Veterinary drug residue detection method

Abstract

The invention discloses a veterinary drug residue detection method, and relates to the technical field of veterinary drug residue detection, and the veterinary drug residue detection method comprises the following steps: firstly, based on preset sensitivity, carrying out veterinary drug residue detection on a detection sample, and obtaining veterinary drug residue information in the sample in real time; according to the method, the analysis set is established for the veterinary drug residue information obtained in real time, and the data information in the analysis set is preprocessed, so that the residue data is more accurate, and the reliability of subsequent feature extraction and analysis is ensured. According to the method, the dynamic sensitivity adjustment of the low-concentration veterinary drug residues is realized by combining the preset sensitivity preliminary detection and the machine learning model intelligent evaluation. The system flexibly adjusts sensitivity according to actual conditions of samples to ensure an optimal working state. Through real-time data acquisition, feature extraction and quantification and in combination with machine learning evaluation, low-concentration residues can be accurately identified, and the risk of missing detection is avoided. According to the method, the detection accuracy and reliability in a complex sample matrix are improved, and food safety and consumer health are effectively guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of veterinary drug residue detection, and in particular to a veterinary drug residue detection method. Background Art

[0002] Veterinary drug residue testing refers to the testing of animals and their products (such as meat, dairy products, eggs, etc.) to ensure the presence of residues of veterinary drugs (such as antibiotics, hormones, anthelmintics, etc.) used in the breeding process. These residues may be harmful to the health of consumers, so veterinary drug residue testing is very important. Through scientific testing methods, such as liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS), veterinary drug components in animal products can be detected and their concentrations can be quantified. The test results help to determine whether they meet national or regional safety standards and ensure food safety.

[0003] The existing technology has the following deficiencies: The existing technology usually uses a fixed detection sensitivity when testing the test samples, but some low-concentration veterinary drug residues may be missed. The distribution and metabolism of veterinary drug residues in animals are complex. Some veterinary drug residues may exist at extremely low concentrations. Fixed sensitivity may fail to identify these trace residues, resulting in the failure to timely discover potential hazards. If such trace residues accumulate over a long period of time or become toxic under certain environmental conditions, they may cause long-term chronic effects on consumers' health.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide a method for detecting veterinary drug residues, which realizes dynamic sensitivity adjustment of low-concentration veterinary drug residues by combining preliminary detection with preset sensitivity and intelligent evaluation of machine learning models. The system flexibly adjusts the sensitivity according to the actual situation of the sample to ensure the best working state. Through real-time data acquisition, feature extraction and quantification, combined with machine learning evaluation, low-concentration residues can be accurately identified to avoid the risk of missed detection. This method improves the detection accuracy and reliability in complex sample matrices, effectively guarantees food safety and consumer health, and solves the problems in the above-mentioned background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a method for detecting veterinary drug residues, comprising the following steps:

[0007] First, based on the preset sensitivity, the test sample is tested for veterinary drug residues to obtain the residual information of veterinary drugs in the sample in real time;

[0008] The real-time acquisition of veterinary drug residue information is used to establish an analysis set, and the data information in the analysis set is preprocessed to make the residue data more accurate and ensure the reliability of subsequent feature extraction and analysis;

[0009] Extract key features reflecting low-concentration veterinary drug residues from the pre-processed data, analyze and process the extracted key features under the detection window, convert low-concentration veterinary drug residues into measurable digital signals, and quantify the concentration of veterinary drug residues;

[0010] The key features after feature analysis and quantification are input into the pre-trained machine learning model for further intelligent evaluation. The machine learning model is used to evaluate the concentration of veterinary drug residues in the current sample and intelligently determine whether it is a low-concentration residue.

[0011] When the machine learning model assesses the presence of low-concentration veterinary drug residues in the current sample, the detection sensitivity will be dynamically improved based on the evaluation results to ensure that the detection system is always in the best working state in different detection scenarios and adapts to the complexity of different samples, thereby improving the accuracy and reliability of the detection.

[0012] Preferably, key features reflecting low-concentration veterinary drug residues are extracted from the preprocessed data, and the extracted features include the chromatographic peak width of the residue and the ratio of different isomers. Under the detection window, the obtained chromatographic peak width of the residue and the ratio of different isomers are analyzed to generate a residue chromatographic peak width reference value and an isomer ratio reference value, respectively. The concentration level of veterinary drug residues is quantified by the residue chromatographic peak width reference value and the isomer ratio reference value, thereby providing data support for the accurate assessment of low-concentration veterinary drug residues.

