages or aptamers on the surface of an electrode; after injecting a low background culture medium and controlling the temperature, periodically scanning in a preset frequency range by a broadband impedance analysis module to obtain complex impedance spectrum data; the embedded unit analyzes data through a pre-trained lightweight model, and outputs a microorganism classification result and content prediction; according to a preset threshold value, whether standard exceeding early warning is generated or not is judged, after early warning, a chip is automatically cleaned, equipment is calibrated, a system integrates a corresponding module to execute the method, rapid, accurate and real-time monitoring of dairy product microorganisms is achieved, and the in-situ early warning requirement of a production line is met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of microorganism monitoring, and more particularly to a method, system and device for real-time monitoring of microbial content in dairy products based on impedance method. BACKGROUND

[0002] In the industrial production of dairy products, microbial contamination is a key risk factor affecting product quality and food safety, especially specific pathogenic bacteria such as Escherichia coli and Listeria, which can cause food safety incidents even in trace amounts. Current detection of microbial contamination in dairy products mainly relies on traditional culture method, colloidal gold immunochromatography and PCR technology. However, these technologies have limitations in industrial real-time monitoring. Traditional culture method is accurate, but requires a culture period of several hours to several days, which cannot meet the needs of real-time feedback and rapid intervention on the production line, and may lead to large-scale production of contaminated products that are not detected until after production, resulting in significant economic losses. Colloidal gold immunochromatography is fast, but can only achieve qualitative or semi-quantitative detection, cannot accurately obtain microbial content, and has a high detection limit for low-concentration microorganisms, which may miss the detection of trace contamination. PCR technology has high sensitivity and specificity, but requires a complex nucleic acid extraction process and is dependent on professional operation. In addition, the equipment is large in size and high in cost, making it difficult to integrate into the production line for in-situ real-time monitoring. At the same time, existing impedance detection technologies have many adaptability problems in the application of dairy products. Traditional impedance sensors are mostly non-microfluidic structures, which require large sample amounts and have low detection efficiency, making it difficult to adapt to the needs of microsampling on the production line. The biological sensitive membrane lacks specificity and is easily interfered by complex matrix such as protein and fat in dairy products, leading to misjudgment of non-target microorganisms. Wideband impedance analysis is mostly limited to a single frequency or a narrow frequency range, which cannot fully capture the impedance spectrum changes during the growth and metabolism of microorganisms, thereby affecting the prediction accuracy of the content. In addition, impedance data processing mostly relies on traditional statistical methods, which have weak pattern recognition ability for complex spectrum data, making it difficult to simultaneously achieve microbial classification and content quantification, and cannot efficiently run on embedded devices, restricting the miniaturization and industrialization of real-time monitoring systems. These technical bottlenecks make it difficult for existing detection methods to meet the needs of real-time, in-situ monitoring of microbial contamination on the production line.

[0003] In summary, the existing technology has the problems of slow detection, low accuracy, poor adaptability, difficulty in real-time, and inability to meet the needs of production line monitoring. SUMMARY

[0004] In order to overcome the problems of slow detection, low accuracy, poor adaptability, difficulty in real-time, and inability to meet the needs of production line monitoring in the prior art, the present application discloses a method, system and device for real-time monitoring of microbial content in dairy products based on impedance method, which can effectively solve the above technical problems.

[0005] To solve the above technical problems, the technical solution of the present application is as follows:

[0006] A real-time monitoring method for microbial content in dairy products based on impedance analysis is proposed. This method utilizes a microfluidic chip sensor modified with a specifically biosensitive membrane, a broadband impedance analysis module, and an embedded machine learning analysis unit. The method includes:

[0007] A dairy product sample to be tested is obtained, and the sample is injected into a microfluidic chip sensor modified with a specific biosensitive membrane. The target microorganisms in the sample are specifically captured and enriched through the biosensitive membrane of the sensor.

[0008] Inject low-background culture medium into the enriched microfluidic chip sensor, control the chip to heat to the preset culture temperature, and start the broadband impedance analysis module.

[0009] The broadband impedance analysis module performs periodic impedance scanning on the enriched target microorganisms within a preset frequency range to obtain complex impedance spectrum data containing impedance magnitude and phase angle.

[0010] The complex impedance spectrum data is input into the embedded machine learning analysis unit, and the data is pattern recognized by a pre-trained lightweight machine learning model to output the classification results and content prediction results of the target microorganisms.

[0011] Based on the classification results and content prediction results, determine whether to generate a microbial exceedance warning.

[0012] Preferably, the microfluidic chip sensor modified with a specific biosensitive membrane includes a microfluidic chip body and gold interdigitated electrodes integrated within the chip, wherein specific bacteriophages or aptamers are chemically immobilized on the electrode surface; the step of injecting the sample into the microfluidic chip sensor modified with a specific biosensitive membrane, and specifically capturing and enriching the target microorganisms in the sample through the biosensitive membrane of the sensor, specifically includes:

[0013] The automatic sampler is controlled to extract a certain volume of sample to be tested from the dairy production line and inject it into the sample inlet of the microfluidic chip sensor.

[0014] When the sample flows over the surface of the gold interdigitated electrode, the specific bacteriophages or aptamers on the electrode surface bind specifically to the target microorganisms in the sample, thereby capturing and enriching the target microorganisms.

