An automated detection method and system based on ATP fluorescence detection

Abstract

The application provides an automatic detection method and system based on ATP fluorescence detection. Wherein, the sample to be detected is obtained, uniformly mixed, and a mixed solution is generated; the fluorescence signal intensity is detected, the background fluorescence value is synchronously collected and recorded, the environmental interference noise is eliminated, the signal-to-noise ratio is improved, and the optimized fluorescence signal intensity change rate is obtained; the fluorescence signal intensity change rate is predicted and processed, the preliminary concentration prediction value of ATP in the sample to be detected is generated, the detection deviation caused by the reagent batch difference is effectively corrected, and the corrected preliminary concentration prediction value is generated; the dynamic range expansion algorithm is implemented, the fluorescence detection parameter is automatically adjusted, and the final optimized detection result is obtained; after the detection is completed, the final optimized detection result is protected, sent to the cloud server, and a comprehensive biological safety monitoring report is generated. The technical scheme provided by the application improves the overall performance and reliability of ATP fluorescence detection.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of biological detection, and particularly relate to an automatic detection method and system based on ATP fluorescence detection. BACKGROUND

[0002] In the fields of food safety, environmental monitoring, and medical diagnosis, it is crucial to quickly and accurately detect the level of microbial contamination. Adenosine triphosphate (ATP) serves as the energy carrier for all living cells, and its concentration can indirectly reflect the activity level of microorganisms in a sample.

[0003] There are various handheld or laboratory-level detection devices based on the ATP fluorescence principle on the market. These devices usually use disposable swabs or test tubes to collect samples, and through the action of pre-installed enzyme reaction reagents with ATP in the sample, fluorescence signals are generated, and then photoelectric sensors are used to measure the fluorescence intensity to calculate the ATP content. Some advanced models also integrate automatic sampling, data processing software, and other functions, aiming to simplify the operation process and improve analysis efficiency.

[0004] Although existing solutions have met the basic detection needs to some extent, they still face some challenges: especially under low concentration ATP conditions, traditional methods have difficulty in ensuring sufficient detection accuracy. External light, temperature changes, and other factors may interfere with test results, leading to poor repeatability. Differences in reagents between batches may cause deviations in detection results, requiring frequent manual correction. When encountering extremely high or low concentration samples, existing detection devices often cannot provide accurate results. Lack of advanced data analysis tool support makes it difficult to realize automated trend analysis and abnormal warning functions.

[0005] In summary, the present application proposes a new generation of ATP fluorescence detection scheme based on microfluidic technology and machine learning algorithm optimization, aiming to overcome the current technical limitations and provide faster, more accurate, and more stable services. SUMMARY

[0006] Embodiments of the present application provide an automatic detection method and system based on ATP fluorescence detection to solve the problem of low overall performance and reliability of ATP fluorescence detection in the prior art.

[0007] In a first aspect, embodiments of the present application provide an automatic detection method based on ATP fluorescence detection, comprising:

[0008] Obtaining a sample to be tested, controlling the flow speed and direction of the liquid on the microfluidic chip, and uniformly mixing the sample to be tested with the pre-mixed fluorescently labeled ATP detection reagent to generate a mixed solution;

[0009] Based on the mixed solution, the fluorescence signal intensity is detected at a specific wavelength by using a high-sensitivity photomultiplier tube, the background fluorescence value is synchronously collected and recorded, environmental interference noise is eliminated by applying digital signal processing technology to the collected signal, the signal-to-noise ratio of the signal is improved, and the optimized fluorescence signal intensity change rate is obtained;

[0010] The optimized fluorescence signal intensity change rate is analyzed by using an adaptive threshold algorithm, a machine learning model trained based on a large amount of historical detection data is combined, the fluorescence signal intensity change rate is predicted and processed, a preliminary concentration prediction value of ATP in the sample to be detected is generated, and detection deviation caused by reagent batch difference is effectively corrected to generate a corrected preliminary concentration prediction value;

[0011] According to the corrected preliminary concentration prediction value, a dynamic range expansion algorithm is implemented, the fluorescence detection parameters are automatically adjusted to adapt to the ATP detection requirements in different concentration ranges, the fluorescence signal is re-detected and optimized, and the final optimized detection result is obtained to ensure that the detection accuracy is not limited by the change of sample concentration;

[0012] After the detection is completed, the final optimized detection result is protected, the encrypted data is sent to a cloud server through a safe wireless communication protocol, real-time monitoring, trend prediction and abnormal alarm functions are realized, and a comprehensive biosafety monitoring report is generated.

[0013] Optionally, the adaptive threshold algorithm is used to analyze the optimized fluorescence signal intensity change rate, a machine learning model trained based on a large amount of historical detection data is combined, the fluorescence signal intensity change rate is predicted and processed, a preliminary concentration prediction value of ATP in the sample to be detected is generated, and detection deviation caused by reagent batch difference is effectively corrected to generate a corrected preliminary concentration prediction value, including:

[0014] The optimized fluorescence signal is used to perform time series analysis on the optimized fluorescence signal, the trend characteristics of the fluorescence signal changing with time are extracted, and time series data of the fluorescence signal intensity change rate are obtained;

[0015] Based on the time series data of the fluorescence signal intensity change rate, an adaptive threshold algorithm is applied to dynamically set a threshold to distinguish the effective change part from the noise part in the signal, and the fluorescence signal intensity change rate of the effective change part is obtained;

[0016] A machine learning model trained based on a large amount of historical detection data is combined to predict and process the fluorescence signal intensity change rate of the effective change part, and a preliminary concentration prediction value of ATP in the sample to be detected is generated;

[0017] In the machine learning model, based on the difference in fluorescence response characteristics between different batches of reagents in historical data, the preliminary concentration prediction value of ATP in the to-be-tested sample is corrected to generate a corrected preliminary concentration prediction value, so as to reduce the influence of reagent batch difference on the detection result and improve the accuracy and consistency of the detection result.

[0018] Optionally, based on the time series data of the fluorescence signal intensity change rate, an adaptive threshold algorithm is applied to dynamically set a threshold to distinguish the effective change part from the noise part in the signal, and the fluorescence signal intensity change rate of the effective change part is obtained, including:

[0019] Using the time series data of the fluorescence signal intensity change rate, the statistical characteristics of the time series data of the fluorescence signal intensity change rate are calculated, including the mean and the standard deviation, and the basic parameters are obtained;

[0020] According to the basic parameters, an adaptive threshold algorithm is applied to dynamically calculate the threshold of each time point to obtain a dynamic threshold, and the threshold of each time point can be adjusted with the fluctuation of the time series data to adaptively identify the effective change part from the noise part in the signal;

[0021] Based on the dynamically calculated threshold of each time point, the time series data of the fluorescence signal intensity change rate is filtered to remove data points below the threshold, and the data points below the threshold are considered to belong to the noise part, and the preliminary filtered fluorescence signal intensity change rate is obtained;

[0022] The data points higher than or equal to the dynamic threshold are retained, and the data points higher than or equal to the dynamic threshold are considered to represent the effective change part in the signal, and the fluorescence signal intensity change rate of the effective change part is generated from the preliminary filtered fluorescence signal intensity change rate.

[0023] Optionally, based on the time series data of the fluorescence signal intensity change rate, an adaptive threshold algorithm is applied to dynamically set a threshold to distinguish the effective change part from the noise part in the signal, and the fluorescence signal intensity change rate of the effective change part is obtained, including:

[0024] Using the time series data of the fluorescence signal intensity change rate, the local statistical characteristics of the time series data of the fluorescence signal intensity change rate in the time window [t-Δt, t+Δt] are calculated, including the local mean μ loc (t) and the local standard deviation σ loc (t), and the local basic parameters are obtained;

[0025] According to the local basic parameters μ loc (t) and σ loc(t), combined with global statistical characteristics μ and σ, and considering the importance of different time points in time series data, a time weight factor w(t) and a nonlinear adjustment factor f(x) are applied to dynamically calculate the threshold T(t) of each time point t, and the threshold of each time point is calculated by the following formula:

[0026] T(t) = f(μ + w(t)·k·(σ loc (t)) α )

[0027] Wherein, T(t) represents the threshold of each time point t, μ represents the global mean, and σ represents the average value of the change rate of fluorescence signal intensity in the whole time series; σ represents the global standard deviation, and σ represents the volatility of the change rate of fluorescence signal intensity in the whole time series; k is an adjustable coefficient for controlling the sensitivity of the threshold, and α is an adjustable index for adjusting the influence degree of local standard deviation on the threshold, is a time weight factor, which is used to reflect the importance of data at different time points, and β represents a parameter for controlling the change rate of weight with time, which is adjusted according to actual application requirements; t0 represents a reference time point, which is usually selected as the center point or starting point of the time series; f(x) = x + γ·(x-μ) 2 is a nonlinear function for increasing the nonlinear adjustment ability of the threshold, and γ is an adjustable coefficient for controlling the degree of nonlinear adjustment, which is adjusted according to actual application requirements, and T(t) can be adjusted with the fluctuation of time series data to adaptively identify the effective change part and noise part in the signal;

[0028] Based on the dynamically calculated threshold T(t) of each time point t, the time series data of the change rate of fluorescence signal intensity is filtered, and the data points below the threshold T(t) are removed, which are considered to belong to the noise part, and the preliminary filtered change rate of fluorescence signal intensity is obtained.

[0029] The data points higher than or equal to the dynamic threshold T(t) are reserved, which are considered to represent the effective change part in the signal, and the change rate of fluorescence signal intensity of the effective change part is generated from the preliminary filtered change rate of fluorescence signal intensity.

[0030] Optionally, the dynamic range expansion algorithm is implemented according to the corrected preliminary concentration prediction value, the fluorescence detection parameters are automatically adjusted to adapt to the ATP detection requirements in different concentration ranges, the fluorescence signal is re-detected and optimized, and the final optimized detection result is obtained to ensure that the detection precision is not limited by the change of sample concentration, which comprises:

[0031] Using the preliminary predicted concentration of ATP in the sample to be tested, the preliminary predicted concentration is evaluated to obtain the approximate concentration range of ATP in the sample to be tested.