[0013] Preferably, the specific steps of analyzing the chromatographic peak width of the residue under the detection window to generate the reference value of the chromatographic peak width of the residue are as follows:

[0014] First, the chromatogram of the target substance is obtained through the chromatograph, and the chromatographic peak of the target residue in the sample is extracted. First, the starting point and the end point of the peak need to be determined, and the full width of the peak is measured by the peak shape integration method. The calculation expression is as follows:

[0015]

[0016] , where FWHM is the peak width, t max is the time point when the peak appears in the chromatogram, that is, the time corresponding to the maximum intensity of the chromatographic peak, t min It is the time point from the beginning of the chromatographic peak to the time when it reaches half of its maximum height;

[0017] Considering the complexity and interference of the samples, the chromatographic peak shape contrast factor is introduced to correct the chromatographic peak. The chromatographic peak shape contrast factor is used to compensate for the peak shape changes caused by the complexity of the sample matrix. The calculation expression is as follows:

[0018]

[0019] , where PSF is the peak contrast factor, I(t) is the signal intensity of the chromatographic peak at time point t, and I 0 is the baseline intensity, i.e., the background signal in the absence of target substances, I max is the maximum signal intensity of the chromatographic peak, t start and t end are the start and end times of the chromatographic peak, respectively;

[0020] Finally, by combining the peak width FWHM and the peak shape contrast factor PSF, the reference value of the residual chromatographic peak width is generated, and the generation formula is as follows:

[0021]

[0022] , where CPWR is the reference value of the residue chromatographic peak width, α and β are adjustment coefficients, α is used to calibrate the chromatographic peak width, and β is used to adjust the influence of the peak shape factor.

[0023] Preferably, the specific steps of analyzing the ratios of different isomers in the detection window to generate isomer ratio reference values ​​are as follows:

[0024] In the detection window, the signal peaks of each isomer in the sample are first obtained by chromatographic analysis technology, and the peak area of ​​each isomer is extracted from the original chromatographic data. The peak area represents the abundance of each isomer. All peak areas A are extracted. i ={A 1 , A 2 ,……,A n}, where A i represents the peak area of ​​the ith isomer, n is the total number of isomers, and the concentration data of the isomer is proportional to its peak area;

[0025] According to the concentration data of different isomers extracted, the ratio between each pair of isomers is calculated. The calculation expression is as follows:

[0026]

[0027] , where R ij is the ratio between isomer i and isomer j, A i and A j are the peak areas of isomers i and j, respectively, C i and C jare the concentration data of isomers i and j, respectively. γ and δ are adjustment parameters. γ is used to control the concentration difference (C i -C j ) is used to determine the degree of influence of the contrast value, and δ is used to balance the influence of the concentration difference on the contrast value;

[0028] Calculate the ratio R between each pair of isomers ij After that, all the ratios are weighted to generate the isomer ratio reference value. The generation formula is as follows:

[0029]

[0030] , where IR is the isomer ratio reference value, w ij is the weighting factor, |ΔC ij | is the difference in isomer concentrations, C max is the normalized value for the maximum concentration among all isomers.

[0031] Preferably, the analyzed residue chromatographic peak width reference value and isomer ratio reference value are input into a pre-learned machine learning model, and a veterinary drug residue assessment coefficient is generated through machine learning. The veterinary drug residue assessment coefficient is used to intelligently assess the veterinary drug residue concentration in the test sample.

[0032] Preferably, the veterinary drug residue assessment coefficient generated when evaluating the veterinary drug residue concentration in the current sample through the pre-learned machine learning model is compared and analyzed with the pre-set veterinary drug residue assessment coefficient reference threshold, and an intelligent judgment is made as to whether it belongs to low concentration residue. The specific steps are as follows:

[0033] If the veterinary drug residue assessment coefficient is greater than the pre-set veterinary drug residue assessment coefficient reference threshold, the current test sample is classified as having low-concentration veterinary drug residues; if the veterinary drug residue assessment coefficient is less than or equal to the pre-set veterinary drug residue assessment coefficient reference threshold, the current test sample is classified as having no low-concentration veterinary drug residues.

[0034] Preferably, when the machine learning model evaluates that the current sample contains low concentrations of veterinary drug residues, the specific steps of dynamically improving the detection sensitivity, thereby improving the accuracy and reliability of the detection are as follows:

[0035] When it is determined that there are low-concentration veterinary drug residues in the sample, the detection sensitivity will be dynamically improved based on the preset detection sensitivity. The new sensitivity will be calculated based on the value of the veterinary drug residue assessment coefficient VDRA. The calculation expression is as follows:

[0036]

[0037] , where S base is the preset detection sensitivity, VDRA refis the reference threshold of the veterinary drug residue assessment coefficient, ω is the basic adjustment coefficient for sensitivity improvement, which controls the adjustment range of sensitivity. is the exponential adjustment factor, θ is the environmental adjustment coefficient, C complex is the complexity coefficient of the current sample, C base is the complexity coefficient of the standard sample;

[0038] After the sensitivity adjustment is completed, another test will be conducted based on the new sensitivity setting, and the original test results will be compared with the test results under the new sensitivity setting. In order to verify the effectiveness of the sensitivity adjustment, the detection accuracy improvement coefficient is used to quantify the improvement of the test results after the adjustment. The calculation expression is as follows:

[0039]

[0040] , where A new is the adjusted detection accuracy, A base is the original detection accuracy, ∈ is the adjustment coefficient of sensitivity on accuracy improvement, and κ is the influence coefficient of the ratio of veterinary drug residue assessment coefficient to reference threshold on accuracy.