[0015] The target microorganism is a specific microorganism.

[0016] Preferably, the step of injecting low-background culture medium into the enriched microfluidic chip sensor and controlling the chip to heat to a preset culture temperature specifically includes:

[0017] Through the medium channel of the microfluidic chip, a certain volume of special low-background medium is injected into the chip area rich in target microorganisms;

[0018] The temperature control module built in the chip is started, and the temperature of the chip is stably controlled at a preset culture temperature in a certain temperature range.

[0019] Preferably, the wideband impedance analysis module adopts a specific impedance analysis chip; the periodic impedance scanning of the enriched target microorganisms in the preset frequency range by the wideband impedance analysis module obtains complex impedance spectrum data containing impedance modulus and phase angle, specifically including:

[0020] The scanning frequency range and scanning period of the wideband impedance analysis module are set;

[0021] During each scanning process, the impedance modulus |Z| and phase angle θ at different frequencies are synchronously collected to generate complex impedance data Z = Z' + jZ"

[0022] Wherein, Z is the complex impedance, Z' is the real part representing the resistance part of the impedance, and Z" is the imaginary part representing the reactance part of the impedance, and j is the imaginary unit;

[0023] The complex impedance data obtained by each scanning is integrated into a complex impedance spectrum data set and stored in the cache module of the embedded machine learning analysis unit.

[0024] Preferably, the embedded machine learning analysis unit adopts a specific series of MCU with DSP function, and the lightweight machine learning model is a convolutional neural network or support vector machine after pruning and quantization processing; the pattern recognition of data by the pre-trained lightweight machine learning model outputs the classification result and content prediction result of the target microorganism, specifically including:

[0025] The real-time collected complex impedance spectrum data set is input into the pre-trained machine learning model, and the model matches and identifies the feature pattern of the spectrum data;

[0026] Two types of results are output: one is the classification result, including whether it is a target microorganism and the corresponding confidence; the other is the regression result, including the estimated content of the target microorganism;

[0027] Wherein, the pre-trained machine learning model is obtained by the following method: collecting impedance spectrum data of a plurality of known microorganisms and different content dairy product samples in the cloud, constructing a training data set, training, pruning and quantizing the initial model, and deploying it to the embedded machine learning analysis unit.

[0028] Preferably, according to the classification result and content prediction result, whether to generate a microorganism over-limit warning information, specifically including:

[0029] a preset content safety threshold of the target microorganism and a classification confidence threshold;

[0030] when the confidence of the target microorganism in the classification result is greater than or equal to the preset confidence threshold, and the content prediction result is greater than or equal to the preset content threshold, it is determined that the microorganism is over-standard, and early warning information containing the over-standard microorganism species, the estimated content and the detection time is generated;

[0031] uploading the early warning information to a central control system of a dairy production line in real time, and triggering local audible and visual alarms.

[0032] Preferably, after the early warning information containing the over-standard microorganism species, the estimated content and the detection time is generated, it further includes:

[0033] controlling the microfluidic chip sensor to start an automatic cleaning process, and flushing the chip channel and the electrode surface with high-pressure cleaning liquid to remove residual microorganisms and culture medium;

[0034] After cleaning, a calibration solution is injected to calibrate the parameters of the wideband impedance analysis module, ensuring the accuracy of the next detection.

[0035] Preferably, the real-time monitoring system for the content of microorganisms in dairy products based on impedance method includes a microfluidic chip sensor modified by a specific biological sensitive film, a wideband impedance analysis module, an embedded machine learning analysis unit, an automatic sampler, a temperature control module, a central communication module and an alarm module.

[0036] The output end of the automatic sampler is connected with the sample inlet of the microfluidic chip sensor, and is used for extracting and conveying the dairy product sample to be detected;

[0037] The microfluidic chip sensor integrates a gold interdigital electrode, and the electrode surface is fixed with specific phages or aptamers, which are used for capturing and enriching target microorganisms;

[0038] The temperature control module is connected with the microfluidic chip sensor, and is used for controlling the temperature of the chip to a preset culture temperature;

[0039] The detection end of the wideband impedance analysis module is connected with the electrode of the microfluidic chip sensor, and is used for collecting complex impedance spectrum data;

[0040] The embedded machine learning analysis unit is connected with the wideband impedance analysis module and the central communication module respectively, and is built-in with a pre-trained lightweight machine learning model, which is used for analyzing impedance data and outputting detection results;

[0041] The central communication module is used for uploading the detection results and early warning information to the central control system;

[0042] The alarm module is connected with the embedded machine learning analysis unit, and is used for triggering sound and light alarm when the microorganism exceeds the standard.

[0043] The system performs the real-time monitoring method of the microbial content of the dairy product based on the impedance method.

[0044] Preferably, an electronic device comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the real-time monitoring method of the microbial content of the dairy product based on the impedance method when executing the computer program.

[0045] Preferably, a computer readable storage medium has a computer program stored thereon, and the processor implements the real-time monitoring method of the microbial content of the dairy product based on the impedance method when executing the computer program.