[0032] Based on the approximate concentration range of ATP in the sample to be tested, a corresponding dynamic range expansion algorithm is selected, and the fluorescence detection parameters are selected and adjusted to determine the appropriate exposure time and gain settings for the current concentration range.

[0033] Based on the operating parameters, the operating state of the high-sensitivity photomultiplier tube is automatically adjusted, and the fluorescence signal intensity of the mixture is re-detected to obtain the fluorescence signal under optimized detection conditions.

[0034] Using the fluorescence signal under the optimized detection conditions, digital signal processing technology is applied again to eliminate environmental interference noise, improve the signal-to-noise ratio, and generate a further optimized fluorescence signal.

[0035] Based on the further optimized fluorescence signal, the adaptive threshold algorithm and machine learning model are applied again to predict the rate of change of fluorescence signal intensity, generating the final optimized detection result and ensuring that the detection accuracy is not limited by the change of sample concentration.

[0036] Optionally, the step of selecting a corresponding dynamic range expansion algorithm based on the approximate concentration range of ATP in the sample to be tested, selecting and adjusting the fluorescence detection parameters, and determining the operating parameters of exposure time and gain settings suitable for the current concentration range includes:

[0037] Using the approximate concentration range of ATP in the sample to be tested, the dynamic range expansion algorithm library is searched to find the most suitable dynamic range expansion algorithm for the current concentration range, and the selected dynamic range expansion algorithm is generated.

[0038] Based on the selected dynamic range expansion algorithm, the fluorescence detection parameters are simulated and analyzed to determine the exposure time and gain settings that provide the best detection performance within the current ATP concentration range, and the working parameters are generated.

[0039] Optionally, the step of selecting a corresponding dynamic range expansion algorithm based on the approximate concentration range of ATP in the sample to be tested, selecting and adjusting the fluorescence detection parameters, and determining the operating parameters of exposure time and gain settings suitable for the current concentration range includes:

[0040] Using the approximate concentration range C of ATP in the sample to be tested range The dynamic range expansion algorithm library is searched to find the most suitable dynamic range expansion algorithm for the current concentration range, and the selected dynamic range expansion algorithm is generated.

[0041] Based on the selected dynamic range expansion algorithm, the fluorescence detection parameters are simulated and analyzed to determine the exposure time and gain setting that can provide the best detection performance in the current ATP concentration range, and working parameters are generated, specifically including:

[0042] Using the approximate concentration range C of ATP in the sample to be tested range , select the most suitable algorithm A(C range ) from the dynamic range expansion algorithm library;

[0043] Based on the selected dynamic range expansion algorithm A(C range ), simulate and analyze the fluorescence detection parameters, and calculate the performance index P(E, G) of each parameter combination, where E represents the exposure time and G represents the gain setting;

[0044] The performance index P(E, G) is calculated by the following formula:

[0045]

[0046] Where S(E, G) represents the signal intensity, which is obtained by simulation and reflects the intensity of the fluorescence signal under the given exposure time and gain setting; D(E, G) represents the signal-to-noise ratio, which is obtained by simulation and reflects the signal-to-noise ratio of the fluorescence signal under the given exposure time and gain setting; N(E, G) represents the noise level, which is obtained by simulation and reflects the noise level of the fluorescence signal under the given exposure time and gain setting; T(E, G) represents the detection time, which is obtained by simulation and reflects the time required to complete one detection under the given exposure time and gain setting; w1(C range ), w2(C range ), w3(C range ) and w4(C range ) are dynamic weight factors used to dynamically adjust the influence of different performance indicators according to the ATP concentration range C range ; α(C range ), β(C range ), γ(C range ) and δ(C range ) are adjustable exponents used to dynamically adjust the nonlinear influence of each performance indicator according to the ATP concentration range C range ; and according to the actual application requirements, the nonlinear influence of each performance indicator is adjusted;

[0047] Optimization algorithm is used to find the exposure time and gain setting that maximizes the performance index P(E, G), and the best working parameters (E opt , G opt ) are generated;

[0048] Based on the optimal operating parameters (E) opt G opt This generates the final operating parameters, which are used to automatically adjust the operating status of the high-sensitivity photomultiplier tube to ensure optimal detection performance within the current ATP concentration range.

[0049] Optionally, based on the further optimized fluorescence signal, the adaptive threshold algorithm and machine learning model are applied again to predict the rate of change of fluorescence signal intensity, generating the final optimized detection result to ensure that the detection accuracy is not limited by changes in sample concentration, including:

[0050] Using the further optimized fluorescence signal, time series analysis was performed to extract the characteristics of fluorescence signal changes over time, and optimized fluorescence signal feature data was obtained.

[0051] Based on the optimized fluorescence signal feature data, the adaptive threshold algorithm is applied again to dynamically set the threshold to distinguish the effective change part from the noise part in the signal, and the fluorescence signal intensity change rate of the effective change part is obtained.

[0052] Using the fluorescence signal intensity change rate of the effective change portion as input feature, and combined with a machine learning model trained based on a large amount of historical detection data, the fluorescence signal intensity change rate is predicted to generate the final predicted value of ATP concentration in the sample to be tested.

[0053] In the machine learning model, the final concentration prediction value is corrected based on the differences in fluorescence response characteristics between different reagent batches in historical data, and a corrected final concentration prediction value is generated to reduce the impact of reagent batch differences on the detection results and improve the accuracy and consistency of the detection results.

[0054] Based on the corrected final concentration prediction, the final optimized detection result is generated to ensure that the detection result remains highly accurate and reliable regardless of how the concentration of ATP in the sample changes.

[0055] Optionally, based on the mixture, the fluorescence signal intensity is detected at a specific wavelength using a high-sensitivity photomultiplier tube, background fluorescence values ​​are simultaneously acquired and recorded, and environmental interference noise is eliminated by applying digital signal processing technology to the acquired signal to improve the signal-to-noise ratio, thereby obtaining an optimized fluorescence signal intensity change rate, including:

[0056] Using the mixture, a high-sensitivity photomultiplier tube is controlled to work at a specific wavelength, and the fluorescence signal intensity of the mixture is detected. At the same time, the background fluorescence value is collected and recorded to obtain the original fluorescence signal data.

[0057] Based on the original fluorescence signal data, signal processing is performed to remove environmental interference noise and improve the clarity of the signal, and a preliminary processed fluorescence signal is generated.

[0058] Using the preliminary processed fluorescence signal and the background fluorescence value, background fluorescence correction processing is performed to eliminate the influence of background fluorescence on the signal, and a corrected fluorescence signal is obtained.

[0059] Based on the corrected fluorescence signal, the signal quality is further optimized to ensure that the signal-to-noise ratio of the fluorescence signal reaches the best, and an optimized fluorescence signal is generated.

[0060] In a second aspect, the embodiments of the present application provide an automatic detection system based on ATP fluorescence detection, comprising:

[0061] An acquisition control module is configured to acquire a sample to be tested, control the flow speed and direction of liquid on a microfluidic chip, and uniformly mix the sample to be tested with a premixed fluorescence-labeled ATP detection reagent to generate a mixed liquid.

[0062] A detection and collection module is configured to detect the fluorescence signal intensity at a specific wavelength based on the mixed liquid using a high-sensitivity photomultiplier tube, synchronously collect and record the background fluorescence value, eliminate environmental interference noise by applying digital signal processing technology to the collected signal, improve the signal-to-noise ratio of the signal, and obtain the optimized fluorescence signal intensity change rate.

[0063] An analysis and prediction module is configured to analyze the optimized fluorescence signal intensity change rate using an adaptive threshold algorithm, combine a machine learning model trained based on a large amount of historical detection data, and perform prediction processing on the fluorescence signal intensity change rate to generate a preliminary concentration prediction value of ATP in the sample to be tested, while effectively correcting the detection deviation caused by reagent batch differences to generate a corrected preliminary concentration prediction value.

[0064] An adjustment and detection module is configured to implement a dynamic range expansion algorithm according to the corrected preliminary concentration prediction value, automatically adjust the fluorescence detection parameters to adapt to the ATP detection requirements in different concentration ranges, re-detect and optimize the fluorescence signal, and obtain the final optimized detection result to ensure that the detection accuracy is not limited by the change of sample concentration.

[0065] A protection generation module is configured to protect the final optimized detection result after detection, send the encrypted data to a cloud server through a secure wireless communication protocol, realize real-time monitoring, trend prediction and abnormal alarm functions, and generate a comprehensive biological safety monitoring report.

[0066] In the embodiments of the present application, the sample to be tested is obtained, the flow speed and direction of the liquid on the microfluidic chip are controlled, the sample to be tested and the premixed fluorescently labeled ATP detection reagent are uniformly mixed to generate a mixed solution; based on the mixed solution, the high-sensitivity photomultiplier tube is used to detect the fluorescence signal intensity at a specific wavelength, the background fluorescence value is synchronously collected and recorded, the environmental interference noise is eliminated by applying digital signal processing technology to the collected signal, the signal-to-noise ratio is improved, and the optimized fluorescence signal intensity change rate is obtained; the adaptive threshold algorithm is used to analyze the optimized fluorescence signal intensity change rate, the machine learning model trained based on a large amount of historical detection data is combined to predict the fluorescence signal intensity change rate, generate the preliminary concentration prediction value of ATP in the sample to be tested, effectively correct the detection deviation caused by reagent batch difference, and generate the corrected preliminary concentration prediction value; according to the corrected preliminary concentration prediction value, a dynamic range expansion algorithm is implemented to automatically adjust the fluorescence detection parameters to adapt to the ATP detection requirements in different concentration ranges, re-detect and optimize the fluorescence signal, and obtain the final optimized detection result to ensure that the detection accuracy is not limited by the change of sample concentration; after the detection is completed, the final optimized detection result is protected, the encrypted data is sent to the cloud server through a safe wireless communication protocol, the real-time monitoring, trend prediction and abnormal alarm functions are realized, and a comprehensive biological safety monitoring report is generated.