[0041] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0042] The present invention realizes dynamic sensitivity adjustment for low-concentration veterinary drug residues by combining preliminary detection with preset sensitivity and intelligent evaluation of machine learning models. This method can flexibly adjust the detection sensitivity according to the actual situation of the sample, ensuring that it is always in the best working state in different detection scenarios. By acquiring sample data in real time and performing feature extraction and quantification, combined with the intelligent evaluation of machine learning models, low-concentration residues can be effectively identified and accurately quantified, avoiding the risk of missed detection by traditional fixed sensitivity methods. Ultimately, the system can more reliably detect trace veterinary drug residues in complex sample matrices, thereby improving the accuracy, flexibility and reliability of detection, and ensuring food safety and consumer health. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0044] Figure 1 The present invention is a method flow chart of a veterinary drug residue detection method. DETAILED DESCRIPTION

[0045] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0046] The present invention provides Figure 1 A method for detecting veterinary drug residues shown comprises the following steps:

[0047] First, based on the preset sensitivity, the test sample is tested for veterinary drug residues to obtain the residual information of veterinary drugs in the sample in real time;

[0048] The main purpose of this step is to ensure that all samples receive basic testing, regardless of the residue concentration. At the preset sensitivity, the system quickly screens the samples and obtains real-time information on the residues of veterinary drugs in the samples. This process ensures that preliminary data collection can be completed regardless of the sample concentration, providing raw information for subsequent steps. In addition, the preset sensitivity provides a benchmark for subsequent dynamic adjustments, allowing the system to decide whether to adjust the sensitivity based on subsequent evaluations.

[0049] The test samples can come from a variety of animals and their products, including but not limited to meat (such as pork, beef, chicken, etc.), dairy products (such as milk, goat milk), eggs (such as chicken eggs, duck eggs) and aquatic products. These samples usually come from farms, slaughterhouses or food processing plants. In order to detect veterinary drug residues, a variety of analytical techniques can be used, the most common ones include liquid chromatography-mass spectrometry (LC-MS / MS), gas chromatography-mass spectrometry (GC-MS / MS), immunoassay technology (such as enzyme-linked immunosorbent assay ELISA), high performance liquid chromatography (HPLC), etc. These technologies are highly sensitive and highly selective, and can effectively detect veterinary drug components in animals.

[0050] The real-time acquisition of veterinary drug residue information is used to establish an analysis set, and the data information in the analysis set is preprocessed to make the residue data more accurate and ensure the reliability of subsequent feature extraction and analysis;

[0051] The purpose of this step is to organize all detected veterinary drug residue data into a standardized data set, which provides a basis for further data processing and analysis. Through this analysis set, the system can uniformly manage the residue data of all samples, and mark the key features related to low-concentration residues in the data. This lays the foundation for subsequent feature extraction and dynamic sensitivity adjustment, making data processing more efficient and accurate.

[0052] The goal of preprocessing is to clean and standardize the test data, remove noise and interference in the sample, and ensure data quality. This process may involve operations such as removing background impurities, correcting instrument errors, and filling missing data. In veterinary drug residue detection, samples may contain multiple interfering components, such as moisture, fat, protein, etc., which may affect the final residue detection results. Through preprocessing, this interference can be minimized, making the residue data more accurate and ensuring the reliability of subsequent feature extraction and analysis.

[0053] Extract key features reflecting low-concentration veterinary drug residues from the pre-processed data, analyze and process the extracted key features under the detection window, convert low-concentration veterinary drug residues into measurable digital signals, and quantify the concentration of veterinary drug residues;

[0054] Key features reflecting low-concentration veterinary drug residues are extracted from the preprocessed data, and the extracted features include the chromatographic peak width of the residue and the ratio of different isomers. Under the detection window, the obtained chromatographic peak width of the residue and the ratio of different isomers are analyzed to generate reference values ​​for the chromatographic peak width of the residue and reference values ​​for the isomer ratio, respectively. The concentration levels of veterinary drug residues are quantified by the reference values ​​for the chromatographic peak width of the residue and the isomer ratio, thereby providing data support for the accurate assessment of low-concentration veterinary drug residues.

[0055] The chromatographic peak width of the residue becomes wider as the concentration decreases, which can usually be used as an important indicator of low-concentration veterinary drug residues. In chromatographic analysis, the lower the concentration of the target substance in the sample, the wider the chromatographic peak and the longer the tail. This is because the separation efficiency of low-concentration substances on the chromatographic column is low, and the interaction between the molecules and the mobile phase is weak, resulting in a longer sample transfer time and a slower separation process. This phenomenon is common in low-concentration samples, especially in complex sample matrices, where the interference between the target substance and other components is more significant, resulting in an increase in the chromatographic peak width. In addition, the increase in chromatographic peak width may also be related to the coexistence of metabolites or multi-component residual substances of veterinary drug residues in the sample. The presence of these metabolites or residual substances may lead to the expansion of the peak shape, thereby affecting the final test results. Therefore, the increase in chromatographic peak width can reflect the presence of low-concentration veterinary drug residues in the sample, especially when the veterinary drug residues in the test sample are close to the minimum sensitivity of the detection method. Monitoring through peak width changes can effectively identify and quantify low-concentration residues and ensure the accuracy and reliability of the test results.