[0046] Compared with the prior art, the beneficial effects of the present application are: the present application aims to detect slow problems, and the microfluidic chip sensor modified by specific biological sensitive film is adopted to realize rapid capture and enrichment by the specific binding ability of phage or aptamer to target microorganisms, without the long waiting time of traditional culture method, and the wideband impedance analysis module can efficiently complete periodic scanning in the preset frequency range, cooperate with the light weight model quantified by pruning in the embedded machine learning analysis unit, compress the time consumption of the whole process from sample injection to result output, and effectively shorten the detection period; for the problem of low precision, because the biological sensitive film only specifically binds with the target microorganism, the interference of complex matrix such as casein and milk fat in dairy products can be avoided, the wideband impedance analysis module synchronously collects impedance modulus and phase angle to generate multi-dimensional complex impedance spectrum data, and then the model trained based on known microbial spectrum data is used for pattern recognition, which can accurately distinguish the target microorganism and realize quantitative prediction, and improve the detection precision; for the problem of poor adaptation, because the microfluidic chip sensor is small in size and integrates the automatic sampling function, it can be directly embedded into the dairy product production line to realize in-situ sampling, the temperature control module built-in the chip can adapt to the temperature environment of the production line site, and after the warning, the cleaning and calibration process can be automatically started to ensure that the equipment continuously meets the production line monitoring requirements; for the problem of difficult real-time, because the whole process does not need manual intervention, sample collection, microbial enrichment, impedance scanning, data analysis and warning are all automatically completed, the detection result and over-limit warning information can be uploaded to the production line central control system in real time and trigger sound and light alarm, realizing the change from traditional post-detection to real-time intervention, and through the link of rapid enrichment-accurate scanning-intelligent analysis-real-time warning, the defects of the prior art are solved, the needs of rapid, accurate, adaptive and real-time microbial monitoring of dairy product production line are met, and efficient and reliable technical support is provided for food safety control. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only exemplary, and for those skilled in the art, other drawings can be obtained without creative labor based on the provided drawings.

[0048] Figure 1 For the method steps of the present application;

[0049] Figure 2 For the system structure diagram. DETAILED DESCRIPTION

[0050] The drawings are only used for illustrative description and cannot be understood as a limitation of the patent;

[0051] In order to better illustrate the embodiments, some components in the drawings may be omitted, enlarged or reduced, and do not represent the actual product size;

[0052] For those skilled in the art, it is understandable that some well-known structures in the drawings and their descriptions may be omitted.

[0053] The technical solutions of the present application will be further described below in combination with the drawings and embodiments.

[0054] Embodiment 1

[0055] Please refer to Figure 1 , the real-time monitoring method for the microbial content of dairy products based on impedance method, the method is realized based on specific biosensitive film modified microfluidic chip sensor, wideband impedance analysis module and embedded machine learning analysis unit, the method comprises:

[0056] Obtain the dairy product sample to be detected, inject the sample into the specific biosensitive film modified microfluidic chip sensor, capture the target microorganism in the sample through the biosensitive film of the sensor and complete the enrichment;

[0057] Inject low background medium into the enriched microfluidic chip sensor, control the chip to heat to the preset culture temperature, and start the wideband impedance analysis module;

[0058] Periodic impedance scanning of the enriched target microorganism is carried out in the preset frequency range through the wideband impedance analysis module, and complex impedance spectrum data containing impedance modulus and phase angle are obtained;

[0059] The complex impedance spectrum data are input into the embedded machine learning analysis unit, the data are pattern recognized through the pre-trained lightweight machine learning model, and the classification result and content prediction result of the target microorganism are output;

[0060] According to the classification result and the content prediction result, it is judged whether to generate a microorganism over-limit early warning information.

[0061] The specific biosensitive membrane modified microfluidic chip sensor comprises a microfluidic chip body and a gold interdigital electrode integrated in the chip, and the electrode surface is chemically fixed with specific phages or aptamers; the sample is injected into the specific biosensitive membrane modified microfluidic chip sensor, the target microorganism in the sample is specifically captured by the biosensitive membrane of the sensor and enrichment is completed, and the specific steps include:

[0062] The automatic sampler is controlled to extract a certain volume of sample to be detected from the dairy product production line and inject the sample into the sample inlet of the microfluidic chip sensor;

[0063] When the sample flows through the surface of the gold interdigital electrode, the specific phages or aptamers on the surface of the electrode specifically bind to the target microorganism in the sample, realizing the capture and enrichment of the target microorganism;

[0064] The target microorganism is a specific microorganism.

[0065] The low-background medium is injected into the enriched microfluidic chip sensor, and the chip is controlled to heat to a preset culture temperature, and the specific steps include:

[0066] A certain volume of special low-background medium is injected into the chip region enriched with the target microorganism through the medium channel of the microfluidic chip;

[0067] The built-in temperature control module of the chip is started, and the temperature of the chip is stably controlled at a preset culture temperature in a certain temperature range.

[0068] The wideband impedance analysis module adopts a specific impedance analysis chip; the wideband impedance analysis module is used to perform periodic impedance scanning on the enriched target microorganism in a preset frequency range, and complex impedance spectrum data containing impedance modulus and phase angle are obtained, and the specific steps include:

[0069] The scanning frequency range and scanning period of the wideband impedance analysis module are set;

[0070] During each scanning process, the impedance modulus |Z| and phase angle θ at different frequencies are synchronously collected, and complex impedance data Z=Z'+jZ" are generated

[0071] Wherein, Z is the complex impedance, Z' is the resistance part of the real part representing the impedance, and Z" is the imaginary part representing the reactance part of the impedance, and j is the imaginary unit;

[0072] The complex impedance data obtained by each scanning are integrated into a complex impedance spectrum data set and stored in the cache module of the embedded machine learning analysis unit.