[0067] The technical scheme of the present application has the following beneficial effects:

[0068] The present application provides an automatic method based on ATP fluorescence detection, which significantly improves the sensitivity, accuracy and dynamic range of ATP detection through the combination of microfluidic technology, high-sensitivity photoelectric detection, digital signal processing and machine learning algorithm. This method can effectively eliminate environmental interference noise, improve signal-to-noise ratio, and correct the deviation caused by reagent batch difference through adaptive threshold algorithm and machine learning model, ensuring the consistency and accuracy of the detection results. In addition, through the dynamic range expansion algorithm to automatically adjust the detection parameters, the system can adapt to the sample detection in different concentration ranges, so as to ensure that accurate results can be obtained in a wide concentration range. Finally, through a safe data transmission protocol, the encrypted data is sent to the cloud, realizing real-time monitoring, trend prediction and abnormal alarm functions, and enhancing the ability of biological safety monitoring.

[0069] Further, the application further refines the process of analyzing the rate of change of fluorescence signal intensity using adaptive threshold algorithms and machine learning models. By performing time series analysis on the optimized fluorescence signal, extracting the trend characteristics over time, and applying adaptive threshold algorithms to distinguish between valid changes and noise, the true fluorescence signal changes can be more accurately identified. Subsequently, using a machine learning model trained based on a large amount of historical data, the effective change part is processed, not only generating a preliminary concentration prediction value, but also reducing the influence of the response characteristic differences between different reagent batches through correction processing, greatly improving the accuracy and consistency of the detection results. This method enhances the robustness of the system, ensuring that even when using different batches of reagents, high levels of detection accuracy can be maintained.

[0070] Further, the application describes specific steps for implementing dynamic range expansion based on the corrected preliminary concentration prediction value. By evaluating the approximate concentration range of ATP in the sample to be tested, selecting the appropriate dynamic range expansion algorithm and adjusting the fluorescence detection parameters, the system can make the best detection configuration for the current concentration range. This dynamic adjustment mechanism allows the device to maintain good detection performance when facing extremely low or extremely high concentration samples, avoiding the detection accuracy decline problem caused by traditional fixed parameter settings. After further signal processing and analysis, the final detection results are optimized, ensuring that reliable and accurate ATP concentration measurement values can be obtained regardless of sample concentration changes. This not only improves the quality of single detection, but also provides a solid foundation for subsequent data analysis.

[0071] These aspects or other aspects of the application will be more apparent in the following description of the embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0072] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0073] Figure 1 A flowchart of an automatic detection method based on ATP fluorescence detection provided by an embodiment of the application;

[0074] Figure 2 A structural schematic diagram of an automatic detection system based on ATP fluorescence detection provided by an embodiment of the application;

[0075] Figure 3 A structural schematic diagram of a computing device provided by an embodiment of the application. DETAILED DESCRIPTION

[0076] In order for those skilled in the art to better understand the technical solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described in the following with reference to the accompanying drawings.

[0077] In some of the processes described in the specification and the claims of the present application and in the above-described accompanying drawings, a plurality of operations appearing in a specific order are included, but it should be clearly understood that these operations can be executed or in parallel without the order in which they appear in this text, and the serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes can include more or fewer operations, and these operations can be executed in sequence or in parallel. It should be noted that the "first", "second", etc. described herein are used to distinguish different messages, devices, modules, etc., and do not represent the order of precedence. Also, "first" and "second" are different types.

[0078] The technical solutions in the embodiments of the present application will be clearly and completely described in the following with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0079] Figure 1 A flowchart of an automatic detection method based on ATP fluorescence detection is provided for the embodiments of the present application, as shown in Figure 1 The method comprises:

[0080] 101, obtaining a sample to be tested, controlling the flow rate and direction of the liquid on the microfluidic chip, and uniformly mixing the sample to be tested with the pre-mixed fluorescently labeled ATP detection reagent to generate a mixed solution;

[0081] In this step, the sample to be tested refers to a liquid sample, such as water, that needs to be detected by ATP fluorescence. The microfluidic chip is a technology platform that can accurately control the flow of micro-liquid. It uses micro-channels, valves and pumps to achieve fine control of the liquid. The fluorescently labeled ATP detection reagent is a chemical substance that can specifically bind to ATP and emit a fluorescent signal.

[0082] First, collect the water sample that needs to be detected for ATP concentration as the sample to be tested. Then, introduce the sample into the microfluidic chip and add the pre-mixed fluorescently labeled ATP detection reagent. Finally, adjust the flow rate and flow direction on the microfluidic chip to ensure that the sample and the reagent are mixed uniformly to form a mixed solution, which is ready for subsequent fluorescence signal detection.

[0083] In the present application example, it is assumed that at the water quality monitoring station, the staff uses a specially designed microfluidic chip system after sampling from the river or lake. This system automatically completes the pretreatment of the sample, including mixing with fluorescently labeled ATP detection reagents, to quickly and accurately assess the level of microbial activity in the water body.

[0084] 102. Based on the mixed solution, the fluorescence signal intensity is detected at a specific wavelength using a high-sensitivity photomultiplier tube, the background fluorescence value is synchronously collected and recorded, environmental interference noise is eliminated by applying digital signal processing technology to the collected signal, the signal-to-noise ratio is improved, and the optimized fluorescence signal intensity change rate is obtained.

[0085] In this step, the photomultiplier tube is an extremely sensitive photodetector that can convert the received optical signal into an electrical signal and amplify the output. The specific wavelength refers to the specific light wavelength used to excite the fluorescently labeled substance to emit light. The background fluorescence value is the fluorescence signal caused by environmental factors without the sample to be tested.

[0086] First, after the mixed solution enters the detection area, the fluorescently labeled substance will emit fluorescence under specific wavelength illumination. Second, the high-sensitivity photomultiplier tube is responsible for capturing these fluorescence signals and converting them into quantifiable electrical signals. At the same time, the system also records the background fluorescence value so that subsequent digital signal processing techniques can remove environmental noise and improve the signal-to-noise ratio, resulting in a clearer fluorescence signal intensity change rate.

[0087] In the present application example, when the water sample is preliminarily treated and mixed with the fluorescently labeled ATP detection reagent, it is sent into a closed detection chamber. Here, a specific light source is used to excite the fluorescence reaction, and a photomultiplier tube is used to capture the generated fluorescence signal. This process helps the monitoring station to monitor the real-time microbial pollution situation in the water, especially in the water source protection area.

[0088] Optionally, step 102, which involves using a high-sensitivity photomultiplier tube to detect fluorescence signal intensity at a specific wavelength based on the mixture, simultaneously acquiring and recording background fluorescence values, and applying digital signal processing technology to the acquired signals to eliminate environmental interference noise and improve the signal-to-noise ratio to obtain an optimized fluorescence signal intensity change rate, includes: using the mixture to control the high-sensitivity photomultiplier tube to operate at a specific wavelength, detecting the fluorescence signal intensity of the mixture, and simultaneously acquiring and recording background fluorescence values ​​to obtain raw fluorescence signal data; performing signal processing based on the raw fluorescence signal data to remove environmental interference noise and improve signal clarity, generating a pre-processed fluorescence signal; using the pre-processed fluorescence signal and background fluorescence values ​​to perform background fluorescence correction processing to eliminate the influence of background fluorescence on the signal, obtaining a corrected fluorescence signal; and further optimizing the signal quality based on the corrected fluorescence signal to ensure that the signal-to-noise ratio of the fluorescence signal reaches the optimal level, generating an optimized fluorescence signal.

[0089] In this step, a high-sensitivity photomultiplier tube is a high-sensitivity photodetector capable of detecting weak light signals, commonly used in fluorescence detection. A specific wavelength refers to the specific wavelength of light used to excite the fluorescently labeled substance to emit light. Background fluorescence value is the fluorescence signal caused by environmental factors or the system itself in the absence of a sample. Digital signal processing technology is a series of techniques used to process and analyze digital signals, such as filtering and noise reduction, to improve signal quality. The signal-to-noise ratio (SNR) is the ratio of signal strength to noise strength; a higher SNR indicates a clearer and more reliable signal.

[0090] First, the high-sensitivity photomultiplier tube is controlled to operate at a specific wavelength by using a mixed liquid, and the fluorescence signal intensity of the mixed liquid is detected. At the same time, the background fluorescence value is collected and recorded to obtain the original fluorescence signal data.

[0091] Secondly, based on the original fluorescence signal data, environmental interference noise is removed by applying digital signal processing technology to improve the clarity of the signal and generate a pre-processed fluorescence signal.

[0092] Next, using the pre-processed fluorescence signal and background fluorescence value, background fluorescence correction processing is performed to eliminate the influence of background fluorescence on the signal, and the corrected fluorescence signal is obtained.

[0093] Finally, based on the corrected fluorescence signal, the signal quality was further optimized to ensure that the signal-to-noise ratio of the fluorescence signal reached the optimal state, generating an optimized fluorescence signal to provide a high-quality data foundation for subsequent analysis.

[0094] In this embodiment, it is assumed that in the field of water quality monitoring, such as drinking water safety testing, water samples are collected from a water supply system as test samples. First, the water sample is mixed with fluorescently labeled ATP detection reagent to form a mixture. A high-sensitivity photomultiplier tube is used to detect the fluorescence signal intensity in this mixture at a specific wavelength (e.g., 485nm excitation light), while simultaneously recording the background fluorescence value. Digital signal processing techniques, such as wavelet transform or Kalman filtering, are used to remove environmental interference noise and improve signal clarity. Next, the fluorescence signal is corrected using the background fluorescence value to eliminate the influence of background fluorescence on the measurement results. Finally, by further optimizing signal processing, such as using adaptive filtering technology, the signal-to-noise ratio of the fluorescence signal is ensured to be optimal, thereby obtaining high-quality fluorescence signal data. This data can be used to accurately assess the activity level of microorganisms in the water sample, and thus determine the safety of the water quality.