[0056] The specific steps for analyzing the chromatographic peak width of the residue under the detection window to generate the reference value of the chromatographic peak width of the residue are as follows:

[0057] First, the chromatogram of the target substance is obtained through the chromatograph, and the chromatographic peak of the target residue in the sample is extracted. First, the starting point (the intersection of the baseline and the peak) and the end point (another intersection of the peak and the baseline) of the peak need to be determined, and the full width of the peak is measured by the peak shape integration method. The calculation expression is as follows:

[0058]

[0059] , where FWHM is the peak width, which represents the width of the chromatographic peak from one side to the other, specifically the horizontal width at half the maximum height of the peak, t max It is the time point when the peak appears in the chromatogram, that is, the time corresponding to the maximum intensity (concentration) of the chromatographic peak, t min It is the time point from the beginning of the chromatographic peak from the baseline to the time when it reaches half of its maximum height. This point marks the starting position of the peak width;

[0060] The above steps are used to measure the width of the peak, which reflects the concentration of veterinary drug residues in the sample. A wider width means that low concentrations of residues exist; a smaller width indicates a higher concentration or better peak separation efficiency.

[0061] Considering the complexity and interference of the sample, the chromatographic peak shape contrast factor (PeakShapeFactor) is introduced to correct the chromatographic peak. The chromatographic peak shape contrast factor is used to compensate for the peak shape changes caused by the complexity of the sample matrix. Especially in the case of low concentration, the peak shape distortion is more obvious. The calculation expression is as follows:

[0062]

[0063] , where PSF is the peak contrast factor, which is used to quantify the morphological characteristics of the chromatographic peak, especially the relationship between the peak width and the peak area, I(t) is the signal intensity of the chromatographic peak at time point t, and I 0 is the baseline intensity, i.e., the background signal in the absence of target substances, I max is the maximum signal intensity of the chromatographic peak, representing the intensity value at the maximum concentration of the target substance, t start and t end are the start and end times of the chromatographic peak, respectively;

[0064] This step can more accurately reflect the actual situation of low-concentration residues in the sample by calculating the ratio of the integrated area of ​​the peak shape to the peak width WHM. As the low-concentration residue increases, the change in peak shape will intensify, so this factor helps to determine whether there are low-concentration veterinary drug residues.

[0065] Finally, by combining the peak width FWHM and the peak shape contrast factor PSF, the reference value of the residual chromatographic peak width is generated, and the generation formula is as follows:

[0066]

[0067] , where CPWR is the reference value of the residual chromatographic peak width, α and β are adjustment coefficients, α is used to calibrate the chromatographic peak width and adjust the peak width measurement to adapt it to the particularity of the experimental equipment, analytical method or sample, and β is used to adjust the influence of the peak shape factor, which is to enhance or weaken the influence of PSF on the reference value of the chromatographic peak width. It is usually adjusted according to the characteristics of the sample to ensure that under different experimental conditions, the influence of the peak shape factor on the width reference value can be appropriately amplified or reduced.

[0068] Through the above steps, the chromatographic peak width reference value not only considers the width of the peak, but also integrates the changes in peak shape, thereby more accurately reflecting the situation of low-concentration veterinary drug residues. When the peak width increases and the peak shape factor is high, the value of the chromatographic peak width reference value will also increase, indicating that a lower concentration of veterinary drug residues may exist in the sample. This reference value provides a basis for dynamically adjusting the detection sensitivity, which helps to improve the detection sensitivity of low-concentration veterinary drug residues.

[0069] The larger the reference value of the residue chromatographic peak width generated after analyzing the chromatographic peak width of the residue under the detection window, it usually means that there is a low concentration of veterinary drug residues in the test sample. Because at low concentrations, the separation effect of veterinary drug molecules in the chromatographic column is poor, resulting in an increase in the width of the chromatographic peak. When the residue concentration is low, the distribution of the target substance on the column and the interaction with the mobile phase are weakened, the separation process slows down, and a wider chromatographic peak is produced. On the contrary, when the concentration of veterinary drugs is high, the chromatographic peak is sharper and more concentrated, and the width is smaller, indicating that the separation is more efficient. Therefore, the larger the numerical value of the reference value of the residue chromatographic peak width, the lower the concentration of veterinary drug residues in the sample, and vice versa, it means that the residue concentration is high or there is no residue.

[0070] Changes in the abundance ratios of different isomers may indicate the presence of low-concentration veterinary drug residues in the test sample, especially after the veterinary drug is metabolized or degraded. The changes in the ratios of isomers can reflect the metabolic pathways and residual characteristics of the veterinary drug. Many veterinary drugs are converted into different isomers or metabolites during metabolism in animals, and these isomers usually have stable abundance ratios when the concentrations are high. However, when the veterinary drug residues are at low concentrations, the abundance ratios of the isomers may change slightly, and such changes are often not easy to detect by conventional detection methods, but through fine abundance ratio analysis, the trace characteristics of the residues can be revealed. For example, if the abundance ratio of a certain isomer changes significantly at low concentrations compared with the standard concentration, it may indicate that the veterinary drug still remains in the sample and the concentration is low. This is because at low concentrations, the residual veterinary drug may exist in different metabolites or incompletely metabolized original forms, and these changes reflect the generation and distribution patterns of different isomers during metabolism. Therefore, changes in isomer ratios provide important clues for low-concentration residues, especially when the metabolic pathways of veterinary drugs are complex and the residue concentrations are extremely low. This method can effectively identify and quantify these trace residues, thereby improving the sensitivity and accuracy of detection.