[0073] The embedded machine learning analysis unit adopts a specific series of MCUs with DSP function, and the lightweight machine learning model is a convolutional neural network or a support vector machine after pruning and quantization processing; the pre-trained lightweight machine learning model is used for pattern recognition of data, and the classification result and content prediction result of the target microorganism are output, which specifically includes:

[0074] The real-time collected complex impedance spectrum data set is input into the pre-trained machine learning model, and the model matches and recognizes the feature mode of the spectrum data;

[0075] Two types of results are output: one is the classification result, including whether it is the target microorganism and the corresponding confidence; the other is the regression result, including the estimated content of the target microorganism;

[0076] The pre-trained machine learning model is obtained by the following method: collecting impedance spectrum data of a plurality of known microorganisms and different content dairy product samples in the cloud, constructing a training data set, training, pruning and quantizing the initial model, and deploying it to the embedded machine learning analysis unit.

[0077] The classification result and content prediction result are used to determine whether to generate a microorganism over-limit warning information, which specifically includes:

[0078] The content safety threshold and classification confidence threshold of the target microorganism are preset;

[0079] When the confidence of the target microorganism in the classification result is greater than or equal to the preset confidence threshold, and the content prediction result is greater than or equal to the preset content threshold, it is determined that the microorganism is over-limit, and a warning information containing the over-limit microorganism type, estimated content and detection time is generated;

[0080] The warning information is uploaded to the dairy production line central control system in real time, and a local audible and visual alarm is triggered.

[0081] After the warning information containing the over-limit microorganism type, estimated content and detection time is generated, it further includes:

[0082] The microfluidic chip sensor starts an automatic cleaning process, and the chip channel and electrode surface are flushed with high-pressure cleaning liquid to remove residual microorganisms and culture medium;

[0083] After cleaning, a calibration liquid is injected to calibrate the parameters of the wideband impedance analysis module, ensuring the accuracy of the next detection.

[0084] In a specific implementation, in a dairy product production workshop, an automatic sampler is installed, which automatically extracts a certain volume (such as 5 milliliters) of sample from the flowing dairy product on the production line every certain time (such as every 30 minutes) through a pre-set program. The sampling action of the sampler is completed by a high-precision micro pump, which ensures the accuracy and stability of the volume of sample extracted each time.

[0085] The extracted sample is transported through a pipeline to the sample inlet of a microfluidic chip sensor modified with specific biological sensitive membranes. The microfluidic chip sensor has a micro channel system inside. The sample flows in the channel. When the sample flows through the surface of the gold interdigital electrode, the specific phage or aptamer on the surface of the electrode (which are fixed on the surface of the electrode by chemical bonding) will specifically bind to the target microorganism in the sample. If the target microorganism is a certain specific lactic acid bacteria, the specific phage or aptamer fixed on the surface of the electrode will tightly bind to the surface antigen or specific structure of the lactic acid bacteria, thereby achieving the capture and enrichment of the target microorganism. The enrichment process is based on the specific recognition between biological molecules, has high specificity and high efficiency, and can separate the target microorganism from the complex dairy product sample and enrich it on the electrode surface in a short time.

[0086] After the target microorganism in the sample is enriched on the electrode surface of the microfluidic chip sensor, a certain volume (such as 3 milliliters) of low-background medium is injected into the chip area where the target microorganism is enriched through the culture medium channel of the microfluidic chip. This low-background medium is specially prepared and can provide the necessary nutrients for the growth of microorganisms while minimizing interference with subsequent impedance measurements. The injection of the medium is completed by a micro-injection pump, which can accurately control the injection speed and volume of the medium, ensuring that the medium can uniformly cover the electrode surface and the enriched target microorganism.

[0087] Start the built-in temperature control module of the chip to stably control the chip temperature at a pre-set incubation temperature (such as 37 degrees Celsius) within a certain temperature range. The temperature control module is composed of a high-precision temperature sensor and a micro-heater. The temperature sensor can monitor the temperature of the chip in real time and feed back the temperature signal to the temperature control system. The temperature control system controls the heating power of the micro-heater according to the feedback temperature signal, thereby accurately maintaining the chip temperature within the pre-set incubation temperature range. This can ensure that the temperature of the chip remains stable throughout the incubation and impedance scanning process, providing suitable temperature conditions for the growth and metabolic activity of microorganisms.

[0088] The scanning frequency range (e.g. from 100 Hz to 100 kHz) and scanning period (e.g. scanning every 10 minutes) of the wideband impedance analysis module are set, and the wideband impedance analysis module adopts a specific impedance analysis chip with high precision and high sensitivity, which can accurately measure the impedance values at different frequencies. The setting of the scanning frequency range is determined according to the impedance characteristics of the target microorganism and the physicochemical properties of the dairy sample, to ensure that the impedance information of the target microorganism can be comprehensively obtained.