[0095] 103. The optimized fluorescence signal intensity change rate is analyzed using an adaptive threshold algorithm. Combined with a machine learning model trained based on a large amount of historical detection data, the fluorescence signal intensity change rate is predicted to generate a preliminary concentration prediction value of ATP in the sample to be tested. At the same time, the detection deviation caused by the difference between reagent batches is effectively corrected to generate a corrected preliminary concentration prediction value.

[0096] In this step, the adaptive thresholding algorithm is a method that dynamically adjusts the threshold based on the actual situation to distinguish between valid signals and noise. The machine learning model is a predictive model trained on historical data, capable of recognizing patterns and making predictions.

[0097] First, an adaptive thresholding algorithm is used to analyze the optimized fluorescence signal intensity change rate and separate the true signal component. Then, a pre-trained machine learning model is used to predict the signal intensity change rate, generating a preliminary estimate of ATP concentration. Furthermore, considering potential differences between reagent batches, a correction mechanism is implemented to further refine the prediction results, yielding a more accurate and reliable preliminary concentration prediction.

[0098] In this application example, it is assumed that during water quality testing, the system employs an adaptive threshold algorithm to eliminate interference from non-target signals, ensuring the accuracy of ATP concentration measurement. Simultaneously, leveraging machine learning technology, the system can continuously optimize its predictive capabilities based on a large amount of past experimental data, maintaining consistent detection accuracy even with different batches of reagents. This is crucial for long-term monitoring of water quality changes.

[0099] Optionally, step 103, which involves analyzing the optimized fluorescence signal intensity change rate using an adaptive threshold algorithm and combining it with a machine learning model trained on a large amount of historical detection data to predict the fluorescence signal intensity change rate and generate a preliminary predicted value of ATP concentration in the sample, while effectively correcting detection deviations caused by reagent batch differences to generate a corrected preliminary concentration prediction value, includes: using the optimized fluorescence signal to perform time series analysis, extracting the trend characteristics of fluorescence signal changes over time, and obtaining time series data of fluorescence signal intensity change rate; based on the time series data of fluorescence signal intensity change rate, applying an adaptive threshold algorithm to predict the fluorescence signal intensity change rate and generating a preliminary predicted value of ATP concentration, the following steps are taken: An adaptive threshold algorithm dynamically sets a threshold to distinguish between effective and noise components in the signal, obtaining the fluorescence signal intensity change rate of the effective change component. Combined with a machine learning model trained on a large amount of historical detection data, the fluorescence signal intensity change rate of the effective change component is predicted to generate a preliminary predicted value of ATP concentration in the sample. Within the machine learning model, based on the differences in fluorescence response characteristics between different reagent batches in historical data, the preliminary predicted value of ATP concentration in the sample is corrected to generate a corrected preliminary predicted value, thereby reducing the impact of reagent batch differences on the detection results and improving the accuracy and consistency of the detection results.

[0100] Optionally, step 103, which involves applying an adaptive thresholding algorithm to dynamically set a threshold based on the time-series data of the fluorescence signal intensity change rate to distinguish between the effective change portion and the noise portion of the signal, and obtaining the fluorescence signal intensity change rate of the effective change portion, includes: using the time-series data of the fluorescence signal intensity change rate to calculate the statistical characteristics of the time-series data, including the mean and standard deviation, to obtain basic parameters; applying the adaptive thresholding algorithm based on the basic parameters to dynamically calculate the threshold at each time point to obtain a dynamic threshold, wherein the threshold at each time point can be adjusted with the fluctuation of the time-series data to adaptively identify the effective change portion and the noise portion of the signal; based on the dynamically calculated threshold at each time point, filtering the time-series data of the fluorescence signal intensity change rate, removing data points below the threshold, which are considered to belong to the noise portion, to obtain the fluorescence signal intensity change rate after preliminary filtering; retaining data points higher than or equal to the dynamic threshold, which are considered to represent the effective change portion of the signal, and generating the fluorescence signal intensity change rate of the effective change portion from the fluorescence signal intensity change rate after preliminary filtering.

[0101] In this step, the adaptive thresholding algorithm is an algorithm that dynamically adjusts the threshold based on data characteristics to distinguish between effective changes and noise in the signal. Time series analysis analyzes data that changes over time to extract its trend characteristics. Time series data of fluorescence signal intensity change rate records the data sequence of fluorescence signal intensity changes over time. The machine learning model is a model trained on a large amount of historical detection data, capable of predicting the relationship between fluorescence signal intensity change rate and ATP concentration. Reagent batch variation refers to the variation in fluorescence response characteristics caused by subtle differences in production conditions, composition, etc., between different batches of reagents.

[0102] First, using the optimized fluorescence signal, time series analysis is performed on the fluorescence signal to extract the trend characteristics of the fluorescence signal changing over time, and time series data of the fluorescence signal intensity change rate are generated.

[0103] Secondly, based on the time series data of fluorescence signal intensity change rate, an adaptive threshold algorithm is applied. By calculating statistical characteristics such as mean and standard deviation, the threshold at each time point is dynamically set to distinguish the effective change part from the noise part in the signal, and the fluorescence signal intensity change rate of the effective change part is obtained.

[0104] Next, using a machine learning model trained on a large amount of historical detection data, the rate of change of fluorescence signal intensity in the effective change portion is predicted to generate a preliminary predicted value of ATP concentration in the sample to be tested.

[0105] Finally, in the machine learning model, based on the differences in fluorescence response characteristics between different reagent batches in historical data, the preliminary predicted value of ATP concentration in the test sample is corrected to generate a corrected preliminary predicted value, thereby reducing the impact of reagent batch differences on the test results and improving the accuracy and consistency of the test results.

[0106] In this embodiment, it is assumed that in the field of water quality monitoring, such as drinking water safety testing, water samples are collected from a water supply system as test samples. First, a high-sensitivity photomultiplier tube is used to detect the fluorescence signal intensity in the mixture at a specific wavelength, and after optimization using digital signal processing technology, an optimized fluorescence signal is obtained. Next, time-series analysis is performed on the optimized fluorescence signal to extract the trend characteristics of fluorescence signal changes over time, forming time-series data of the fluorescence signal intensity change rate. Using this data, an adaptive threshold algorithm is applied, dynamically setting the threshold for each time point by calculating statistical characteristics such as the mean and standard deviation, to screen out the effective change portion of the fluorescence signal intensity change rate. Then, combined with a pre-trained machine learning model, the fluorescence signal intensity change rate of the effective change portion is predicted, generating a preliminary predicted value of the ATP concentration in the water sample. Finally, considering the possible differences in fluorescence response characteristics between different batches of reagents, a correction mechanism is introduced into the machine learning model to correct the preliminary predicted value of the ATP concentration in the test sample, generating a corrected preliminary concentration prediction value. This ensures that consistent and accurate detection results can be obtained even when using different batches of reagents, thereby better assessing the safety of water quality.

[0107] This application considers that, in automated detection methods based on ATP fluorescence detection, an adaptive thresholding algorithm is employed to distinguish the effective variation portion from the noise portion in time-series data of fluorescence signal intensity change rates. This algorithm achieves this goal by dynamically calculating the threshold at each time point. This method can dynamically adjust the threshold based on the local and global characteristics of the time-series data, thereby more accurately identifying the effective components in the signal.

[0108] Optionally, step 103, which involves applying an adaptive thresholding algorithm to dynamically set a threshold based on the time-series data of the fluorescence signal intensity change rate to distinguish between the effective change portion and the noise portion of the signal, and obtaining the fluorescence signal intensity change rate of the effective change portion, includes:

[0109] Using the time series data of the fluorescence signal intensity change rate, calculate the local statistical characteristics of the time series data of the fluorescence signal intensity change rate within the time window [t-Δt, t+Δt], including the local mean μ. loc (t) and local standard deviation σ loc (t), to obtain the local basic parameters;

[0110] According to the local basic parameter μ loc (t) and σ loc(t), combining the global statistical properties μ and σ, and considering the different importance of different time points in the time series data, an adaptive threshold algorithm is applied, introducing a time weight factor w(t) and a nonlinear adjustment factor f(x), to dynamically calculate the threshold T(t) for each time point t. The threshold for each time point is calculated using the following formula:

[0111] T(t)=f(μ+w(t)·k·(σ loc (t)) α )

[0112] Where T(t) represents the threshold at each time point t, μ represents the global mean, which is the average rate of change of fluorescence signal intensity throughout the entire time series; σ represents the global standard deviation, which represents the degree of fluctuation in the rate of change of fluorescence signal intensity throughout the entire time series; k is an adjustable coefficient used to control the sensitivity of the threshold, and α is an adjustable exponent used to adjust the influence of the local standard deviation on the threshold. This is the time weighting factor, used to reflect the importance of data at different time points. β represents a parameter that controls the rate of change of the weight over time, and is adjusted according to actual application needs. t0 represents the reference time point, usually chosen as the center or starting point of the time series. f(x) = x + γ·(x - μ) 2 It is a nonlinear function used to increase the nonlinear adjustment capability of the threshold. γ is an adjustable coefficient used to control the degree of nonlinear adjustment, which is adjusted according to the actual application requirements. T(t) can be adjusted with the fluctuation of time series data to adaptively identify the effective change part and noise part in the signal.

[0113] Based on the threshold T(t) of each time point t calculated dynamically, the time series data of the fluorescence signal intensity change rate is filtered to remove data points below the threshold T(t), which are considered to belong to the noise part, thus obtaining the fluorescence signal intensity change rate after preliminary filtering.

[0114] Data points that are higher than or equal to the dynamic threshold T(t) are retained, and these data points are considered to represent the effective change portion of the signal. The fluorescence signal intensity change rate of the effective change portion is generated from the fluorescence signal intensity change rate after the initial screening.