[0071] The specific steps for analyzing the ratios of different isomers in the detection window to generate isomer ratio reference values ​​are as follows:

[0072] Under the detection window, the signal peaks of each isomer in the sample are first obtained by chromatographic analysis technology (such as liquid chromatography or gas chromatography). Each isomer will correspond to a specific peak in the chromatogram, reflecting the abundance of the isomer. The peak area of ​​each isomer is extracted from the original chromatographic data. The peak area represents the abundance of each isomer. All peak areas A are extracted. i ={A 1 , A 2 ,……,A n}, where A i represents the peak area of ​​the ith isomer, n is the total number of isomers, and the concentration data of the isomer is proportional to its peak area;

[0073] According to the concentration data of different isomers extracted, the ratio between each pair of isomers is calculated. The calculation expression is as follows:

[0074]

[0075] , where R ij is the ratio between isomer i and isomer j, A i and A j are the peak areas of isomers i and j, respectively, C i and C jare the concentration data of isomers i and j, respectively. γ and δ are adjustment parameters. γ is used to control the concentration difference (C i -C j ) has an impact on the contrast value. When C i and C j When the concentration difference is large, it determines the impact of the difference contrast value. δ is used to balance the impact of the concentration difference contrast value. Its function is to prevent the ratio from being distorted when the concentration difference is too large.

[0076] The exponential part of the formula This means that at low concentrations of residues, the concentration difference between isomers will have a greater impact on the ratio, thereby more sensitively reflecting low concentrations of veterinary drug residues. This formula improves the detection capability of trace residues by enhancing the sensitivity of isomer ratios at low concentrations.

[0077] Calculate the ratio R between each pair of isomers ij After that, all the ratios are weighted to generate the isomer ratio reference value. The generation formula is as follows:

[0078]

[0079] , where IR is the isomer ratio reference value, w ij is the weight factor used to adjust the isomer ratio R ij Contribution to the final isomer ratio reference value IR, |ΔC ij | is the difference in isomer concentrations, which is the difference in concentration between isomer i and isomer j, indicating the absolute concentration difference between them, C max is the normalized value for the maximum concentration among all isomers.

[0080] In this way, the differences in isomer concentrations, changes in ratios and their relative importance in the overall detection are taken into account, and an isomer ratio reference value IR is ultimately generated, which can accurately reflect the situation of low-concentration veterinary drug residues in samples.

[0081] The larger the isomer ratio reference value generated after analyzing the ratio of different isomers under the detection window, the more likely it is that low-concentration veterinary drug residues may be present in the test sample, whereas a smaller isomer ratio reference value may indicate that low-concentration residues are not present in the sample. The reason is that veterinary drugs are converted into different isomers during metabolism in animals, and in the case of low-concentration residues, the abundance ratios of these isomers change. The isomer ratio reference value reflects the concentration level of the residue by quantifying these changes. When the concentration of veterinary drug residues is low, the abundance ratios of isomers tend to deviate from normal levels and appear as larger reference values ​​due to incomplete conversion or trace residues during metabolism. This change is usually small, but the presence of low-concentration residues can be effectively revealed by accurately calculating the isomer ratio reference value. Conversely, it may indicate that low-concentration residues are not present in the sample.

[0082] The key features after feature analysis and quantification are input into the pre-trained machine learning model for further intelligent evaluation. The machine learning model is used to evaluate the concentration of veterinary drug residues in the current sample and intelligently determine whether it is a low-concentration residue.

[0083] The analyzed residue chromatographic peak width reference value and isomer ratio reference value are input into the pre-learned machine learning model, and the veterinary drug residue assessment coefficient is generated through machine learning. The veterinary drug residue assessment coefficient is used to perform an intelligent assessment of the veterinary drug residue concentration in the test sample.

[0084] A pre-learned machine learning model refers to a mathematical model that can identify and predict the concentration of veterinary drug residues by training the machine learning algorithm with a large amount of known data in the initial stage. The model learns from a large amount of sample data and masters the relationship between different veterinary drug residue concentrations and their related characteristics (such as chromatographic peak width and isomer ratio). The training data set usually includes samples of various veterinary drug residues, covering different concentration ranges, and these sample data have been annotated with the actual residue concentration. The machine learning algorithm will learn the inherent laws between concentration and characteristics by analyzing the characteristic data of these samples (such as chromatograms, chemical composition, signal intensity, etc.). This learning process usually involves deep learning, support vector machine (SVM), random forest or other supervised learning algorithms. Through training, the machine learning model can extract useful patterns and information from the input sample features, and finally generate an evaluation model that can accurately predict the concentration of veterinary drug residues.

[0085] In practical applications, the pre-learned machine learning model can input the sample's residue chromatographic peak width reference value and isomer ratio reference value into the model in real time. The model will analyze and predict based on the patterns learned during its training process to generate a veterinary drug residue assessment coefficient. This assessment coefficient reflects the concentration level of veterinary drug residues in the current sample. Through continuous "learning" and "adjustment", the model can accurately assess the residual concentration of veterinary drugs in the sample based on the input feature data, avoiding the problems of insufficient sensitivity or missed detection that may exist in a single detection technology. In addition, the advantage of the machine learning model lies in its adaptability and intelligence. By continuously accumulating new data, updating and optimizing the model, it can dynamically adjust the analysis strategy, so as to flexibly respond to different detection scenarios, improve detection accuracy, and ensure timely detection of trace residues.

