Method and system for detecting foodborne microorganisms based on spectroscopy

Abstract

The application relates to a foodborne microorganism detection method and system based on spectral technology. The method comprises the following steps: performing correlation analysis on the content sequence of each chemical substance in a plurality of blank matrix samples and the absorbance sequence of each wave band in the spectral data of the plurality of blank matrix samples to obtain an aliasing coefficient of each wave band; performing analysis on the content of each strain in a plurality of strain samples and the absorbance of each wave band in the corresponding strain spectral data to obtain an absorbance influence coefficient between any two strains under each wave band; determining a strain group representative coefficient of each wave band according to the number of the plurality of strains, the aliasing coefficient of each wave band and the absorbance influence coefficient of other strains on the first strain under the corresponding wave band; and screening a characteristic wave band from each wave band according to the strain group representative coefficient of each wave band and the absorbance sequence of the corresponding wave band in the spectral data of a plurality of target samples. The method simplifies the analysis difficulty of the coexistence of multiple strain groups in a complex matrix, and improves the generalization ability and practical application value of the model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of microorganism detection, in particular to a foodborne microorganism detection method and system based on spectral technology. BACKGROUND

[0002] With the rapid development of food industry and globalization of supply chain, foodborne microbial contamination has become an important factor affecting food safety and public health. Pathogenic bacteria such as Escherichia coli, Salmonella, Staphylococcus aureus and Listeria are the main microorganisms causing foodborne diseases, and their contamination risks exist widely in meat, aquatic products, dairy products, fruits and vegetables, and instant foods. Food safety regulatory agencies and related enterprises in various countries urgently need efficient, rapid and non-destructive detection methods to ensure food quality and consumer health. In this context, detection methods based on spectral technology have gradually become an important research direction and application trend in the field of food safety detection due to their advantages of non-destructive, rapid response and no need for complex sample pretreatment.

[0003] In the detection of foodborne microorganisms using spectra, the traditional method needs to extract and analyze the characteristic waveband of the obtained spectral data, and then detect the content of microorganisms according to the size of the content of different substances. However, in actual food detection scenarios, there are often multiple different types of microorganisms existing at the same time. Since similar components such as proteins, lipids and polysaccharides are commonly contained in different microorganism cells, the characteristic wavebands of multiple similar components in spectral response will overlap, making it difficult to accurately distinguish and quantify the content of different microorganisms by relying solely on spectral data, which will result in insufficient detection accuracy. SUMMARY

[0004] In order to solve the above technical problems, the purpose of the present application is to provide a foodborne microorganism detection method and system based on spectral technology, and the technical solution adopted is as follows:

[0005] In a first aspect, a foodborne microorganism detection method based on spectral technology is provided, the method comprising:

[0006] correlating the content sequence of each chemical substance in the plurality of blank matrix samples with the absorbance sequence of each waveband in the spectral data of the plurality of blank matrix samples to obtain an aliasing coefficient of each waveband;

[0007] analyzing the content of each strain in the plurality of strain samples and the absorbance of each waveband in the corresponding strain spectral data to obtain an absorbance influence coefficient between any two strains under each waveband; the strain samples are obtained by culturing a plurality of strains in each blank matrix sample, and the strain spectral data is obtained by analyzing the difference between the spectral data of the plurality of strain samples and the spectral data of the corresponding blank matrix samples;

[0008] determining a bacteria group representative coefficient of each wave band according to the number of the plurality of bacteria species, the overlap coefficient of each wave band, and the absorbance influence coefficient of other bacteria species on the first bacteria species in the corresponding wave band; the first bacteria species is any bacteria species in the plurality of bacteria species; the bacteria group representative coefficient indicates a specificity degree of each wave band to the bacteria species that is most strongly represented in the wave band;

[0009] screening a characteristic wave band from each wave band according to the bacteria group representative coefficient of each wave band and the absorbance sequence of the corresponding wave band in the spectral data of the plurality of target samples; the target samples include a plurality of blank matrix samples and a plurality of bacteria species samples; the spectral data of the target samples in the characteristic wave band is used to train the bacteria content detection model.

[0010] Optionally, the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each wave band in the spectral data of the plurality of blank matrix samples are subjected to correlation analysis to obtain an overlap coefficient of each wave band, including:

[0011] Pearson correlation analysis is performed on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each wave band in the spectral data of the plurality of blank matrix samples to obtain a plurality of Pearson correlation coefficients between each chemical substance and each wave band; each wave band corresponds to an absorbance sequence;

[0012] For any wave band, the overlap coefficient of the wave band is calculated according to the maximum value of the Pearson correlation coefficients between the wave band and all chemical substances and the average value of the Pearson correlation coefficients between the wave band and all chemical substances.

[0013] Optionally, the content of each bacteria species in the plurality of bacteria species samples and the absorbance of each wave band in the corresponding bacteria spectral data are analyzed to obtain an absorbance influence coefficient between any two bacteria species in each wave band, including:

[0014] The content of each bacteria species in the plurality of single bacteria species samples, the content of each bacteria species in the plurality of mixed bacteria species samples, and the absorbance of each wave band in the corresponding bacteria spectral data are analyzed to obtain an absorbance influence coefficient between any two bacteria species in each wave band; wherein the bacteria species samples include single bacteria species samples and mixed bacteria species samples, one bacteria species is cultured in the single bacteria species samples, and at least two bacteria species are cultured in the mixed bacteria species samples.

[0015] Optionally, the content of each bacteria species in the plurality of single bacteria species samples, the content of each bacteria species in the plurality of mixed bacteria species samples, and the absorbance of each wave band in the corresponding bacteria spectral data are analyzed to obtain an absorbance influence coefficient between any two bacteria species in each wave band, including:

[0016] According to the product of the single-strain ratio and the content of the first strain in the plurality of mixed-strain samples, the content of the second strain in the plurality of mixed-strain samples, and the absorbance of each wave band in the corresponding mixed-strain spectrum data, the absorbance influence coefficient of the second strain on the first strain at each wave band is obtained; the single-strain ratio indicates the reciprocal of the ratio of the content of the first strain in the plurality of single-strain samples and the absorbance of each wave band in the corresponding single-strain spectrum data, the first strain and the second strain are any two different strains in the plurality of strains, the single-strain spectrum data is the spectrum data corresponding to the single-strain sample in the strain spectrum data, and the mixed-strain spectrum data is the spectrum data corresponding to the mixed-strain sample in the strain spectrum data.