[0115] By calculating the local mean and local standard deviation within a time window, local characteristics of the data can be better captured. Incorporating the global mean and standard deviation ensures that threshold setting depends not only on local information but also on the overall characteristics of the entire time series. Introducing a time weighting factor w(t) reflects the importance of different time points, making more recent data points have a greater impact on the threshold. Using a nonlinear function f(x) increases the nonlinear adjustment capability of the threshold, making it more adaptable to changes in signal trends.

[0116] The following is a brief introduction to the design rationale behind each term of the formula:

[0117] In the threshold formula T(t) at each time point, w(t)·k·(σ) loc (t)) α This sub-item dynamically adjusts the threshold by combining time weights, adjustable coefficients, and powers of local standard deviations to adapt to the importance and local fluctuation characteristics of the signal at different time points.

[0118] The following is a brief introduction to how the parameters of this formula are obtained:

[0119] Wherein, the local mean μ loc (t) and local standard deviation σ loc (t) is calculated from the data within the time window [t-Δt, t+Δt]. The global mean μ and global standard deviation σ are calculated from the entire time series data. The adjustable coefficient k and exponent α are determined based on experimental or historical data analysis. In the time weighting factor, parameters β and t0 are used; β can be set according to actual application requirements, and t0 is usually selected as the center point or starting point of the time series. The parameter γ in the nonlinear function is determined based on experimental or historical data analysis.

[0120] Suppose we have a time series data of fluorescence signal intensity change rate with a length of 100 time points, a time window width Δt = 10, a global mean μ = 5.0, a global standard deviation σ = 2.0, an adjustable coefficient k = 2.0, an exponent α = 1.5, a parameter β = 0.1 in the time weighting factor, a reference time point t0 = 50, and a parameter γ = 0.5 in the nonlinear function.

[0121] Calculate the local mean and local standard deviation:

[0122] For each time point t, calculate the local mean μ within the time window [t-10, t+10]. loc (t) and local standard deviation σ loc (t).

[0123] Calculate the time weighting factor w(t):

[0124] w(t) = e -0.1|t-50|;

[0125] Calculate the threshold T(t) at each time point:

[0126] T(t)=f(5.0+w(t)·2.0·(σ loc (t)) 1.5 );

[0127] Where f(x) = x + 0.5 * (x - 5.0) 2

[0128] Filtering process:

[0129] Remove data points below the threshold T(t) and retain data points above or equal to the threshold T(t).

[0130] Suppose that at time point t = 50, the local mean μ loc (50) = 5.5, local standard deviation σ loc (50) = 2.5;

[0131] Suppose that at time point t = 50, the local mean μ loc (50) = 5.5, local standard deviation σ loc (50) = 2.5;

[0132] Time weighting factor w(50)=e -0.1|50-50| =e 0 =1.0

[0133] Threshold T(50) = f(5.0 + 1.0 · 2.0 · (2.5)) 1.5 )

[0134] T(50)=f(5.0+2.0·2.5 1.5 )

[0135] T(50)=f(5.0+2.0·3.9528)=f(5.0+7.9056)=f(12.9056)

[0136] T(50)=12.9056+0.5·(12.9056-5.0) 2

[0137] T(50) = 12.9056 + 0.5 * 7.9056 2

[0138] T(50) = 12.9056 + 0.5·62.496

[0139] T(50) = 12.9056 + 31.248

[0140] T(50)=44.1536

[0141] At time point t=50, the calculated threshold T(50)=44.1536.

[0142] If the rate of change of fluorescence signal intensity at this time point is less than 44.1536, then this part of the data is considered noise; if it is greater than or equal to 44.1536, then it is considered a valid change.

[0143] By setting such a threshold, noise can be effectively removed while retaining useful signals, thereby improving the accuracy of subsequent analysis.

[0144] In the field of water quality testing, this threshold setting method can help to more accurately identify changes in microbial activity in water samples, thereby assessing water safety and the degree of pollution. For example, in drinking water safety monitoring, this method can more reliably detect microbial contamination in water, ensuring the safety of the water supply system.

[0145] 104. Based on the corrected preliminary concentration prediction value, implement the dynamic range expansion algorithm to automatically adjust the fluorescence detection parameters to adapt to the ATP detection requirements in different concentration ranges, re-detect and optimize the fluorescence signal to obtain the final optimized detection result, so as to ensure that the detection accuracy is not limited by the change of sample concentration.

[0146] In this step, the dynamic range extension algorithm is a technique that automatically adjusts detection parameters to expand the effective measurement range of the detection device. Exposure time and gain settings are key parameters affecting the sensitivity of fluorescence signal detection.

[0147] First, based on the preliminary concentration prediction obtained in the previous step, the system determines the approximate concentration range of the current sample. Then, an appropriate dynamic range expansion algorithm is selected, and key detection parameters such as exposure time and gain are adjusted to ensure optimal detection performance. Next, fluorescence signal detection is re-executed to obtain better data quality, ultimately yielding accurate ATP concentration measurements.

[0148] In this application example, for water quality testing, the dynamic range expansion algorithm allows the system to flexibly handle water samples with different levels of pollution. For example, when processing clear water sources with low pollution levels, the system may increase the exposure time to capture weak fluorescence signals; while when processing turbid water sources with high pollution levels, the exposure time may be appropriately reduced to avoid signal overload. In this way, the reliability and accuracy of the test results can be guaranteed regardless of changes in water quality.

[0149] Optionally, step 104, which involves implementing a dynamic range expansion algorithm based on the corrected preliminary concentration prediction value to automatically adjust the fluorescence detection parameters to adapt to the ATP detection requirements within different concentration ranges, re-detecting and optimizing the fluorescence signal to obtain the final optimized detection result, and ensuring that the detection accuracy is not limited by changes in sample concentration, includes: evaluating the preliminary concentration prediction value of ATP in the sample to obtain an approximate concentration range of ATP in the sample; selecting a corresponding dynamic range expansion algorithm based on the approximate concentration range of ATP in the sample to select and adjust the fluorescence detection parameters to determine a suitable dynamic range expansion algorithm for the current situation. The system sets the exposure time and gain parameters within the initial concentration range. Based on these parameters, it automatically adjusts the operating state of the high-sensitivity photomultiplier tube to re-detect the fluorescence signal intensity of the mixture, obtaining a fluorescence signal under optimized detection conditions. Using this optimized fluorescence signal, it performs digital signal processing again to eliminate environmental interference noise, improve the signal-to-noise ratio, and generate a further optimized fluorescence signal. Based on this further optimized fluorescence signal, it applies an adaptive threshold algorithm and a machine learning model to predict the rate of change of fluorescence signal intensity, generating the final optimized detection result and ensuring that detection accuracy is not limited by changes in sample concentration.

[0150] Optionally, step 104, which involves selecting a suitable dynamic range expansion algorithm based on the approximate concentration range of ATP in the sample to be tested, selecting and adjusting the fluorescence detection parameters, and determining the working parameters for exposure time and gain settings suitable for the current concentration range, includes: using the approximate concentration range of ATP in the sample to be tested to search the dynamic range expansion algorithm library, finding the most suitable dynamic range expansion algorithm for the current concentration range, and generating the selected dynamic range expansion algorithm; and based on the selected dynamic range expansion algorithm, performing simulation analysis on the fluorescence detection parameters to determine the exposure time and gain settings that can provide the best detection performance within the current ATP concentration range, and generating the working parameters.

[0151] Optionally, step 104, which involves applying an adaptive threshold algorithm and machine learning model again to predict the rate of change of fluorescence signal intensity based on the further optimized fluorescence signal, and generating the final optimized detection result to ensure that the detection accuracy is not limited by changes in sample concentration, includes: performing time series analysis on the further optimized fluorescence signal to extract the features of fluorescence signal changes over time, obtaining optimized fluorescence signal feature data; applying the adaptive threshold algorithm again based on the optimized fluorescence signal feature data to dynamically set a threshold to distinguish between the effective change portion and the noise portion of the signal, obtaining the rate of change of fluorescence signal intensity of the effective change portion; and using the effective change portion... The rate of change in fluorescence signal intensity is used as an input feature. Combined with a machine learning model trained on a large amount of historical detection data, the rate of change in fluorescence signal intensity is predicted to generate a final predicted concentration of ATP in the sample. Within the machine learning model, the final predicted concentration is corrected based on the differences in fluorescence response characteristics between different reagent batches in historical data, generating a corrected final predicted concentration to reduce the impact of reagent batch differences on the detection results and improve the accuracy and consistency of the results. Based on the corrected final predicted concentration, the final optimized detection result is generated, ensuring that the detection results maintain high accuracy and reliability regardless of changes in the concentration of ATP in the sample.

[0152] In this step, the dynamic range extension algorithm is an algorithm that automatically adjusts the detection parameters according to the sample concentration range to ensure high-quality fluorescence signals are obtained across different concentration ranges. Exposure time is the length of time the photomultiplier tube acquires the fluorescence signal, affecting signal intensity and signal-to-noise ratio. Gain setting is the factor by which the photomultiplier tube amplifies the signal, affecting signal sensitivity and noise level. Operating parameters refer to the specific parameters used to control the operation of the photomultiplier tube, including exposure time and gain settings.

[0153] First, the preliminary concentration prediction of ATP in the sample is used to evaluate and process the preliminary concentration prediction to obtain the approximate concentration range of ATP in the sample.

[0154] Secondly, based on the approximate concentration range of ATP in the sample to be tested, the most suitable dynamic range expansion algorithm is searched and selected from the dynamic range expansion algorithm library.

[0155] Next, the fluorescence detection parameters are simulated and analyzed using the selected dynamic range extension algorithm to determine the exposure time and gain settings that provide the best detection performance, and to generate working parameters suitable for the current concentration range.

[0156] Furthermore, based on the generated operating parameters, the operating state of the high-sensitivity photomultiplier tube is automatically adjusted, and the fluorescence signal intensity of the mixture is re-detected to obtain the fluorescence signal under optimized detection conditions.

[0157] Furthermore, by utilizing the fluorescence signal under optimized detection conditions, digital signal processing technology is applied again to eliminate environmental interference noise, improve the signal-to-noise ratio, and generate a further optimized fluorescence signal.