[0089] During each scanning process, the impedance analysis module synchronously collects the impedance modulus |Z| and phase angle θ at different frequencies to generate complex impedance data Z = Z' + jZ", where Z is the complex impedance, Z' is the real part representing the resistance part of the impedance, and Z" is the imaginary part representing the reactance part of the impedance, and j is the imaginary unit. The impedance analysis module can measure the impedance modulus and phase angle at each frequency point through a high-precision measurement circuit and a signal processing unit, and convert these data into digital signals. Then, the complex impedance data obtained in each scanning is integrated into a complex impedance spectrum data set, which is stored in the cache module of the embedded machine learning analysis unit, to ensure the integrity and accuracy of the data and provide a data basis for subsequent data analysis.

[0090] The real-time collected complex impedance spectrum data set is input into the pre-trained lightweight machine learning model, which is implemented using a specific series of MCUs with DSP function. The model is built-in with a pruned and quantized convolutional neural network or support vector machine. The training process of the model is completed in the cloud. By collecting impedance spectrum data of multiple known microorganisms and dairy samples with different contents, a training data set is constructed, and then the initial model is trained, pruned and quantized to improve the calculation efficiency and accuracy of the model. In the actual detection process, the model matches and identifies the feature patterns of the input complex impedance spectrum data, and judges whether the target microorganism exists in the sample and the content of the target microorganism by analyzing various characteristic information in the data, such as the change trend of the impedance modulus and the distribution law of the phase angle.

[0091] The model outputs two types of results: one is the classification result, including whether it is the target microorganism and the corresponding confidence; the other is the regression result, including the estimated content of the target microorganism. For example, the model identifies that the target microorganism exists in the sample, and outputs a classification result such as "target microorganism: Escherichia coli, confidence: 95%", indicating that the model has 95% confidence that the sample contains Escherichia coli. At the same time, the model also outputs a regression result such as "estimated content: 1200 per milliliter", indicating that the model estimates the content of Escherichia coli in the sample to be 1200 per milliliter. These results are obtained through the internal algorithm of the model and a large amount of training data, and have high accuracy and reliability.

[0092] The preset content safety threshold (e.g. 100 per milliliter) and the classification confidence threshold (e.g. 90%) of the target microorganism are determined. When the confidence of the target microorganism in the classification result is greater than or equal to the preset confidence threshold, and the content prediction result is greater than or equal to the preset content threshold, it is determined that the microorganism is over-standard. If the classification result output by the model is the target microorganism: Escherichia coli, the confidence is 95%, the content prediction result is the estimated content: 1200 per milliliter, the preset content safety threshold is 100 per milliliter, and the classification confidence threshold is 90%, then since the confidence 95% is greater than or equal to 90%, and the content 1200 per milliliter is greater than or equal to 100 per milliliter, it is determined that the microorganism is over-standard.

[0093] Generate a warning information containing the over-standard microorganism species (such as Escherichia coli), the estimated content (such as 1200 per milliliter), and the detection time (such as May 18, 2024, 14:30). Upload the warning information to the central control system of the dairy production line in real time. The central control system will take appropriate measures according to the warning information, such as suspending production, cleaning and disinfecting the production line, etc., to prevent further spread of microbial contamination, and at the same time trigger local sound and light alarm to remind on-site workers to take timely measures, so that they can quickly discover the situation of microbial over-standard and take appropriate handling measures.

[0094] After generating the warning information, control the microfluidic chip sensor to start the automatic cleaning process. High-pressure cleaning liquid (such as deionized water containing surfactant) is used to flush the chip channel and electrode surface to remove residual microorganisms and culture medium. After cleaning, inject calibration liquid (such as standard solution with known impedance characteristics) to calibrate the parameters of the wideband impedance analysis module to ensure the accuracy of the next detection. The automatic cleaning process is completed by a miniature cleaning pump and a miniature calibration pump. The cleaning pump can generate high-pressure water flow to flush the residual substances on the chip channel and electrode surface, while the calibration pump can accurately inject calibration liquid to calibrate the impedance analysis module, ensuring that the system can recover to the best state after each detection and provide accurate measurement results for the next detection.

[0095] Example 2

[0096] Please refer to Figure 2 , the real-time monitoring system for the content of microorganisms in dairy products based on impedance method, comprising a specific biosensitive membrane modified microfluidic chip sensor, a wideband impedance analysis module, an embedded machine learning analysis unit, an automatic sampler, a temperature control module, a central communication module and an alarm module;

[0097] The output end of the automatic sampler is connected with the sample inlet of the microfluidic chip sensor, which is used to extract and transport the dairy product sample to be detected;

[0098] The microfluidic chip sensor is integrated with a gold interdigital electrode, and a specific phage or aptamer is fixed on the surface of the electrode to capture and enrich target microorganisms.

[0099] The temperature control module is connected with the microfluidic chip sensor to control the temperature of the chip to a preset culture temperature.

[0100] The detection end of the wideband impedance analysis module is connected with the electrode of the microfluidic chip sensor to collect complex impedance spectrum data.

[0101] The embedded machine learning analysis unit is connected with the wideband impedance analysis module and the central communication module, and a pre-trained lightweight machine learning model is built in, which is used to analyze impedance data and output detection results.

[0102] The central communication module is used to upload the detection results and warning information to the central control system.

[0103] The alarm module is connected with the embedded machine learning analysis unit to trigger sound and light alarms when the microorganisms exceed the standard.

[0104] The system executes the real-time monitoring method of the microbial content in dairy products based on the impedance method.