[0017] Optionally, the determination of the strain group representative coefficient of each wave band according to the number of the plurality of strains, the aliasing coefficient of each wave band, and the absorbance influence coefficient of the other strains on the first strain at the corresponding wave band comprises:

[0018] The spectral representation intensity of the first strain at the corresponding wave band is calculated according to the product of the negative of the aliasing coefficient of each wave band and the absorbance influence coefficient of each second strain on the first strain at the corresponding wave band, the difference between the number of the plurality of strains and 1.

[0019] The strain group representative coefficient of the corresponding wave band is determined according to the spectral representation intensity of the first strain at the corresponding wave band.

[0020] Optionally, the determination of the strain group representative coefficient of the corresponding wave band according to the spectral representation intensity of the first strain at the corresponding wave band comprises:

[0021] The maximum spectral representation intensity is selected from the plurality of spectral representation intensities of the first strain at the corresponding wave band as the strain group representative coefficient of the corresponding wave band.

[0022] Optionally, the selection of the characteristic wave band from each wave band according to the strain group representative coefficient of each wave band and the absorbance sequence of the corresponding wave band in the spectrum data of the plurality of target samples comprises:

[0023] a. The wave band corresponding to the maximum strain group representative coefficient is selected from the plurality of strain group representative coefficients of the plurality of wave bands as the first characteristic wave band in the selected characteristic wave band set.

[0024] b. The gain coefficient of each candidate wave band is calculated according to the strain group representative coefficient of each candidate wave band, the absorbance sequence of the corresponding candidate wave band in the spectrum data of the plurality of target samples, and the absorbance sequence of each characteristic wave band in the selected characteristic wave band set in the spectrum data of the plurality of target samples; the candidate wave band is a wave band other than the characteristic wave band; each candidate wave band corresponds to an absorbance sequence, and each characteristic wave band corresponds to an absorbance sequence.

[0025] c. selecting a candidate waveband with the largest gain coefficient from the plurality of gain coefficients of all candidate wavebands, and adding the candidate waveband to the selected characteristic waveband set as a new characteristic waveband;

[0026] d. repeating steps b and c until the gain coefficient of each candidate waveband is less than the preset threshold.

[0027] Optionally, the gain coefficient of each candidate waveband is calculated according to the colony representative coefficient of each candidate waveband, the absorbance sequence of the corresponding candidate waveband in the spectral data of the plurality of target samples, and the absorbance sequence of each characteristic waveband in the selected characteristic waveband set in the spectral data of the plurality of target samples, and the gain coefficient of each candidate waveband is calculated according to the product of the colony representative coefficient of each candidate waveband and the weight coefficient, and the plurality of Pearson correlation coefficients between the absorbance sequence of the corresponding candidate waveband in the spectral data of the plurality of target samples and the absorbance sequence of each characteristic waveband in the selected characteristic waveband set in the spectral data of the plurality of target samples.

[0028] The gain coefficient of each candidate waveband is calculated according to the product of the colony representative coefficient of each candidate waveband and the weight coefficient, and the plurality of Pearson correlation coefficients between the absorbance sequence of the corresponding candidate waveband in the spectral data of the plurality of target samples and the absorbance sequence of each characteristic waveband in the selected characteristic waveband set in the spectral data of the plurality of target samples.

[0029] In a second aspect, a foodborne microorganism detection system based on spectral technology is provided, and the system comprises:

[0030] A first analysis module is configured to perform correlation analysis on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each waveband in the spectral data of the plurality of blank matrix samples, to obtain an overlap coefficient of each waveband.

[0031] A second analysis module is configured to analyze the content of each strain in the plurality of strain samples and the absorbance of each waveband in the corresponding strain spectral data, to obtain an absorbance influence coefficient between any two strains under each waveband; the strain samples are obtained by culturing a plurality of strains in each blank matrix sample, and the strain spectral data is obtained according to the difference between the spectral data of the plurality of strain samples and the spectral data of the corresponding blank matrix sample.

[0032] A determination module is configured to determine a colony representative coefficient of each waveband according to the number of the plurality of strains, the overlap coefficient of each waveband, and the absorbance influence coefficient of other strains on a first strain under the corresponding waveband; the first strain is any strain in the plurality of strains; the colony representative coefficient indicates the specificity degree of each waveband to the strain that is most strongly represented by the waveband.

[0033] A screening module is configured to screen a characteristic waveband from each waveband according to the colony representative coefficient of each waveband and the absorbance sequence of the corresponding waveband in the spectral data of the plurality of target samples; the target samples include the plurality of blank matrix samples and the plurality of strain samples, and the spectral data of the target samples under the characteristic waveband is used to train a strain content detection model.

[0034] Optionally, the first analysis module is further configured to:

[0035] performing Pearson correlation analysis on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each waveband in the spectral data of the corresponding blank matrix sample, to obtain a plurality of Pearson correlation coefficients between each chemical substance and each waveband;

[0036] for any waveband, calculating an aliasing coefficient of the waveband according to the maximum value in the Pearson correlation coefficients between the waveband and all chemical substances, and the average value of the Pearson correlation coefficients between all wavebands and all chemical substances.

[0037] On the basis of common general knowledge in the art, the above-mentioned preferred conditions can be combined arbitrarily, i.e. to obtain each preferred example of the present application.

[0038] The present application has the following beneficial effects: The present application has a complete process from spectral data acquisition, waveband aliasing degree determination, extraction of bacterial flora characteristic spectrum, selection of characteristic waveband, to modeling of multi-bacterial flora content. By introducing the joint evaluation of the waveband aliasing degree and the bacterial flora representative coefficient in the characteristic waveband selection stage, the misjudgment problem caused by waveband redundancy or overlapping of material signals in the traditional method can be effectively avoided. By deducting the spectral data of the blank matrix to extract the bacterial spectrum data, the corresponding relationship between the spectral characteristics and the target bacterial flora is further enhanced. Finally, the direct estimation of the multi-bacterial flora content is realized by combining multi-output modeling, which improves the accuracy and robustness of the detection. The overall scheme not only simplifies the analysis difficulty of the coexistence of multi-bacterial flora in complex matrix, but also improves the generalization ability and practical application value of the model, and can be widely applied to food safety detection and rapid quality control links. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed in the following embodiment or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0040] Figure 1 Flow chart of a foodborne microorganism detection method based on spectral technology in an embodiment;