[0158] Finally, based on the further optimized fluorescence signal, the adaptive threshold algorithm and machine learning model are applied again to predict the rate of change of fluorescence signal intensity, generating the final optimized detection result and ensuring that the detection accuracy is not limited by changes in sample concentration.

[0159] In this embodiment, it is assumed that in the field of water quality monitoring, such as drinking water safety testing, water samples are collected from a water supply system as test samples. First, the corrected preliminary concentration prediction value generated in step 103 is used to assess the approximate concentration range of ATP in the test sample. Based on this approximate concentration range, an algorithm best suited to the current concentration range is selected from the dynamic range expansion algorithm library. For example, a high gain and low exposure time setting is used for the low concentration range, while a low gain and high exposure time setting is used for the high concentration range.

[0160] Secondly, the fluorescence detection parameters are simulated and analyzed using the selected dynamic range extension algorithm to determine the optimal exposure time and gain settings, thus generating the operating parameters. For example, for the low concentration range, a longer exposure time and higher gain may be selected; while for the high concentration range, a shorter exposure time and lower gain may be selected.

[0161] Furthermore, the operating parameters are used to automatically adjust the working state of the high-sensitivity photomultiplier tube, and the fluorescence signal in the mixture is re-detected to obtain the fluorescence signal under optimized detection conditions. Next, digital signal processing technology is used to further remove environmental interference noise, improve the signal-to-noise ratio, and generate a further optimized fluorescence signal.

[0162] Finally, based on the further optimized fluorescence signal, an adaptive thresholding algorithm and a pre-trained machine learning model are applied again to predict the rate of change of fluorescence signal intensity, generating the final optimized detection result. This ensures accurate and reliable detection results regardless of changes in the ATP concentration in the water sample, thereby better assessing water quality safety.

[0163] This application considers that in automated detection methods based on ATP fluorescence detection, it is necessary to dynamically adjust fluorescence detection parameters to adapt to the detection requirements of ATP within different concentration ranges. By selecting a suitable dynamic range expansion algorithm and combining it with simulation analysis to determine the optimal operating parameters, optimal detection performance can be ensured across different concentration ranges. The performance index P(E, G) is used to comprehensively evaluate the effects of different parameter combinations.

[0164] Optionally, step 104, which involves selecting a suitable dynamic range expansion algorithm based on the approximate concentration range of ATP in the sample to be tested, and selecting and adjusting the fluorescence detection parameters to determine the appropriate exposure time and gain settings for the current concentration range, includes:

[0165] Using the approximate concentration range C of ATP in the sample to be tested range， The dynamic range expansion algorithm library is searched to find the most suitable dynamic range expansion algorithm for the current concentration range, and the selected dynamic range expansion algorithm is generated.

[0166] Based on the selected dynamic range expansion algorithm, the fluorescence detection parameters are simulated and analyzed to determine the exposure time and gain settings that provide optimal detection performance within the current ATP concentration range, generating working parameters, specifically including:

[0167] Using the approximate concentration range C of ATP in the sample to be tested range Select the most suitable algorithm A(C) from the dynamic range expansion algorithm library. range Generate the selected dynamic range expansion algorithm;

[0168] Based on the selected dynamic range expansion algorithm A(C) range The fluorescence detection parameters were simulated and analyzed to calculate the performance index P(E, G) for each parameter combination, where E represents the exposure time and G represents the gain setting.

[0169] The performance index P(E, G) is calculated using the following formula:

[0170]

[0171] Where S(E, G) represents the signal intensity, obtained through simulation, reflecting the intensity of the fluorescence signal under a given exposure time and gain setting; D(E, G) represents the signal-to-noise ratio, obtained through simulation, reflecting the signal-to-noise ratio of the fluorescence signal under a given exposure time and gain setting; N(E, G) represents the noise level, obtained through simulation, reflecting the noise level of the fluorescence signal under a given exposure time and gain setting; T(E, G) represents the detection time, obtained through simulation, reflecting the time required to complete one detection under a given exposure time and gain setting; w1(C range w2(C) range w3(C) range ) and w4(C range ) is a dynamic weighting factor used to determine the ATP concentration range C. range The impact of different performance indicators is dynamically adjusted according to actual application requirements; α(C range ), β(C range ), γ(C range ) and δ(C range () is an adjustable index used to adjust ATP concentrations within a range C. range The nonlinear effects of each performance index are dynamically adjusted according to actual application requirements;

[0172] The optimal operating parameters (E, G) are generated by using an optimization algorithm to find the exposure time and gain settings that maximize the performance index P(E, G). opt G opt );

[0173] Based on the optimal operating parameters (E) opt G opt This generates the final operating parameters, which are used to automatically adjust the operating status of the high-sensitivity photomultiplier tube to ensure optimal detection performance within the current ATP concentration range.

[0174] The performance metric P(E, G) comprehensively considers multiple factors such as signal strength, signal-to-noise ratio, noise level, and detection time to achieve multi-objective optimization. A dynamic weighting factor w1(C) is introduced. range w2(C) range w3(C) range ) and w4(C range This allows for adjustment of the importance of different performance indicators based on a range of ATP concentrations. An adjustable exponent α (C) is used. range ), β(C range ), γ(C range ) and δ(C range This allows the performance indicators to better reflect the nonlinear changes of each performance indicator under different concentration ranges.

[0175] The following is a brief introduction to the design rationale behind each term of the formula:

[0176] In the performance index formula P(E, G), In this sub-item, signal intensity S(E, G) is a key indicator of detection sensitivity; high signal intensity means that the light signal can be detected more clearly. This is achieved by introducing a dynamic weighting factor w1(C). range ) and the adjustable exponent α (C range This allows you to adjust the degree to which signal strength affects overall performance indicators; In this sub-item, the signal-to-noise ratio (SNR) D(E, G) reflects the ratio of signal to noise. A higher SNR means a clearer signal, which helps improve detection accuracy. This is achieved through a dynamic weighting factor w2(C). range ) and adjustable exponent β (C range This allows for adjustment of the signal-to-noise ratio's impact on overall performance metrics; The noise level N(E, G) in this sub-item reflects the background noise at a given exposure time and gain setting. A lower noise level helps improve detection accuracy. This is achieved through a dynamic weighting factor w3(C). range ) and adjustable exponential γ (C range This allows for adjustment of the impact of noise levels on overall performance indicators; In this sub-item, the detection time T(E, G) reflects the time required to complete one detection. A shorter detection time improves detection efficiency, especially in applications requiring rapid response. This is achieved through a dynamic weighting factor w4(C). range ) and the adjustable exponent δ(C) range This allows for adjustment of the impact of testing time on overall performance indicators;

[0177] The following is a brief introduction to how the parameters of this formula are obtained:

[0178] Among them, the signal strength S(E,G), signal-to-noise ratio D(E,G), noise level N(E,G), and detection time T(E,G) are obtained through simulation or experimental data; the dynamic weighting factor w1(C range w2(C) range w3(C) range ) and w4(C range ): Determined based on actual application needs and historical data analysis; Adjustable index α (C range ), β(C range ), γ(C range ) and δ(C range Similarly, this is determined based on actual application needs and historical data analysis;

[0179] Suppose there is a set of water samples to be tested with an approximate ATP concentration range C. range =10 -6 M to 10 -5 M, and a dynamic range extension algorithm A(C) suitable for this concentration range has been selected. range The optimal exposure time and gain settings will be calculated using the following steps:

[0180] Dynamic weighting factor:

[0181] w1(C range ) = 0.5; w2(C range ) = 0.3; w3(C range ) = 0.1; w4(C range ) = 0.1;

[0182] Adjustable index:

[0183] α(C range )=1.5;β(C range )=1.0;γ(C range )=1.2;δ(C range =1.0;

[0184] Data obtained from simulation:

[0185] E=100ms, G=1000; S(100,1000)=1000; D(100,1000)=50; N(100,1000)=10; T(100,1000)=0.1s;

[0186] Calculate the performance metrics P(E, G):

[0187] P(100, 1000) = 0.5·1000 1.5 +0.3·50 1.0 -0.1·10 1.2 -0.1·0.1 1.0

[0188] P(100,1000)=0.5·1000000+0.3·50-0.1·15.849-0.1·0.1

[0189] P(100,1000)=500000+15-1.5849-0.01

[0190] P(100, 1000) = 500013.4151

[0191] The optimization algorithm finds the maximized P(E, G):

[0192] By iterating through different combinations of exposure time and gain settings, the E and G values ​​that maximize P(E, G) are found. Assume that after optimization, the algorithm finds that P(E, G) reaches its maximum value when E = 120ms and G = 1200.

[0193] Generate optimal operating parameters:

[0194] The optimal operating parameter is E opt =120ms and G opt =1200.

[0195] In the ATP concentration range C range =10 -6 M to 10 -5 Within M, the dynamic range expansion algorithm A(C) is used. range Through optimization of the exposure time and performance metrics P(E, G), the optimal exposure time was determined to be 120ms and the gain was set to 1200.

[0196] These parameters provide optimal detection performance, including high signal strength, good signal-to-noise ratio, low noise level, and reasonable detection time.

[0197] By automatically adjusting the operating status of the high-sensitivity photomultiplier tube, optimal detection performance is ensured within the current ATP concentration range, thereby improving the accuracy and reliability of water quality detection.

[0198] 105. After the test is completed, the final optimized test results are protected, and the encrypted data is sent to the cloud server through a secure wireless communication protocol to realize real-time monitoring, trend prediction and abnormal alarm functions, and generate a comprehensive biosafety monitoring report.

[0199] In this step, the wireless communication protocol is a standard protocol used for data transmission, ensuring the security and reliability of data transmission. The cloud server is a remote server used to store and process data, supporting real-time monitoring, trend prediction, and anomaly alarm functions.

[0200] After the testing is completed, the final optimized test results are protected, and the encrypted data is sent to the cloud server via a secure wireless communication protocol. The cloud server enables real-time monitoring, trend prediction, and anomaly alarm functions, generating a comprehensive biosafety monitoring report.