[0105] The automatic sampler is installed at a suitable position of the dairy product production line, which mainly consists of a high-precision micro pump, a sample extraction pipeline and a sample storage container. The micro pump can automatically extract a certain volume (such as 5 milliliters) of sample from the flowing dairy product on the production line every certain time (such as every 30 minutes) according to the pre-set program, and deliver the sample to the sample storage container through the sample extraction pipeline. The sample storage container has a certain capacity and can temporarily store the extracted sample for subsequent detection.

[0106] The output end of the automatic sampler is connected with the sample inlet of the microfluidic chip sensor through a pipeline made of corrosion-resistant and high-temperature-resistant materials, which can ensure the stability and safety of the sample during transportation. The connection of the pipeline adopts a joint with good sealing performance to prevent the sample from leaking or being contaminated during transportation.

[0107] The microfluidic chip sensor is composed of a microfluidic chip body and a gold interdigital electrode integrated in the chip. The microfluidic chip body has a micro flow channel system inside for the flow of samples and culture medium. The gold interdigital electrode is made by micro-nano processing technology and has the characteristics of high precision and high sensitivity. The electrode surface is fixed with specific phages or aptamers by chemical bonding. These biological sensitive elements can specifically bind to target microorganisms in the sample, thereby realizing the capture and enrichment of target microorganisms. When the sample flows through the electrode surface, the target microorganisms will bind to the biological sensitive elements on the electrode surface and adhere to the electrode surface, completing the enrichment process.

[0108] The electrode of the microfluidic chip sensor is connected to the detection end of the wideband impedance analysis module for transmitting the impedance signal on the electrode surface to the impedance analysis module. Meanwhile, the microfluidic chip sensor is also connected to a temperature control module. The temperature control module can control the temperature of the chip and provide suitable temperature conditions for the growth and metabolic activity of microorganisms. In addition, the culture medium channel of the microfluidic chip sensor is connected to a micro injection pump. The injection pump can inject special low-background culture medium into the culture medium channel of the chip to provide necessary nutrients for the growth of microorganisms.

[0109] The temperature control module is composed of a high-precision temperature sensor and a micro heater. The temperature sensor can monitor the temperature of the microfluidic chip sensor in real time and feed back the temperature signal to the temperature control system. The temperature control system controls the heating power of the micro heater according to the feedback temperature signal, thereby accurately maintaining the chip temperature within the preset culture temperature range (for example, 37 degrees Celsius). The temperature control module has high temperature control accuracy and can ensure that the temperature of the chip remains stable throughout the culture and impedance scanning process, thereby providing suitable temperature conditions for the growth and metabolic activity of microorganisms.

[0110] The temperature control module is connected to the microfluidic chip sensor and connected to the temperature control part of the chip through the temperature sensor and the heater. The temperature sensor is installed at a suitable position of the chip and can monitor the temperature change of the chip in real time. The heater can heat or keep the chip warm according to the instruction of the temperature control system.

[0111] The wideband impedance analysis module adopts an impedance analysis chip, which has high precision and high sensitivity, and can accurately measure the impedance value at different frequencies. The main function of the impedance analysis module is to periodically scan the impedance of the enriched target microorganisms within a pre-set frequency range (e.g. from 100 Hz to 100 kHz) and obtain complex impedance spectrum data containing impedance modulus and phase angle. During each scanning process, the impedance analysis module synchronously collects the impedance modulus |Z| and phase angle θ at different frequencies, generates complex impedance data Z = Z' + jZ", and then integrates the complex impedance data obtained each time into a complex impedance spectrum data set and stores it in the cache module of the embedded machine learning analysis unit.

[0112] The detection end of the wideband impedance analysis module is connected to the electrodes of the microfluidic chip sensor for receiving the impedance signals on the electrode surface. The output end of the impedance analysis module is connected to the embedded machine learning analysis unit to transmit the collected complex impedance spectrum data to the analysis unit for subsequent data analysis and processing.

[0113] The embedded machine learning analysis unit adopts a specific series of MCU with DSP function, which has a built-in lightweight machine learning model (such as convolutional neural network or support vector machine) after pruning and quantization processing. The main function of the analysis unit is to input the real-time collected complex impedance spectrum data set into the pre-trained machine learning model, which matches and identifies the feature patterns of the spectrum data, and outputs the classification results and content prediction results of the target microorganisms. The pre-trained machine learning model is constructed by collecting impedance spectrum data of various known microorganisms and different content dairy product samples in the cloud, training data set, training the initial model, pruning and quantization processing, and deploying to the embedded machine learning analysis unit. The analysis unit can quickly and accurately analyze and process the input data and output reliable detection results.

[0114] The embedded machine learning analysis unit is connected to the wideband impedance analysis module and the central communication module. It receives complex impedance spectrum data from the impedance analysis module, analyzes and processes the data, and uploads the detection results and warning information to the dairy production line central control system through the central communication module. The analysis unit is also connected to the alarm module, which can trigger the alarm module to send sound and light alarm signals when the microorganisms exceed the standard.

[0115] The main function of the central communication module is to upload the detection results and warning information output by the embedded machine learning analysis unit to the dairy production line central control system in real time. The central communication module uses high-speed and stable communication protocols such as Ethernet or wireless communication protocols to ensure the accuracy and timeliness of data transmission. Through the central communication module, the central control system can obtain real-time detection data and warning information, thereby effectively monitoring and managing the production process.