[0041] Figure 2 Structural schematic diagram of a foodborne microorganism detection system based on spectral technology in an embodiment;

[0042] Figure 3 Structural schematic diagram of an electronic device in an embodiment;

[0043] The labels in the drawings are respectively: 21, first analysis module; 22, second analysis module; 23, determination module; 24, screening module; 30, electronic device; 31, processor; 32, memory; 321, random access memory; 322, cache memory; 323, read-only memory; 324, program module; 325, program tool; 33, bus; 34, external device; 35, input / output interface; 36, network adapter. DETAILED DESCRIPTION

[0044] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined object, the specific embodiments, structure, characteristics and effects of the foodborne microorganism detection method and system based on spectrum technology according to the present application are described in detail as follows in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0045] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0046] The specific scheme of the foodborne microorganism detection method based on spectrum technology provided by the present application is specifically described below in combination with the drawings. As shown in Figure 1 The method comprises:

[0047] S11, performing correlation analysis on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each waveband in the spectrum data of the plurality of blank matrix samples, to obtain the aliasing coefficient of each waveband.

[0048] The blank matrix sample is a food matrix sample without cultivation of bacterial strains, that is, a sterile or bacteriostatic treated matrix sample.

[0049] Microorganism refers to a small organism observed by microscope, including bacteria, fungi, viruses and algae. Bacterial strain refers to a specific species or strain of bacteria or fungi identified, named and preserved in taxonomy, such as Escherichia coli O157:H7, Salmonella and lactic acid bacteria. The bacterial strain mentioned in the present application is included in microorganism.

[0050] Chemical substances include proteins, lipids, polysaccharides, nucleic acids and metabolic products (organic acids, alcohols, esters) etc. After the bacterial strain is cultured in the matrix, the content of the chemical substance can be obtained by standard chemical quantitative method (gravimetric method, volumetric method, spectroscopy / chromatography method etc.).

[0051] The present application obtains the spectral data of various microorganisms and the spectral data of blank matrix in a test environment:

[0052] ①Identify the object and food matrix:

[0053] Foodborne microorganisms include: Escherichia coli O157:H7, Salmonella, Staphylococcus aureus, Listeria, lactic acid bacteria, etc.

[0054] Food matrix includes: fresh meat or poultry, instant vegetables, dairy products (covering three types of high moisture, high protein and high fiber).

[0055] ②Sample and grouping design:

[0056] Single strain sample: each strain x each matrix x 5 CFU (Colony-Forming Unit, colony-forming unit) gradient (10 4 , 10 5 , 10 6 CFU / g) x 6 technical repeats.

[0057] Mixed strain sample: two or three microorganisms (i.e. strains) are mixed in proportion, for example 1:1, 1:9, 3:7, etc., covering different proportions under the same total CFU, each proportion x 3 technical repeats.

[0058] Negative control: sterile or bacteriostatic treated matrix (blank matrix), blank matrix is used to measure the background of the matrix.

[0059] ③Spectrum collection configuration:

[0060] Mid-infrared: collected by mid-infrared spectrometer, wavelength range is 4000-650cm -1 , resolution is 4cm -1 . The abscissa of the obtained spectral data is wave number, and the ordinate is absorption or absorbance.

[0061] ④Label structure:

[0062] Classification label: sample bacterial population category, mixed strain sample can label multiple strains as positive at the same time.

[0063] Quantitative label: record the absolute content of each microorganism CFU / g (solid / semi-solid), the content of microorganisms in mixed strain sample is saved in vector form .

[0064] After the above spectral data acquisition and sample labeling, a complete data set covering single-species samples and mixed-species samples, and including different substrate types and different CFU gradients, is obtained. The data set contains both the classification information of each microorganism and the corresponding quantitative content, providing a basis for subsequent spectral preprocessing, feature band selection, and modeling analysis.

[0065] Chemical substances from the inherent components of food substrates (protein, lipid, polysaccharide, water, etc.) constitute the chemical background noise in the spectral data, and the cell components of microorganisms themselves (protein, nucleic acid, lipopolysaccharide, metabolite) are themselves a class of biochemical substances. When the chemical substances in the substrate are similar to the functional groups or molecular structures contained in the microorganisms, the two will overlap in the same waveband, causing waveband aliasing. In the process of spectral data analysis, the strength of the signal at different wavebands depends on the content of different chemical substances or functional groups, but when the spectral waveband range influenced by multiple chemical substances or functional groups is similar or overlapping, the correlation between multiple substances and a single waveband will be high, i.e., waveband aliasing will occur.

[0066] In the detection of foodborne microorganisms using spectral technology, the spectral signal not only comes from the target microorganism itself, but also contains a large amount of background substances in the food substrate. Since the growth and metabolism of the strain depend on the nutrients provided by the substrate, the spectral characteristics of the strain are mainly established on the basis of the spectral characteristics of the food substrate. If a waveband in the substrate has a high correlation with multiple chemical substances, the probability of strong aliasing effect in the subsequent microorganism detection is also high, making it difficult to use as a specific waveband for strain differentiation. When the waveband aliasing is serious, it will lead to redundant features in the input of the strain content detection model, resulting in poor model fitting effect, increasing the risk of overfitting and computational complexity. Therefore, it is necessary to first determine the aliasing degree of each waveband.

[0067] Therefore, in an embodiment, the content sequence of each chemical substance in the plurality of blank substrate samples and the absorbance sequence of each waveband in the spectral data of the plurality of blank substrate samples are subjected to correlation analysis to obtain the aliasing coefficient of each waveband, including:

[0068] The content sequence of each chemical substance in the plurality of blank substrate samples and the absorbance sequence of each waveband in the spectral data of the plurality of blank substrate samples are subjected to Pearson correlation analysis to obtain a plurality of Pearson correlation coefficients between each chemical substance and each waveband; each waveband corresponds to an absorbance sequence;

[0069] For any waveband, the aliasing coefficient of the waveband is calculated according to the maximum value of the Pearson correlation coefficients of the waveband and all chemical substances, and the average value of the Pearson correlation coefficients of the waveband and all chemical substances.

[0070] The content sequence of each chemical substance in the plurality of blank matrix samples indicates: for each chemical substance, a content sequence of the chemical substance in the plurality of blank matrix samples is obtained, and each chemical substance corresponds to a content sequence.