[0201] In this application example, after ATP concentration detection is completed in water quality monitoring, the encrypted data is sent to a cloud server via a secure wireless communication protocol. The cloud server can display the detection results in real time, perform trend analysis, and issue alarms when anomalies are detected, helping management departments to take timely measures to ensure water quality safety.

[0202] Figure 2This application provides a schematic diagram of the structure of an automated detection system based on ATP fluorescence detection, as shown in the embodiments below. Figure 2 As shown, the system includes:

[0203] The acquisition control module 21 is used to acquire the sample to be tested, control the flow speed and direction of the liquid on the microfluidic chip, and fully and uniformly mix the sample to be tested with the premixed fluorescently labeled ATP detection reagent to generate a mixture.

[0204] The detection and acquisition module 22 is used to detect the fluorescence signal intensity at a specific wavelength based on the mixture using a high-sensitivity photomultiplier tube, simultaneously acquire and record the background fluorescence value, and apply digital signal processing technology to the acquired signal to eliminate environmental interference noise, improve the signal-to-noise ratio of the signal, and obtain the optimized fluorescence signal intensity change rate.

[0205] The analysis and prediction module 23 is used to analyze the optimized fluorescence signal intensity change rate using an adaptive threshold algorithm, and combine it with a machine learning model trained based on a large amount of historical detection data to predict the fluorescence signal intensity change rate, generate a preliminary concentration prediction value of ATP in the sample to be tested, and effectively correct the detection deviation caused by the difference between reagent batches to generate a corrected preliminary concentration prediction value.

[0206] The detection module 24 is used to implement a dynamic range expansion algorithm based on the corrected preliminary concentration prediction value, automatically adjust the fluorescence detection parameters to adapt to the ATP detection requirements in different concentration ranges, re-detect and optimize the fluorescence signal, and obtain the final optimized detection result to ensure that the detection accuracy is not limited by the change of sample concentration.

[0207] The protection generation module 25 is used to protect the final optimized detection result after the detection is completed. It sends the encrypted data to the cloud server through a secure wireless communication protocol to realize real-time monitoring, trend prediction and abnormal alarm functions, and generate a comprehensive biosafety monitoring report.

[0208] Figure 2 The aforementioned automated detection system based on ATP fluorescence detection can perform... Figure 1 The implementation principle and technical effects of the automated detection method based on ATP fluorescence detection described in the illustrated embodiments will not be repeated here. The specific operation methods of each module and unit in the automated detection system based on ATP fluorescence detection in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0209] In one possible design, Figure 2 The automated detection system based on ATP fluorescence detection in the illustrated embodiment can be implemented as a computing device, such as... Figure 3As shown, the computing device may include a storage component 31 and a processing component 32;

[0210] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0211] The processing component 32 is used for: acquiring the sample to be tested; controlling the flow rate and direction of the liquid on the microfluidic chip; thoroughly and uniformly mixing the sample to be tested with premixed fluorescently labeled ATP detection reagent to generate a mixture; based on the mixture, detecting the fluorescence signal intensity at a specific wavelength using a high-sensitivity photomultiplier tube, simultaneously acquiring and recording background fluorescence values; eliminating environmental interference noise and improving the signal-to-noise ratio by applying digital signal processing technology to the acquired signal to obtain an optimized fluorescence signal intensity change rate; analyzing the optimized fluorescence signal intensity change rate using an adaptive threshold algorithm, and combining it with a machine learning model trained based on a large amount of historical detection data to further refine the fluorescence signal intensity change rate. Predictive processing generates a preliminary predicted value of ATP concentration in the sample to be tested, while effectively correcting detection deviations caused by batch differences in reagents, generating a corrected preliminary predicted value. Based on the corrected preliminary predicted value, a dynamic range expansion algorithm is implemented to automatically adjust the fluorescence detection parameters to adapt to the ATP detection requirements within different concentration ranges. The fluorescence signal is re-detected and optimized to obtain the final optimized detection result, ensuring that the detection accuracy is not limited by changes in sample concentration. After the detection is completed, the final optimized detection result is protected, and the encrypted data is sent to the cloud server via a secure wireless communication protocol to achieve real-time monitoring, trend prediction, and anomaly alarm functions, generating a comprehensive biosafety monitoring report.

[0212] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0213] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0214] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0215] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0216] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0217] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0218] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The illustrated embodiment is an automated detection method based on ATP fluorescence detection.

[0219] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0220] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0221] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0222] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An automated detection method based on ATP fluorescence detection, characterized in that, include: The sample to be tested is obtained, and the flow rate and direction of the liquid on the microfluidic chip are controlled to fully and uniformly mix the sample to be tested with the premixed fluorescently labeled ATP detection reagent to generate a mixture. Based on the mixture, a high-sensitivity photomultiplier tube is used to detect the fluorescence signal intensity at a specific wavelength, and the background fluorescence value is collected and recorded simultaneously. By applying digital signal processing technology to the collected signal to eliminate environmental interference noise and improve the signal-to-noise ratio, the optimized fluorescence signal intensity change rate is obtained. The optimized fluorescence signal intensity change rate is analyzed using an adaptive threshold algorithm. Combined with a machine learning model trained based on a large amount of historical detection data, the fluorescence signal intensity change rate is predicted to generate a preliminary predicted value of ATP concentration in the sample to be tested. At the same time, the detection deviation caused by the difference between reagent batches is effectively corrected to generate a corrected preliminary predicted value of concentration. Based on the corrected preliminary concentration prediction, a dynamic range expansion algorithm is implemented to automatically adjust the fluorescence detection parameters to adapt to the ATP detection requirements in different concentration ranges. The fluorescence signal is then re-detected and optimized to obtain the final optimized detection result, ensuring that the detection accuracy is not limited by changes in sample concentration. After the test is completed, the final optimized test results are protected, and the encrypted data is sent to the cloud server through a secure wireless communication protocol to realize real-time monitoring, trend prediction and anomaly alarm functions, and generate a comprehensive biosafety monitoring report. The process of analyzing the rate of change of the optimized fluorescence signal intensity using an adaptive threshold algorithm includes: Using the optimized fluorescence signal, time series analysis is performed on the optimized fluorescence signal to extract the trend characteristics of the fluorescence signal changing over time, and time series data of fluorescence signal intensity change rate are obtained; based on the time series data of fluorescence signal intensity change rate, an adaptive threshold algorithm is applied to dynamically set a threshold to distinguish the effective change part and the noise part in the signal, and the fluorescence signal intensity change rate of the effective change part is obtained. The step of applying an adaptive thresholding algorithm to dynamically set a threshold based on the time-series data of the fluorescence signal intensity change rate to distinguish between the effective change portion and the noise portion of the signal, and obtaining the fluorescence signal intensity change rate of the effective change portion, includes: Using the time series data of the fluorescence signal intensity change rate, calculate the local statistical characteristics of the time series data of the fluorescence signal intensity change rate within the time window [t-Δt, t+Δt], including the local mean μ. loc (t) and local standard deviation σ loc (t), to obtain the local basic parameters; According to the local basic parameter μ loc (t) and σ loc (t), combining the global statistical properties μ and σ, and considering the different importance of different time points in the time series data, an adaptive threshold algorithm is applied, introducing a time weight factor w(t) and a nonlinear adjustment factor f(x), to dynamically calculate the threshold T(t) for each time point t. The threshold for each time point is calculated using the following formula: T(t)=f(μ+w(t)·k·(σ loc (t)) α ) Where T(t) represents the threshold at each time point t, μ represents the global mean, which is the average rate of change of fluorescence signal intensity throughout the entire time series; σ represents the global standard deviation, which represents the degree of fluctuation of the rate of change of fluorescence signal intensity throughout the entire time series; k is an adjustable coefficient used to control the sensitivity of the threshold, and α is an adjustable exponent used to adjust the influence of the local standard deviation on the threshold. w(t) = e -β|t-t0| This is the time weighting factor, used to reflect the importance of data at different time points. β represents a parameter that controls the rate of change of the weight over time, and is adjusted according to actual application needs. t0 represents the reference time point, usually chosen as the center or starting point of the time series. f(x) = x + γ·(x - μ) 2 It is a nonlinear function used to increase the nonlinear adjustment capability of the threshold. γ is an adjustable coefficient used to control the degree of nonlinear adjustment, which is adjusted according to the actual application requirements. T(t) can be adjusted with the fluctuation of time series data to adaptively identify the effective change part and noise part in the signal. Based on the threshold T(t) of each time point t calculated dynamically, the time series data of the fluorescence signal intensity change rate is filtered to remove data points below the threshold T(t), which are considered to belong to the noise part, thus obtaining the fluorescence signal intensity change rate after preliminary filtering. Data points that are higher than or equal to the threshold T(t) are retained, and these retained data points are considered to represent the effective change portion of the signal. The fluorescence signal intensity change rate of the effective change portion is generated from the fluorescence signal intensity change rate after the initial screening.

2. The automated detection method based on ATP fluorescence detection according to claim 1, characterized in that, By combining a machine learning model trained on a large amount of historical detection data, the rate of change of fluorescence signal intensity is predicted to generate a preliminary predicted value of ATP concentration in the sample. Simultaneously, the model effectively corrects for detection bias caused by batch-to-batch reagent differences, generating a corrected preliminary concentration prediction value, including: By combining a machine learning model trained on a large amount of historical detection data, the rate of change of fluorescence signal intensity of the effective change portion is predicted to generate a preliminary predicted value of ATP concentration in the sample to be tested. In the machine learning model, the preliminary concentration prediction of ATP in the test sample is corrected based on the differences in fluorescence response characteristics between different reagent batches in historical data, so as to generate a corrected preliminary concentration prediction value, thereby reducing the impact of reagent batch differences on the test results and improving the accuracy and consistency of the test results.