[0086] The machine learning model is not limited here, and any machine learning model that can perform a comprehensive analysis of the residue chromatographic peak width reference value CPWR and the isomer ratio reference value IR to generate the veterinary drug residue assessment coefficient VDRA is acceptable. To realize the technical solution of the present invention, the present invention provides a specific implementation method;

[0087] The formula for generating the veterinary drug residue assessment coefficient VDRA is as follows:

[0088]

[0089] , where f 1 、f 2 are the preset proportional coefficients of the residue chromatographic peak width reference value CPWR and the isomer ratio reference value IR, respectively, and f 1 、f 2 Both are greater than 0.

[0090] Preset proportionality factor f 1 and f 2 They are used to adjust the influence of the residue chromatographic peak width reference value ·CPWR and the isomer ratio reference value IR on the calculation of the veterinary drug residue assessment coefficient VDRA. Specifically, f 1 and f 2 are correlation adjustment factors with chromatographic peak width and isomer ratio, which determine the importance of these two reference values ​​in assessing the concentration of veterinary drug residues. These coefficients ensure that the model accurately reflects the concentration of veterinary drug residues in different samples and situations by weighing the contribution of chromatographic peak width and isomer ratio to the model evaluation results. The preset proportionality coefficients are usually obtained through experience or historical data tuning, and they are optimized during the model training process to improve the accuracy and stability of predictions.

[0091] It can be seen from the veterinary drug residue assessment coefficient that the larger the residue chromatographic peak width reference value generated after analyzing the chromatographic peak width of the residue under the detection window, and the larger the isomer ratio reference value generated after analyzing the ratio of different isomers under the detection window, the larger the veterinary drug residue assessment coefficient generated when evaluating the veterinary drug residue concentration in the current sample by using the pre-learned machine learning model, the greater the presence of low-concentration veterinary drug residues in the current test sample, otherwise it indicates that no low-concentration veterinary drug residues appear in the current test sample.

[0092] The veterinary drug residue assessment coefficient generated when evaluating the veterinary drug residue concentration in the current sample through the pre-learned machine learning model is compared and analyzed with the pre-set veterinary drug residue assessment coefficient reference threshold, and an intelligent judgment is made as to whether it is a low-concentration residue. The specific steps are as follows:

[0093] If the veterinary drug residue assessment coefficient is greater than the pre-set veterinary drug residue assessment coefficient reference threshold, the current test sample is classified as having low-concentration veterinary drug residues; if the veterinary drug residue assessment coefficient is less than or equal to the pre-set veterinary drug residue assessment coefficient reference threshold, the current test sample is classified as having no low-concentration veterinary drug residues.

[0094] When the machine learning model evaluates the presence of low-concentration veterinary drug residues in the current sample, the detection sensitivity will be dynamically improved based on the evaluation results to ensure that the detection system is always in the best working state in different detection scenarios and adapt to the complexity of different samples, thereby improving the accuracy and reliability of the detection;

[0095] When the machine learning model assesses the presence of low-concentration veterinary drug residues in the current sample, the detection sensitivity will be dynamically improved, thereby improving the accuracy and reliability of the detection. The specific steps are as follows:

[0096] When it is determined that there are low-concentration veterinary drug residues in the sample, the detection sensitivity will be dynamically improved based on the preset detection sensitivity. The new sensitivity will be calculated based on the value of the veterinary drug residue assessment coefficient VDRA. The calculation expression is as follows:

[0097]

[0098] , where S base is the preset detection sensitivity, VDRA ref is the reference threshold of the veterinary drug residue assessment coefficient, ω is the basic adjustment coefficient for sensitivity improvement, which controls the adjustment range of sensitivity. is the exponential adjustment factor, which is used for weighted adjustment to amplify the impact of low-concentration residues and ensure that the sensitivity of low-concentration samples is effectively improved. θ is the environmental adjustment coefficient, which is used to fine-tune the sensitivity according to different detection environments (such as temperature, humidity, noise, etc.). complexis the complexity coefficient of the current sample, which measures the influence of complex factors such as interfering substances and matrix effects in the sample. base is the complexity coefficient of the standard sample, used as a reference benchmark;

[0099] This step is based on the veterinary drug residue assessment coefficient VDRA evaluated by the machine learning model and the preset veterinary drug residue assessment coefficient reference threshold VDRA ref , and intelligently adjust the detection sensitivity. After evaluating the concentration of veterinary drug residues in the current sample, if the concentration is higher than the preset threshold, the system will automatically increase the detection sensitivity according to the set sensitivity adjustment mechanism to ensure that low-concentration residues can be captured. If the concentration is lower than the threshold, the sensitivity is maintained or reduced to avoid unnecessary waste of resources. This dynamic adjustment can ensure that the detection system maintains the best working state in a complex sample environment, adapts to the characteristics of different samples, and improves the accuracy and efficiency of detection.