[0071] The absorbance sequence of each wave band in the spectral data of the plurality of blank matrix samples indicates: an absorbance sequence of each wave band in the spectral data of the plurality of blank matrix samples is obtained, that is, an absorbance sequence of each wave band of the plurality of blank matrix samples is obtained, and each wave band corresponds to an absorbance sequence.

[0072] The first wave band The calculation formula of the overlap coefficient of the first wave band is as follows:

[0073] ;

[0074] The calculation formula of the overlap coefficient of the first wave band is as follows: The overlap coefficient of the first wave band, indicates a spectral absorbance sequence of the first wave band in the spectral data of the plurality of blank matrix samples, indicates a content sequence of the first chemical substance in the plurality of blank matrix samples, indicates that Pearson correlation analysis is performed on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each wave band in the spectral data of the plurality of blank matrix samples, indicates an average value of the Pearson correlation coefficients of the first wave band and all chemical substances, indicates a maximum value of the Pearson correlation coefficients of the first wave band and all chemical substances, and the denominator is prevented from being 0. When the correlation of a single wave band with a single chemical substance is high, the overlap degree of the corresponding wave band is low, that is, there is a maximum Pearson correlation coefficient, and the average value of the remaining Pearson correlation coefficients is low, and vice versa. When the correlation of a single wave band with a single chemical substance is high, the overlap degree of the corresponding wave band is low, that is, there is a maximum Pearson correlation coefficient, and the average value of the remaining Pearson correlation coefficients is low, and vice versa.

[0075] When the correlation of a single wave band with a single chemical substance is high, the overlap degree of the corresponding wave band is low, that is, there is a maximum Pearson correlation coefficient, and the average value of the remaining Pearson correlation coefficients is low, and vice versa.

[0076] When the correlation of a single wave band with a single chemical substance is high, the overlap degree of the corresponding wave band is low, that is, there is a maximum Pearson correlation coefficient, and the average value of the remaining Pearson correlation coefficients is low, and vice versa. ​​The waveband shows a lower degree of aliasing in the matrix, and the waveband is better for a single chemical substance and can better reflect the change in the content of the chemical substance. However, in microbial detection, although different microbial populations differ in composition, the cell structures and metabolic products of different microbial populations often contain the same functional groups or substances, such as proteins, lipids, or polysaccharides in different microorganisms. These same components will show similar absorption characteristics in the spectrum. Therefore, when detecting microorganisms, it is also necessary to consider whether the different characteristic wavebands are specific to the microorganisms and whether they are good representatives of each microbial population.

[0077] S12, analyzing the content of each microbial population in the plurality of microbial population samples and the absorbance of each waveband in the corresponding microbial population spectrum data to obtain the absorbance influence coefficient between any two microbial populations under each waveband.

[0078] The microbial population samples are obtained by culturing a plurality of microbial populations in each blank matrix sample. The blank matrix sample in this step is consistent in specification with the blank matrix sample in S11, but is not the same blank matrix sample. Understandably, the blank matrix sample in S11 is not used to culture the microbial population, but the microbial population is cultured in a different blank matrix sample of the same specification.

[0079] The microbial population spectrum data is obtained by analyzing the difference between the spectrum data of the plurality of microbial population samples and the spectrum data of the corresponding blank matrix sample, that is, subtracting the spectrum data of the corresponding blank matrix sample from the spectrum data of the plurality of microbial population samples to obtain the corresponding microbial population spectrum data. The corresponding blank matrix sample of each microbial population sample is the blank matrix used to culture the microbial population in the microbial population sample.

[0080] The spectrum data of the microbial population sample is formed by superimposing the matrix components and the microbial population components. When directly analyzing the original spectrum of the microbial sample, the absorbance of many wavebands in the spectrum changes depending on the change in the chemical substances in the matrix, rather than the characteristics of the microbial population itself. Therefore, the spectrum data of the microbial population sample can be subtracted from the spectrum data of the corresponding matrix to obtain the spectrum data of the microbial population, and then the spectrum data of different microbial populations can be analyzed.

[0081] In one embodiment, the content of each microbial population in the plurality of microbial population samples and the absorbance of each waveband in the corresponding microbial population spectrum data are analyzed to obtain the absorbance influence coefficient between any two microbial populations under each waveband, including:

[0082] The content of each bacterial species in multiple single bacterial samples, the content of each bacterial species in multiple mixed bacterial samples, and the absorbance of each band in the corresponding bacterial spectral data were analyzed to obtain the absorbance influence coefficient between any two bacterial species in each band. Among them, the bacterial samples include single bacterial samples and mixed bacterial samples. One bacterial species is cultured in a single bacterial sample, and at least two bacterial species are cultured in a mixed bacterial sample.

[0083] Furthermore, the analysis of the content of each bacterial species in multiple single-species samples, the content of each bacterial species in multiple mixed-species samples, and the absorbance of each band in the corresponding bacterial spectral data, to obtain the absorbance influence coefficient between any two bacterial species in each band, includes:

[0084] The influence coefficient of the absorbance of the second species on the first species in each band is obtained by multiplying the single-species ratio by the content of the first species in multiple mixed-species samples, the content of the second species in multiple mixed-species samples, and the absorbance of each band in the corresponding mixed-species spectral data. The single-species ratio indicates the reciprocal of the ratio of the content of the first species in multiple single-species samples to the absorbance of each band in the corresponding single-species spectral data. The first species and the second species are any two different species among the multiple species. The single-species spectral data is the spectral data corresponding to the single-species sample in the spectral data of the species. The mixed-species spectral data is the spectral data corresponding to the mixed-species sample in the spectral data of the species.

[0085] Among them, the absorbance indication of each band in the corresponding bacterial species spectral data: the absorbance of each band in the spectral data of each bacterial species sample. Absorbance in each wavelength band.

[0086] Similarly, the absorbance indication of each band in the corresponding mixed bacterial strain spectral data can be obtained as follows: for each mixed bacterial strain sample, the absorbance of each band in the spectral data of each mixed bacterial strain is indicated as follows: Absorbance in each wavelength band.