[0116] The central communication module is connected with the embedded machine learning analysis unit, receives the data output by the analysis unit, and uploads the data to the central control system. Meanwhile, the central communication module is also connected with the central control system of the dairy product production line, realizing bidirectional transmission and communication of data.

[0117] The main function of the alarm module is to trigger sound and light alarm when the microorganism exceeds the standard, reminding the on-site staff to take timely measures. The alarm module is composed of a sound and light alarm device. After receiving the early warning signal sent by the embedded machine learning analysis unit, the device can immediately send dazzling light and piercing alarm sound to attract the attention of the staff. The sound and light alarm device has the characteristics of high brightness and high volume, and can clearly send alarm signals in a noisy production workshop environment.

[0118] The alarm module is connected with the embedded machine learning analysis unit, receives the early warning signal output by the analysis unit, and triggers the alarm device to send sound and light alarm signals according to the signal.

[0119] Through the implementation of the above system, real-time, rapid and accurate monitoring of the microorganism content of dairy products can be realized, the quality and safety of dairy products can be effectively guaranteed, and the automation and intelligent level of the production process can be improved.

[0120] An electronic device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the real-time monitoring method of the microorganism content of dairy products based on the impedance method when executing the computer program.

[0121] A computer readable storage medium, the computer readable storage medium has a computer program stored thereon, and the processor implements the real-time monitoring method of the microorganism content of dairy products based on the impedance method when executing the computer program.

[0122] In the embodiment, the electronic device is equipped with a high-performance processor and a large-capacity memory, and the memory stores a special computer program, which can realize the real-time monitoring method for the microbial content of dairy products based on the impedance method when the processor executes the program. Specifically, the program first controls the automatic sampler to extract samples from the dairy product production line and inject them into the microfluidic chip sensor modified by the specific biosensitive film. The gold interdigital electrode surface of the sensor is fixed with specific phages or aptamers, which can specifically capture target microorganisms and complete enrichment. Subsequently, the program controls the injection of low-background medium into the chip, and the chip is heated to the preset culture temperature through the temperature control module. The broadband impedance analysis module performs periodic impedance scanning on the enriched target microorganisms within the preset frequency range to obtain complex impedance spectrum data. These data are input into the embedded machine learning analysis unit, which uses a MCU with DSP function and a lightweight machine learning model that has been pruned and quantized. The unit performs pattern recognition on the data and outputs the classification result and content prediction result of the target microorganism. If the classification confidence and content prediction result exceed the preset threshold, it is determined that the microorganism is over-standard, and a warning information is generated and uploaded to the central control system, and a local audible and visual alarm is triggered. Finally, the program controls the microfluidic chip sensor to start the automatic cleaning process and inject the calibration liquid to calibrate the impedance analysis module to ensure the accuracy of the next detection. The entire process realizes data interaction with the central control system through the central communication module. The combination of the electronic device and the computer-readable storage medium provides an efficient and accurate solution for real-time monitoring of the microbial content of dairy products.

[0123] The same or similar reference signs correspond to the same or similar components;

[0124] The terms used to describe the positional relationship in the drawings are only used for illustrative purposes and should not be construed as limiting the patent;

[0125] Obviously, the above embodiments of the present application are only examples for clearly illustrating the present application, and are not intended to limit the embodiments of the present application. For those skilled in the art, other different forms of changes or variations can be made on the basis of the above description, and here, all the embodiments need not and cannot be exhausted, and any modification, equivalent replacement and improvement made within the spirit and principle of the present application should be included in the protection scope of the claims of the present application.

Claims

1. A method for real-time monitoring of the microbial content of dairy products based on the impedance method, characterized in that, The method is realized based on a specific biosensitive membrane modified microfluidic chip sensor, a wideband impedance analysis module and an embedded machine learning analysis unit, and the method comprises the following steps: A sample to be detected is obtained, and the sample is injected into the specific biosensitive membrane modified microfluidic chip sensor, so that the target microorganism in the sample is specifically captured by the biosensitive membrane of the sensor and enrichment is completed; Low-background medium is injected into the microfluidic chip sensor after enrichment, and the chip is heated to a preset culture temperature, and the wideband impedance analysis module is started; Periodic impedance scanning of the enriched target microorganism is performed in a preset frequency range by the wideband impedance analysis module, and complex impedance spectrum data containing impedance modulus and phase angle are obtained; The complex impedance spectrum data are input into the embedded machine learning analysis unit, pattern recognition is performed on the data by a pre-trained lightweight machine learning model, and classification results and content prediction results of the target microorganism are output; According to the classification results and content prediction results, it is judged whether to generate a microorganism over-limit warning information.

2. The method for real-time monitoring of microbial content in dairy products based on impedance method as claimed in claim 1, wherein, The specific biosensitive membrane modified microfluidic chip sensor comprises a microfluidic chip body and a gold interdigital electrode integrated in the chip, and the electrode surface is chemically fixed with specific phages or aptamers; the sample is injected into the specific biosensitive membrane modified microfluidic chip sensor, and the target microorganism in the sample is specifically captured by the biosensitive membrane of the sensor and enrichment is completed, which specifically comprises the following steps: An automatic sampler is controlled to extract a certain volume of sample to be detected from a dairy product production line and inject it into the sample inlet of the microfluidic chip sensor; When the sample flows through the surface of the gold interdigital electrode, the specific phages or aptamers on the surface of the electrode specifically bind to the target microorganism in the sample, thereby achieving capture and enrichment of the target microorganism; The target microorganism is a specific microorganism.