[0100] After the sensitivity adjustment is completed, another test will be conducted based on the new sensitivity setting, and the original test results will be compared with the test results under the new sensitivity setting. In order to verify the effectiveness of the sensitivity adjustment, the detection accuracy improvement coefficient is used to quantify the improvement of the test results after the adjustment. The calculation expression is as follows:

[0101]

[0102] , where A new is the adjusted detection accuracy, A base is the original detection accuracy, ∈ is the adjustment coefficient of sensitivity on accuracy improvement, which controls the impact of sensitivity change on detection accuracy, and κ is the influence coefficient of the ratio of veterinary drug residue assessment coefficient to reference threshold on accuracy, which is used to adjust the veterinary drug residue assessment coefficient VDRA and the reference threshold VDRA ref The difference between them affects the accuracy.

[0103] By automatically adjusting the detection sensitivity, the detection system can always operate at its best under different concentration levels and sample complexity scenarios. Through this intelligent sensitivity adjustment, the system can adapt to various sample types, whether it is low-concentration veterinary drug residues or complex sample matrices, thereby minimizing the risk of missed detection and improving the accuracy and reliability of detection.

[0104] Through the intelligent evaluation of the concentration of veterinary drug residues in the sample by the machine learning model, the detection sensitivity is dynamically adjusted to ensure that the detection system can always operate in the best state in different detection scenarios, adapt to the complexity of various samples, and maximize the accuracy and reliability of the detection. Specifically, traditional veterinary drug residue detection methods usually use fixed detection sensitivity, which may lead to missed detection of low-concentration residues in some cases, especially for veterinary drug samples with complex distribution and metabolism. By evaluating the concentration of veterinary drugs in samples through machine learning models, the actual residue level of each sample can be understood in real time, and the detection sensitivity can be automatically adjusted based on this evaluation result. When the model evaluates that the veterinary drug residues in the sample are at a low concentration, the system will automatically increase the sensitivity and enhance the detection ability of trace residues, thereby avoiding missed detection. For samples with higher concentrations, the system can be maintained at a lower sensitivity setting to avoid excessive detection and waste of resources.

[0105] This dynamic adjustment can not only improve the accuracy of detection, but also effectively deal with complex sample matrices or metabolic patterns of different veterinary drugs. For example, in some mixed samples containing multiple veterinary drug residues, traditional methods may be interfered with and unable to accurately distinguish low-concentration residues, but by dynamically adjusting the sensitivity, the system can more accurately identify the residue of each veterinary drug and avoid misjudgment or missed judgment. At the same time, adaptive sensitivity adjustment can also improve the versatility of the system, ensuring that the system can maintain an efficient and stable working state in any detection scenario, thereby providing reliable data support for food safety assurance.

[0106] The present invention realizes dynamic sensitivity adjustment for low-concentration veterinary drug residues by combining preliminary detection with preset sensitivity and intelligent evaluation of machine learning models. This method can flexibly adjust the detection sensitivity according to the actual situation of the sample, ensuring that it is always in the best working state in different detection scenarios. By acquiring sample data in real time and performing feature extraction and quantification, combined with the intelligent evaluation of machine learning models, low-concentration residues can be effectively identified and accurately quantified, avoiding the risk of missed detection by traditional fixed sensitivity methods. Ultimately, the system can more reliably detect trace veterinary drug residues in complex sample matrices, thereby improving the accuracy, flexibility and reliability of detection, and ensuring food safety and consumer health.

[0107] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0108] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0109] It should be noted that, in this article, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0110] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0111] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0112] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0113] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0114] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0115] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0116] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for detecting veterinary drug residues, characterized in that: The following steps are involved: First, based on the preset sensitivity, the test sample is tested for veterinary drug residues to obtain the residual information of veterinary drugs in the sample in real time; The real-time acquisition of veterinary drug residue information is used to establish an analysis set, and the data information in the analysis set is preprocessed to make the residue data more accurate and ensure the reliability of subsequent feature extraction and analysis; Extract key features reflecting low-concentration veterinary drug residues from the pre-processed data, analyze and process the extracted key features under the detection window, convert low-concentration veterinary drug residues into measurable digital signals, and quantify the concentration of veterinary drug residues; The key features after feature analysis and quantification are input into the pre-trained machine learning model for further intelligent evaluation. The machine learning model is used to evaluate the concentration of veterinary drug residues in the current sample and intelligently determine whether it is a low-concentration residue. When the machine learning model assesses the presence of low-concentration veterinary drug residues in the current sample, the detection sensitivity will be dynamically improved based on the evaluation results to ensure that the detection system is always in the best working state in different detection scenarios and adapts to the complexity of different samples, thereby improving the accuracy and reliability of the detection.

2. A veterinary drug residue detection method according to claim 1, characterized in that: Key features reflecting low-concentration veterinary drug residues are extracted from the preprocessed data, and the extracted features include the chromatographic peak width of the residue and the ratio of different isomers. Under the detection window, the obtained chromatographic peak width of the residue and the ratio of different isomers are analyzed to generate reference values ​​for the chromatographic peak width of the residue and reference values ​​for the isomer ratio, respectively. The concentration levels of veterinary drug residues are quantified by the reference values ​​for the chromatographic peak width of the residue and the isomer ratio, thereby providing data support for the accurate assessment of low-concentration veterinary drug residues.