3. The automated detection method based on ATP fluorescence detection according to claim 2, characterized in that, The time-series data based on the fluorescence signal intensity change rate is used, and an adaptive threshold algorithm is applied to dynamically set a threshold to distinguish between the effective change portion and the noise portion of the signal, thereby obtaining the fluorescence signal intensity change rate of the effective change portion, including: Using the time series data of the fluorescence signal intensity change rate, the statistical characteristics of the time series data of the fluorescence signal intensity change rate, including the mean and standard deviation, are calculated to obtain the basic parameters; Based on the aforementioned basic parameters, an adaptive thresholding algorithm is applied to dynamically calculate the threshold at each time point to obtain a dynamic threshold. The threshold at each time point can be adjusted as the time series data fluctuates, so as to adaptively identify the effective changing part and the noise part in the signal. Based on the dynamically calculated threshold for each time point, the time series data of the fluorescence signal intensity change rate is filtered to remove data points below the threshold, which are considered to be noise, thus obtaining the preliminary filtered fluorescence signal intensity change rate. Data points that are higher than or equal to the dynamic threshold are retained, and these data points are considered to represent the effective change portion of the signal. The fluorescence signal intensity change rate of the effective change portion is generated from the fluorescence signal intensity change rate after the initial screening.

4. The automated detection method based on ATP fluorescence detection according to claim 1, characterized in that, The process involves implementing a dynamic range expansion algorithm based on the corrected preliminary concentration prediction value to automatically adjust fluorescence detection parameters to adapt to ATP detection requirements within different concentration ranges, re-detecting and optimizing the fluorescence signal to obtain the final optimized detection result. This ensures that detection accuracy is not limited by changes in sample concentration. The process includes: Using the preliminary predicted concentration of ATP in the sample to be tested, the preliminary predicted concentration is evaluated to obtain the approximate concentration range of ATP in the sample to be tested. Based on the approximate concentration range of ATP in the sample to be tested, a corresponding dynamic range expansion algorithm is selected, and the fluorescence detection parameters are selected and adjusted to determine the appropriate exposure time and gain settings for the current concentration range. Based on the operating parameters, the operating state of the high-sensitivity photomultiplier tube is automatically adjusted, and the fluorescence signal intensity of the mixture is re-detected to obtain the fluorescence signal under optimized detection conditions. Using the fluorescence signal under the optimized detection conditions, digital signal processing technology is applied again to eliminate environmental interference noise, improve the signal-to-noise ratio, and generate a further optimized fluorescence signal. Based on the further optimized fluorescence signal, the adaptive threshold algorithm and machine learning model are applied again to predict the rate of change of fluorescence signal intensity, generating the final optimized detection result and ensuring that the detection accuracy is not limited by the change of sample concentration.

5. The automated detection method based on ATP fluorescence detection according to claim 4, characterized in that, Based on the approximate concentration range of ATP in the sample to be tested, a corresponding dynamic range expansion algorithm is selected to select and adjust the fluorescence detection parameters, determining the appropriate exposure time and gain settings for the current concentration range, including: Using the approximate concentration range of ATP in the sample to be tested, the dynamic range expansion algorithm library is searched to find the most suitable dynamic range expansion algorithm for the current concentration range, and the selected dynamic range expansion algorithm is generated. Based on the selected dynamic range expansion algorithm, the fluorescence detection parameters are simulated and analyzed to determine the exposure time and gain settings that provide the best detection performance within the current ATP concentration range, and the working parameters are generated.

6. The automated detection method based on ATP fluorescence detection according to claim 4, characterized in that, Based on the approximate concentration range of ATP in the sample to be tested, a corresponding dynamic range expansion algorithm is selected to select and adjust the fluorescence detection parameters, determining the appropriate exposure time and gain settings for the current concentration range, including: Using the approximate concentration range C of ATP in the sample to be tested range The dynamic range expansion algorithm library is searched to find the most suitable dynamic range expansion algorithm for the current concentration range, and the selected dynamic range expansion algorithm is generated. Based on the selected dynamic range expansion algorithm, the fluorescence detection parameters are simulated and analyzed to determine the exposure time and gain settings that provide optimal detection performance within the current ATP concentration range, generating working parameters, specifically including: Using the approximate concentration range C of ATP in the sample to be tested range Select the most suitable algorithm A(C) from the dynamic range extension algorithm library. range Generate the selected dynamic range expansion algorithm; Based on the selected dynamic range extension algorithm A(C) range The fluorescence detection parameters were simulated and analyzed to calculate the performance index P(E, G) for each parameter combination, where E represents the exposure time and G represents the gain setting. The performance index P(E, G) is calculated using the following formula: Where S(E, G) represents the signal intensity, obtained through simulation, reflecting the intensity of the fluorescence signal under a given exposure time and gain setting; D(E, G) represents the signal-to-noise ratio, obtained through simulation, reflecting the signal-to-noise ratio of the fluorescence signal under a given exposure time and gain setting; N(E, G) represents the noise level, obtained through simulation, reflecting the noise level of the fluorescence signal under a given exposure time and gain setting; T(E, G) represents the detection time, obtained through simulation, reflecting the time required to complete one detection under a given exposure time and gain setting; w1(C range w2(C) range w3(C) range ) and w4(C range ) is a dynamic weighting factor used to determine the ATP concentration range C. range The impact of different performance indicators is dynamically adjusted according to actual application requirements; α(C range ), β(C range ), γ(C range ) and δ(C range () is an adjustable index used to adjust ATP concentrations within a range C. range The nonlinear effects of each performance index are dynamically adjusted according to actual application requirements; The optimal operating parameters (E, G) are generated by using an optimization algorithm to find the exposure time and gain settings that maximize the performance index P(E, G). opt G opt ); Based on the optimal operating parameters (E) opt G opt This generates the final operating parameters, which are used to automatically adjust the operating status of the high-sensitivity photomultiplier tube to ensure optimal detection performance within the current ATP concentration range.

7. The automated detection method based on ATP fluorescence detection according to claim 4, characterized in that, Based on the further optimized fluorescence signal, an adaptive threshold algorithm and machine learning model are applied again to predict the rate of change of fluorescence signal intensity, generating the final optimized detection result. This ensures that the detection accuracy is not limited by changes in sample concentration, including: Using the further optimized fluorescence signal, time series analysis was performed to extract the characteristics of fluorescence signal changes over time, and optimized fluorescence signal feature data was obtained. Based on the optimized fluorescence signal feature data, the adaptive threshold algorithm is applied again to dynamically set the threshold to distinguish the effective change part from the noise part in the signal, and the fluorescence signal intensity change rate of the effective change part is obtained. Using the fluorescence signal intensity change rate of the effective change portion as input feature, and combined with a machine learning model trained based on a large amount of historical detection data, the fluorescence signal intensity change rate is predicted to generate the final predicted value of ATP concentration in the sample to be tested. In the machine learning model, the final concentration prediction value is corrected based on the differences in fluorescence response characteristics between different reagent batches in historical data, and a corrected final concentration prediction value is generated to reduce the impact of reagent batch differences on the detection results and improve the accuracy and consistency of the detection results. Based on the corrected final concentration prediction, the final optimized detection result is generated to ensure that the detection result remains highly accurate and reliable regardless of how the concentration of ATP in the sample changes.

8. The automated detection method based on ATP fluorescence detection according to claim 1, characterized in that, Based on the mixture, a high-sensitivity photomultiplier tube is used to detect the fluorescence signal intensity at a specific wavelength, while simultaneously acquiring and recording background fluorescence values. Digital signal processing techniques are applied to the acquired signals to eliminate environmental interference noise and improve the signal-to-noise ratio, resulting in an optimized fluorescence signal intensity change rate, including: Using the mixture, a high-sensitivity photomultiplier tube is controlled to work at a specific wavelength, and the fluorescence signal intensity of the mixture is detected. At the same time, the background fluorescence value is collected and recorded to obtain the original fluorescence signal data. Based on the original fluorescence signal data, signal processing is performed to remove environmental interference noise, improve signal clarity, and generate a pre-processed fluorescence signal. Using the pre-processed fluorescence signal and background fluorescence value, background fluorescence correction processing is performed to eliminate the influence of background fluorescence on the signal, and the corrected fluorescence signal is obtained. Based on the corrected fluorescence signal, the signal quality is further optimized to ensure that the signal-to-noise ratio of the fluorescence signal reaches the best, and an optimized fluorescence signal is generated.

9. An automated detection system based on ATP fluorescence detection, used to execute the automated detection method based on ATP fluorescence detection as described in any one of claims 1 to 8, characterized in that, include: The acquisition control module is used to acquire the sample to be tested, control the flow speed and direction of the liquid on the microfluidic chip, and fully and uniformly mix the sample to be tested with the premixed fluorescently labeled ATP detection reagent to generate a mixture. The detection and acquisition module is used to detect the fluorescence signal intensity at a specific wavelength based on the mixture using a high-sensitivity photomultiplier tube, simultaneously acquire and record the background fluorescence value, and apply digital signal processing technology to the acquired signal to eliminate environmental interference noise, improve the signal-to-noise ratio, and obtain the optimized fluorescence signal intensity change rate. The analysis and prediction module is used to analyze the optimized fluorescence signal intensity change rate using an adaptive threshold algorithm, and combined with a machine learning model trained based on a large amount of historical detection data, to predict the fluorescence signal intensity change rate and generate a preliminary concentration prediction value of ATP in the sample to be tested. At the same time, it effectively corrects the detection deviation caused by the difference between reagent batches and generates a corrected preliminary concentration prediction value. The detection module is adjusted to implement a dynamic range expansion algorithm based on the corrected preliminary concentration prediction value, automatically adjust the fluorescence detection parameters to adapt to the ATP detection requirements in different concentration ranges, re-detect and optimize the fluorescence signal to obtain the final optimized detection result, so as to ensure that the detection accuracy is not limited by the change of sample concentration. The protection generation module is used to protect the final optimized detection results after the detection is completed. It sends the encrypted data to the cloud server through a secure wireless communication protocol to realize real-time monitoring, trend prediction and anomaly alarm functions, and generate a comprehensive biosafety monitoring report.

Images