[0087] To analyze the similarities and differences in the spectral data of different bacterial communities, we consider using the spectral data of each bacterial species, including the regions of similarity and difference. The spectral data of a single bacterial species sample corresponding to the first bacterial species and the data containing the first bacterial species are shown in the figure. The spectral data of the mixed bacterial strains were compared with those of the first bacterial strain. When there are significant differences between the spectral data of individual bacterial species, then in the original... Based on the spectral data of individual bacterial species, the bands with larger variations are more significantly affected by mixed bacteria, even if that band is in the [missing information]. The absorbance of a single bacterial species sample was higher than that of the first bacterial species. The contribution of each bacterial species should not be too high; otherwise, in the subsequent regression process, the estimation of microbial colonies may be biased due to the mixed influence of multiple bacterial groups.

[0088] Then the first Under the first band, the first For the first strain of bacteria The influence coefficient of absorbance of the strain for:

[0089] ;

[0090] in, Instruction No. Under the first band, the first For the first strain of bacteria The absorbance influence coefficient of the inoculum. Indicator single-strain sample culture A sample collection of bacterial species. Indicator single-strain sample culture The size of the set of bacterial species, that is, the number of samples in the set. Indicates the number of cultures in the mixed bacterial sample. The bacterial species and the first A sample collection of bacterial species. Indicator of mixed bacterial culture samples individual bacterial species and The size of the collection of individual bacterial species express The set of The serial number of each single bacterial species sample. express The set of The serial number of each mixed bacterial sample. Instruction No. The first mixed bacterial sample The absolute content of each bacterial species, expressed in CFU / g, indicates the actual number of bacterial species in the sample. Instruction No. In the single bacterial species sample, the first The absolute content of each bacterial species Instruction No. The first mixed bacterial sample The absolute content of each bacterial species Instruction No. In the single-species spectral data of the first single-species sample Absorbance in each wavelength band Instruction No. In the single-species spectral data of the first single-species sample Absorbance in each wavelength band.

[0091] Characterization in the In the mixed bacterial sample, the first The theoretical absorbance of each wavelength band This represents the difference between the theoretical absorbance and the actual absorbance, that is, it represents the first... The first bacterial species for the first The magnitude of the influence of each bacterial species This is to measure the unit number The difference in the content of each bacterial species is important for the first... The magnitude of the influence of each bacterial species.

[0092] S13. Determine the representative coefficient of the bacterial community for each band based on the number of multiple bacterial species, the aliasing coefficient of each band, and the influence coefficient of other bacterial species on the absorbance of the first bacterial species in the corresponding band.

[0093] The first bacterial species can be any one of multiple bacterial species.

[0094] The bacterial community representativeness coefficient indicates the degree of specificity of each band for the bacterial species that best represents that band, i.e., the [missing value]. The bacterial community representativeness coefficient of each band indicates the number of bands in the first band. The spectral characterization of the strongest bacterial species is determined by the spectral characteristics of the first band, reflecting the specificity of the first band. The relative dominance of the bacterial species that has the strongest spectral characterization in each band.

[0095] In one embodiment, the representativeness coefficient of the bacterial community in each band is determined based on the number of multiple bacterial species, the aliasing coefficient of each band, and the influence coefficient of other bacterial species on the absorbance of the first bacterial species in the corresponding band, including:

[0096] The spectral characterization intensity of the first species in the corresponding band is calculated by multiplying the negative aliasing coefficient of each band by the product of the influence coefficient of the absorbance of each second species on the first species in the corresponding band, the difference between the number of multiple species and 1.

[0097] The representative coefficient of the bacterial community in the corresponding band is determined based on the spectral intensity of the first bacterial species in the corresponding band.

[0098] Based on the calculated absorbance influence coefficient In utilizing the first When fitting the absolute content of the first bacterial species, the overall content of all other bacterial species for the first species is considered. The magnitude of the influence of the first bacterial species, then the first The bacterial species in the first Spectral characterization intensity in each band for:

[0099] ;

[0100] wherein, indicates the spectral characterization intensity of the th species under the th waveband, the th species is a first species, the th species is a second species, indicates the number of species samples, indicates the aliasing coefficient of the th waveband, indicates the absorbance influence coefficient of the th species on the th species under the th waveband, represents a normalization function.

[0101] Traditional feature waveband selection methods mainly consider the difference in correlation of spectral data under different wavebands. Common methods include SPA (Successive Projections Algorithm) and CARS (Competitive Adaptive Reweighted Sampling). The core logic of such feature waveband selection is to reduce the redundancy between wavebands, retain wavebands with larger information quantity and lower correlation, so as to improve the calculation efficiency and generalization ability of the model.

[0102] However, the traditional method does not consider the characteristics of the regression or fitting object, and does not consider that when a waveband has high absorbance on multiple species, even if the correlation between the waveband and multiple wavebands is low, the characterization intensity of the single bacterial population is also low.

[0103] According to the obtained spectral characterization intensity of different species, the maximum value of the spectral characterization intensity of each different species under each waveband can be calculated when performing feature waveband selection.

[0104] In an embodiment, the bacterial population representative coefficient of the corresponding waveband is determined according to the spectral characterization intensity of the first species under the corresponding waveband, comprising:

[0105] The maximum spectral characterization intensity is selected from the multiple spectral characterization intensities of the multiple first species under the corresponding waveband as the bacterial population representative coefficient of the corresponding waveband.

[0106] The spectral characterization intensity of the th waveband corresponding to the multiple th species can be obtained from the multiple th waveband corresponding to the multiple The maximum spectral characterization intensity is selected from the spectral characterization intensities of the strains as a strain population representative coefficient of the corresponding waveband, that is, the first waveband. The strain population representative coefficient of the first waveband The calculation formula is as follows: wherein, indicates the spectral characterization intensity of the first strain in the first waveband, indicates the spectral characterization intensity of the first strain in the first waveband, indicates the maximum spectral characterization intensity selected from the spectral characterization intensities of the strains corresponding to the first waveband. S14, according to the strain population representative coefficient of each waveband and the absorbance sequence of the corresponding waveband in the spectral data of the plurality of target samples, a characteristic waveband is selected from each waveband. The target samples include a plurality of blank matrix samples and a plurality of strain samples, and the spectral data of the target samples in the characteristic waveband is used to train the strain content detection model.

[0107] The absorbance sequence of the corresponding waveband in the spectral data of the plurality of target samples indicates that for each waveband, an absorbance sequence of the plurality of target samples in the waveband is obtained.