3. A veterinary drug residue detection method according to claim 2, characterized in that: The specific steps for analyzing the chromatographic peak width of the residue under the detection window to generate the reference value of the chromatographic peak width of the residue are as follows: First, the chromatogram of the target substance is obtained through the chromatograph, and the chromatographic peak of the target residue in the sample is extracted. First, the starting point and the end point of the peak need to be determined, and the full width of the peak is measured by the peak shape integration method. The calculation expression is as follows: , where FWHM is the peak width, t max is the time point when the peak appears in the chromatogram, that is, the time corresponding to the maximum intensity of the chromatographic peak, t min It is the time point from the beginning of the chromatographic peak to the time when it reaches half of its maximum height; Considering the complexity and interference of the samples, the chromatographic peak shape contrast factor is introduced to correct the chromatographic peak. The chromatographic peak shape contrast factor is used to compensate for the peak shape changes caused by the complexity of the sample matrix. The calculation expression is as follows: , where PSF is the peak contrast factor, I(t) is the signal intensity of the chromatographic peak at time point t, I0 is the baseline intensity, i.e., the background signal when there is no target substance, and I max is the maximum signal intensity of the chromatographic peak, t start and t end are the start and end times of the chromatographic peak, respectively; Finally, by combining the peak width FWHM and the peak shape contrast factor PSF, the reference value of the residual chromatographic peak width is generated, and the generation formula is as follows: , where CPWR is the reference value of the residue chromatographic peak width, α and β are adjustment coefficients, α is used to calibrate the chromatographic peak width, and β is used to adjust the influence of the peak shape factor.

4. A veterinary drug residue detection method according to claim 2, characterized in that: The specific steps for analyzing the ratios of different isomers in the detection window to generate isomer ratio reference values ​​are as follows: In the detection window, the signal peaks of each isomer in the sample are first obtained by chromatographic analysis technology, and the peak area of ​​each isomer is extracted from the original chromatographic data. The peak area represents the abundance of each isomer. All peak areas A are extracted. i ={A1, A2, ..., A n }, where A i represents the peak area of ​​the ith isomer, n is the total number of isomers, and the concentration data of the isomer is proportional to its peak area; According to the concentration data of different isomers extracted, the ratio between each pair of isomers is calculated. The calculation expression is as follows: , where R ij is the ratio between isomer i and isomer j, A i and A j are the peak areas of isomers i and j, respectively, C i and C j are the concentration data of isomers i and j, respectively. γ and δ are adjustment parameters. γ is used to control the concentration difference (C i -C j ) is used to determine the degree of influence of the contrast value, and δ is used to balance the influence of the concentration difference on the contrast value; Calculate the ratio R between each pair of isomers ij After that, all the ratios are weighted to generate the isomer ratio reference value. The generation formula is as follows: , where IR is the isomer ratio reference value, w ij is the weighting factor, |ΔC ij | is the difference in isomer concentrations, C max is the normalized value for the maximum concentration among all isomers.

5. A veterinary drug residue detection method according to claim 2, characterized in that: The analyzed residue chromatographic peak width reference value and isomer ratio reference value are input into the pre-learned machine learning model, and the veterinary drug residue assessment coefficient is generated through machine learning. The veterinary drug residue assessment coefficient is used to perform an intelligent assessment of the veterinary drug residue concentration in the test sample.

6. A veterinary drug residue detection method according to claim 5, characterized in that: The veterinary drug residue assessment coefficient generated when evaluating the veterinary drug residue concentration in the current sample through the pre-learned machine learning model is compared and analyzed with the pre-set veterinary drug residue assessment coefficient reference threshold, and an intelligent judgment is made as to whether it is a low-concentration residue. The specific steps are as follows: If the veterinary drug residue assessment coefficient is greater than the pre-set veterinary drug residue assessment coefficient reference threshold, the current test sample is classified as having low-concentration veterinary drug residues; if the veterinary drug residue assessment coefficient is less than or equal to the pre-set veterinary drug residue assessment coefficient reference threshold, the current test sample is classified as having no low-concentration veterinary drug residues.

7. A veterinary drug residue detection method according to claim 6, characterized in that: When the machine learning model assesses the presence of low-concentration veterinary drug residues in the current sample, the detection sensitivity will be dynamically improved, thereby improving the accuracy and reliability of the detection. The specific steps are as follows: When it is determined that there are low-concentration veterinary drug residues in the sample, the detection sensitivity will be dynamically improved based on the preset detection sensitivity. The new sensitivity will be calculated based on the value of the veterinary drug residue assessment coefficient VDRA. The calculation expression is as follows: , where S base is the preset detection sensitivity, VDRA ref is the reference threshold of the veterinary drug residue assessment coefficient, ω is the basic adjustment coefficient for sensitivity improvement, which controls the adjustment range of sensitivity. is the exponential adjustment factor, θ is the environmental adjustment coefficient, C complex is the complexity coefficient of the current sample, C base is the complexity coefficient of the standard sample; After the sensitivity adjustment is completed, another test will be conducted based on the new sensitivity setting, and the original test results will be compared with the test results under the new sensitivity setting. In order to verify the effectiveness of the sensitivity adjustment, the detection accuracy improvement coefficient is used to quantify the improvement of the test results after the adjustment. The calculation expression is as follows: , where A new is the adjusted detection accuracy, A base is the original detection accuracy, ∈ is the adjustment coefficient of sensitivity on accuracy improvement, and κ is the influence coefficient of the ratio of veterinary drug residue assessment coefficient to reference threshold on accuracy.

Images