3. The method for real-time monitoring of microbial content in dairy products based on impedance method as claimed in claim 2, wherein, The low-background medium is injected into the microfluidic chip sensor after enrichment, and the chip is heated to a preset culture temperature, which specifically comprises the following steps: A certain volume of special low-background medium is injected into the chip region where the target microorganism is enriched through the medium channel of the microfluidic chip; A built-in temperature control module of the chip is started, and the temperature of the chip is stably controlled at a preset culture temperature in a certain temperature range.

4. The method for real-time monitoring of microbial content in dairy products based on impedance method as claimed in claim 1, wherein, The wideband impedance analysis module uses a specific impedance analysis chip; the complex impedance spectrum data containing impedance modulus and phase angle are obtained by performing periodic impedance scanning of the enriched target microorganism in a preset frequency range by the wideband impedance analysis module, which specifically comprises the following steps: The scanning frequency range and scanning period of the wideband impedance analysis module are set; During each scanning process, the impedance modulus |Z| and phase angle θ at different frequencies are synchronously collected, and complex impedance data Z=Z′+jZ″ are generated Wherein, Z is the complex impedance, Z′ is the real part representing the resistance part of the impedance, Z″ is the imaginary part representing the reactance part of the impedance, and j is the imaginary unit; The complex impedance data obtained by each scanning are integrated into a complex impedance spectrum data set and stored in the cache module of the embedded machine learning analysis unit.

5. The method for real-time monitoring of microbial content in dairy products based on impedance method as claimed in claim 4, wherein, The embedded machine learning analysis unit adopts a specific series of MCUs with DSP function, and the lightweight machine learning model is a convolutional neural network or a support vector machine after pruning and quantization processing; the model is used for pattern recognition of data, and outputs classification results and content prediction results of target microorganisms, specifically including: Input the real-time collected complex impedance spectrum data set into the pre-trained machine learning model, and the model matches and identifies the feature mode of the spectrum data; Output two types of results: one is the classification result, including whether it is a target microorganism and the corresponding confidence; the other is the regression result, including the estimated content of the target microorganism; The pre-trained machine learning model is obtained by the following method: collecting impedance spectrum data of multiple known microorganisms and different content dairy product samples in the cloud, constructing a training data set, training, pruning and quantizing the initial model, and deploying it to the embedded machine learning analysis unit.

6. The method for real-time monitoring of microbial content in dairy products based on impedance method as claimed in claim 1, wherein, The classification result and content prediction result are used to determine whether to generate a microorganism over-limit warning information, specifically including: Preset the content safety threshold and classification confidence threshold of the target microorganism; When the confidence of the target microorganism in the classification result is greater than or equal to the preset confidence threshold, and the content prediction result is greater than or equal to the preset content threshold, it is determined that the microorganism is over-limit, and a warning information including the over-limit microorganism type, estimated content and detection time is generated; The warning information is uploaded to the dairy production line central control system in real time, and a local audible and visual alarm is triggered.

7. The method for real-time monitoring of microbial content in dairy products based on impedance method as claimed in claim 6, wherein, After the generation of the warning information including the over-limit microorganism type, estimated content and detection time, it further includes: Control the microfluidic chip sensor to start the automatic cleaning process, and flush the chip channel and electrode surface with high-pressure cleaning liquid to remove residual microorganisms and culture medium; After cleaning, inject a calibration solution to calibrate the parameters of the wideband impedance analysis module to ensure the accuracy of the next detection.

8. A real-time monitoring system for the microbial content of dairy products based on impedance method, characterized in that, It includes a microfluidic chip sensor modified with specific biological sensitive film, a wideband impedance analysis module, an embedded machine learning analysis unit, an automatic sampler, a temperature control module, a central communication module and an alarm module; The output end of the automatic sampler is connected with the sample inlet of the microfluidic chip sensor, which is used for extracting and conveying the dairy product sample to be detected; The microfluidic chip sensor integrates gold interdigital electrodes, and the electrode surface is fixed with specific phages or aptamers for capturing and enriching target microorganisms; The temperature control module is connected with the microfluidic chip sensor for controlling the chip temperature to a preset culture temperature; The detection end of the wideband impedance analysis module is connected with the electrode of the microfluidic chip sensor for collecting complex impedance spectrum data; The embedded machine learning analysis unit is connected with the wideband impedance analysis module and the central communication module, and has a pre-trained lightweight machine learning model built-in, which is used for analyzing impedance data and outputting detection results; The central communication module is used for uploading the detection results and warning information to the central control system; The alarm module is connected with the embedded machine learning analysis unit, and is used for triggering audible and visual alarm when the microorganism is over-limit. The system performs the real-time monitoring method of microbial content in dairy products based on impedance method according to any one of claims 1-7.

9. An electronic device, comprising: A computer readable storage medium, having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the real-time monitoring method of microbial content in dairy products based on impedance method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer readable storage medium, having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the real-time monitoring method of microbial content in dairy products based on impedance method according to any one of claims 1-7.

Images