[0108] In one embodiment, according to the strain population representative coefficient of each waveband and the absorbance sequence of the corresponding waveband in the spectral data of the plurality of target samples, a characteristic waveband is selected from each waveband, comprising:

[0109] a. From the plurality of strain population representative coefficients of the plurality of wavebands, the waveband corresponding to the maximum strain population representative coefficient is selected as the first characteristic waveband in the selected characteristic waveband set;

[0110] b. According to the strain population representative coefficient of each candidate waveband, the absorbance sequence of the corresponding candidate waveband in the spectral data of the plurality of target samples, and the absorbance sequence of each characteristic waveband in the selected characteristic waveband set in the spectral data of the plurality of target samples, the gain coefficient of each candidate waveband is calculated; the candidate waveband is a waveband other than the characteristic waveband; each candidate waveband corresponds to an absorbance sequence, and each characteristic waveband corresponds to an absorbance sequence;

[0111] c. From the plurality of gain coefficients of all candidate wavebands, the candidate waveband with the maximum gain coefficient is selected, and the candidate waveband is added to the selected characteristic waveband set as a new characteristic waveband;

[0112] d. Repeat steps b and c until the gain coefficient of each candidate waveband is less than a preset threshold.

[0113]

[0114]

[0115] ​​​Further, the gain coefficient of each candidate wave band is calculated according to the colony representative coefficient of each candidate wave band, the absorbance sequence of the corresponding candidate wave band in the spectral data of the plurality of target samples, and the absorbance sequence of each characteristic wave band in the selected characteristic wave band set in the spectral data of the plurality of target samples, including:

[0116] The gain coefficient of each candidate wave band is calculated according to the product of the colony representative coefficient and the weight coefficient of each candidate wave band, and a plurality of Pearson correlation coefficients between the absorbance sequence of the corresponding candidate wave band in the spectral data of the plurality of target samples and the absorbance sequence of each characteristic wave band in the selected characteristic wave band set in the spectral data of the plurality of target samples.

[0117] The preset threshold is set by itself according to the actual situation, for example, 0.6, 0.65.

[0118] In the traditional characteristic wave band selection process, the first characteristic wave band is randomly selected, and the subsequent wave bands are continuously selected based on the first characteristic wave band. The present application considers the difference in colony representation degree, and selects the first characteristic wave band as the wave band with the maximum colony representative coefficient. That is, after obtaining the colony representative coefficients of all wave bands, the wave band with the maximum colony representative coefficient is determined as the first characteristic wave band in the selected characteristic wave band set, and the subsequent characteristic wave bands are selected according to the first characteristic wave band.

[0119] In the selection process of the subsequent wave bands, the gain function is introduced to judge the gain size of the subsequent candidate wave band, and then the selection of the subsequent characteristic wave band is performed. Therefore, the gain function of the candidate wave band can be calculated, the candidate wave band is selected from the plurality of wave bands, and the candidate wave band is represented by The calculation formula of the gain coefficient of the candidate wave band is:

[0120] ;

[0121] wherein, is the gain coefficient of the candidate wave band , the non-redundancy weight is represented by , which can be 0.7, indicates the colony representative coefficient of the first wave band, indicates the absorbance sequence of the first wave band in the plurality of spectral data of the plurality of target samples, the target samples include a plurality of blank matrix samples and a plurality of bacterial samples, indicates the first characteristic wave band in the selected characteristic wave band set, indicates the absorbance sequence of the first characteristic wave band in the plurality of spectral data of the plurality of target samples, and​​ absorbance sequences of the characteristic wavebands, indicating with the Pearson correlation coefficient between, absorbance sequences of the characteristic wavebands of the spectral data of the plurality of target samples, respectively calculating a plurality of Pearson correlation coefficients between the absorbance sequences of each characteristic waveband and the spectral data of the plurality of target samples, each characteristic waveband corresponding to an absorbance sequence, calculating a plurality of absolute values of the plurality of Pearson correlation coefficients, and subtracting 1 from each of the plurality of absolute values to identify the maximum value.

[0122] respectively calculating the gain coefficients of each candidate waveband, selecting a candidate waveband with the largest gain coefficient from all candidate wavebands as the next characteristic waveband, adding it to the selected characteristic wavebands, and calculating the gain coefficients of the remaining candidate wavebands until the gain coefficient of a candidate waveband is less than 0.6, then stopping the selection of characteristic wavebands.

[0123] The gain function for calculating the gain coefficient mainly consists of two parts, i.e., the degree of representing the flora and the non-redundancy, which is based on the traditional selection of characteristic wavebands and adds the selection of wavebands with high degree of representing the flora.

[0124] After screening a plurality of characteristic wavebands, the spectral response values corresponding to the obtained characteristic wavebands can be used as input features to construct a multi-output regression model to realize direct estimation of the contents of a plurality of floras. Specifically, the spectral data of each bacterial sample at the selected characteristic wavebands is combined into a feature vector, and the contents (CFU / g) of different floras are used as corresponding output labels. Through a training process, the mapping relationship between the input and the output is learned, so as to realize simultaneous prediction of the contents of a plurality of floras in unknown samples.

[0125] In the selection of the bacterial content detection model, machine learning or deep learning methods suitable for processing multi-output regression tasks can be used, such as multi-task neural networks, partial least squares regression (PLSR), or random forest regression. In this way, the contents of different floras can be distinguished and estimated simultaneously on the basis of sharing spectral information, effectively avoiding redundant calculation and potential error accumulation caused by independent modeling of single bacteria, and realizing efficient and accurate multi-flora detection.

[0126] ​​The present application runs through the complete process from spectral data acquisition, band aliasing degree determination, bacterial flora characteristic spectrum extraction, characteristic band screening to multi-bacterial flora content modeling. By introducing the joint evaluation of the band aliasing degree and the bacterial flora representative coefficient in the characteristic band selection stage, the misjudgment problem caused by band redundancy or material signal overlap in the traditional method can be effectively avoided. By deducting the spectral data of the blank matrix to extract the bacterial spectrum data, the corresponding relationship between the spectral characteristics and the target bacterial flora is further enhanced. Finally, the direct estimation of the multi-bacterial flora content is realized by combining multi-output modeling, which improves the accuracy and robustness of the detection. The overall scheme not only simplifies the analysis difficulty of multi-bacterial flora coexistence in complex matrix, but also improves the generalization ability and practical application value of the model, which can be widely applied to food safety detection and rapid quality control links.

[0127] It should be understood that although Figure 1 The steps in the flowchart of the present application are displayed in sequence according to the direction of the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps in the flowchart of the present application can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or sub-steps or stages of other steps.

[0128] The present application also provides a foodborne microorganism detection system based on spectral technology, as shown in Figure 2 The system comprises:

[0129] A first analysis module 21 is configured to perform correlation analysis on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each band in the spectral data of the plurality of blank matrix samples, to obtain an aliasing coefficient of each band;

[0130] A second analysis module 22 is configured to analyze the content of each bacterial strain in the plurality of bacterial strain samples and the absorbance of each band in the corresponding bacterial strain spectral data, to obtain an absorbance influence coefficient between any two bacterial strains under each band; the bacterial strain samples are obtained by culturing a plurality of bacterial strains in each blank matrix sample, and the bacterial strain spectral data is obtained according to the difference between the spectral data of the plurality of bacterial strain samples and the spectral data of the corresponding blank matrix samples;

[0131] The determining module 23 is configured to determine a colony representative coefficient of each wave band according to the number of the plurality of bacterial species, the aliasing coefficient of each wave band, and the absorbance influence coefficient of other bacterial species on the first bacterial species in the corresponding wave band; the first bacterial species is any bacterial species in the plurality of bacterial species; the colony representative coefficient indicates the specificity degree of each wave band to the bacterial species that is most characteristic in the wave band.

[0132] The screening module 24 is configured to screen a characteristic wave band from each wave band according to the colony representative coefficient of each wave band and the absorbance sequence of the corresponding wave band in the spectral data of the plurality of target samples; the target samples include a plurality of blank matrix samples and a plurality of bacterial species samples; the spectral data of the target samples in the characteristic wave band is used to train the bacterial content detection model.

[0133] Optionally, the first analyzing module is further configured to:

[0134] performing Pearson correlation analysis on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each wave band in the spectral data of the corresponding blank matrix sample, to obtain a plurality of Pearson correlation coefficients between each chemical substance and each wave band;

[0135] for any wave band, calculating an aliasing coefficient of the wave band according to the maximum value of the Pearson correlation coefficients between the wave band and all chemical substances and the average value of the Pearson correlation coefficients between all wave bands and all chemical substances.

[0136] For the system embodiment, since it basically corresponds to the method embodiment, the relevant parts are described in the method embodiment. The system embodiment described above is only illustrative, and the units described as separate components can or can not be physically separated, and the components of the unit can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present application.

[0137] Figure 3 A structural schematic diagram of an electronic device is shown for an example embodiment of the present application, which includes a memory, a processor, and a computer program stored in the memory and used to run on the processor, and the processor implements the method described in any of the above embodiments when executing the computer program. Figure 3 The electronic device 30 shown is only an example, and should not limit the functions and use range of the embodiments of the present application.

[0138] As Figure 3As shown, the electronic device 30 can be in the form of a general computing device, for example, it can be a server device. The components of the electronic device 30 can include, but are not limited to, the at least one processor 31 described above, the at least one memory 32 described above, a bus 33 that connects the different system components, including the memory 32 and the processor 31.

[0139] The bus 33 includes a data bus, an address bus, and a control bus.

[0140] The memory 32 can include volatile memory, such as random access memory (RAM) 321 and / or cache memory 322, and can further include non-volatile memory, such as read-only memory (ROM) 323.

[0141] The memory 32 can also include a program tool 325 (or utility tool) having a set of one or more program modules 324, such as an operating system, one or more application programs, other program modules, and program data, and can include an implementation of a network environment, for example, in each of these or some combination.

[0142] The processor 31 performs functions of various embodiments by executing computer program instructions stored in the memory 32. The processor 31 can include one or more processors, such as dual-core processors, quad-core processors, or other multi-core processors.

[0143] The electronic device 30 can also communicate with one or more external devices 34 (such as a keyboard or a pointing device, for example) by way of an input / output (I / O) interface 35. Additionally, the electronic device 30 can communicate with one or more networks, such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet, by way of a network adapter 36. As illustrated, the network adapter 36 communicates with the other modules of the electronic device 30 by way of the bus 33. It should be appreciated that other hardware and / or software modules can be used in conjunction with the electronic device 30, such as, for example, microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data archival storage systems, etc.

[0144] It should be noted that although several units / modules or sub-units / modules of the electronic device are mentioned in the foregoing detailed description, such a division is merely exemplary and not mandatory. Indeed, according to embodiments of the present application, features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, features and functions of one unit / module described above can be further divided into a plurality of units / modules.

[0145] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the method provided by any of the above embodiments.

[0146] More specifically, the readable storage medium can include, but is not limited to, a portable disc, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device or any suitable combination of the above.

[0147] It is understood by those skilled in the art that all or part of the processes of the above-mentioned embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, the computer program can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM) and memory bus dynamic RAM (RDRAM) and the like.

[0148] The embodiment of the present application further provides a computer program product, which includes a computer program, and the computer program is executed by a processor to implement the method provided by any of the above embodiments.

[0149] The program code of the computer program product for executing the present application can be written in any combination of one or more programming languages, and can be executed completely on a user device, partially on a user device, as a separate software package, partially on a user device and partially on a remote device, or completely on a remote device.

[0150] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described, however, any combination of the technical features is deemed to be within the scope of the present disclosure.

[0151] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the scope of the application. It should be pointed out that, for ordinary skilled persons in the art, several modifications and improvements can be made without departing from the concept of the present application, and these are within the protection scope of the present application.

[0152] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment mainly describes the difference from other embodiments.

Claims

1. A method for detecting foodborne microorganisms based on spectroscopic techniques, characterized in that, The method comprises: correlation analysis is performed on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each wave band in the spectral data of the plurality of blank matrix samples to obtain an aliasing coefficient of each wave band; the absorbance of each wave band in the spectral data of the corresponding bacteria is analyzed to obtain an absorbance influence coefficient between any two bacteria at each wave band; the bacteria sample is obtained by culturing a plurality of bacteria in each blank matrix sample, and the bacteria spectral data is obtained according to the difference between the spectral data of the plurality of bacteria samples and the spectral data of the corresponding blank matrix samples; a bacteria community representative coefficient of each wave band is determined according to the number of the plurality of bacteria, the aliasing coefficient of each wave band and the absorbance influence coefficient of other bacteria on the first bacteria at the corresponding wave band; the first bacteria is any bacteria in the plurality of bacteria; the bacteria community representative coefficient indicates the specificity degree of each wave band to the bacteria that is most strongly represented at the wave band; a characteristic wave band is selected from each wave band according to the bacteria community representative coefficient of each wave band and the absorbance sequence of the corresponding wave band in the spectral data of the plurality of target samples, the spectral response value corresponding to the obtained characteristic wave band is taken as an input feature, and a multi-output regression model is constructed to realize direct estimation of the content of the plurality of bacteria communities; the target sample includes the plurality of blank matrix samples and the plurality of bacteria samples, and the spectral data of the target sample at the characteristic wave band is used for training a bacteria content detection model; The formula for calculating the aliasing coefficient of the nth waveband is: The formula for calculating the aliasing coefficient of the nth waveband is:​ ； wherein, is the th waveband, indicates a sequence of spectral absorbance of the th waveband in spectral data of a plurality of blank matrix samples, indicates a sequence of content of the th chemical substance in a plurality of blank matrices, indicates a Pearson correlation analysis of the sequence of content of each chemical substance in a plurality of blank matrix samples and the sequence of absorbance of each waveband in spectral data of a plurality of blank matrix samples, indicates an average of the Pearson correlation coefficients of the th waveband with all chemical substances, indicates a maximum of the Pearson correlation coefficients of the th waveband with all chemical substances, prevents the denominator from being zero; No. Under the first band, the first For the first strain of bacteria Influence coefficient of absorbance of the strain for: ； in, Instruction No. Under the first band, the first For the first strain of bacteria The absorbance influence coefficient of the inoculum. Indicator single-strain sample culture A sample collection of bacterial species. Indicator single-strain sample culture The size of the set of bacterial species, that is, the number of samples in the set. Indicates the number of cultures in the mixed bacterial sample. The bacterial species and the first A sample collection of bacterial species. Indicator of mixed bacterial culture samples The bacterial species and the first The size of the collection of individual bacterial species express The set of The serial number of each single bacterial species sample. express The set of The serial number of each mixed bacterial sample. Instruction No. The first mixed bacterial sample The absolute content of each bacterial species, expressed in CFU / g, indicates the actual number of bacterial species in the sample. Instruction No. In the single bacterial species sample, the first The absolute content of each bacterial species Instruction No. The first mixed bacterial sample The absolute content of each bacterial species Instruction No. In the single-species spectral data of the first single-species sample Absorbance in each wavelength band Instruction No. The first mixed bacterial spectral data of the mixed bacterial strain sample Absorbance in each wavelength band; Characterization in the In the mixed bacterial sample, the first The theoretical absorbance of each wavelength band This represents the difference between the theoretical absorbance and the actual absorbance, that is, it represents the first... The first bacterial species for the first The magnitude of the influence of each bacterial species This is to measure the unit number The difference in the content of each bacterial species is important for the first... The magnitude of the influence of each bacterial species; The spectral characterization intensity of the first bacterial species in the first waveband is : ； wherein, indicates the spectral characterization intensity of the th species under the th wavelength band, the th species is the first species, the th species is the second species, indicates the number of species samples, indicates the aliasing coefficient of the th wavelength band, indicates the absorbance influence coefficient of the th species on the th species under the th wavelength band, represents a normalization function; No. Representative coefficient of bacterial community in each band The calculation formula is: ,in, Instruction No. The bacterial species in the first Spectral characterization intensity in each band Instructions from the first The multiple bands corresponding to the first Among the spectral characterization intensities of each bacterial species, the one with the largest spectral characterization intensity was selected. the method comprises: a. selecting the wave band corresponding to the maximum bacteria community representative coefficient from the plurality of bacteria community representative coefficients of the plurality of wave bands as the first characteristic wave band in the selected characteristic wave band set; b. calculating a gain coefficient of each candidate wave band according to the bacteria community representative coefficient of each candidate wave band, the absorbance sequence of the corresponding candidate wave band in the spectral data of the plurality of target samples, and the absorbance sequence of each characteristic wave band in the selected characteristic wave band set in the spectral data of the plurality of target samples; the candidate wave band is a wave band other than the characteristic wave band, each candidate wave band corresponds to an absorbance sequence, and each characteristic wave band corresponds to an absorbance sequence; c. selecting the candidate wave band with the maximum gain coefficient from the plurality of gain coefficients of all candidate wave bands, and adding the candidate wave band to the selected characteristic wave band set as a new characteristic wave band; d. repeating steps b and c until the gain coefficient of each candidate wave band is less than a preset threshold value; to-be-selected wave band gain coefficient The calculation formula is: ； wherein, is a gain coefficient of the selected waveband, is a gain coefficient of the selected waveband, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.7, represents a non-redundancy weight, which can be 0.

7.

2. A foodborne microorganism detection system based on spectroscopic technology, characterized by, the system comprises: a first analysis module configured to perform correlation analysis on the content sequence of each chemical substance in the plurality of blank matrix samples and the absorbance sequence of each wave band in the spectral data of the plurality of blank matrix samples to obtain an aliasing coefficient of each wave band; The second analysis module is configured to analyze the content of each bacterial strain in the plurality of bacterial strain samples and the absorbance of each wavelength band in the corresponding bacterial strain spectrum data, to obtain the absorbance influence coefficient between any two bacterial strains at each wavelength band; the bacterial strain samples are obtained by culturing a plurality of bacterial strains in each blank matrix sample respectively, and the bacterial strain spectrum data is obtained according to the difference between the spectrum data of the plurality of bacterial strain samples and the spectrum data of the corresponding blank matrix sample; The determination module is configured to determine the bacterial community representative coefficient of each wavelength band according to the number of the plurality of bacterial strains, the aliasing coefficient of each wavelength band, and the absorbance influence coefficient of other bacterial strains on the first bacterial strain at the corresponding wavelength band; the first bacterial strain is any bacterial strain in the plurality of bacterial strains; the bacterial community representative coefficient indicates the specificity of each wavelength band to the bacterial strain that is best represented at the wavelength band; The screening module is configured to screen the characteristic wavelength band from each wavelength band according to the bacterial community representative coefficient of each wavelength band and the absorbance sequence of the corresponding wavelength band in the spectrum data of the plurality of target samples, to take the obtained spectrum response value corresponding to the characteristic wavelength band as an input feature, and to construct a multi-output regression model to realize direct estimation of the plurality of bacterial community contents; the target samples include a plurality of blank matrix samples and a plurality of bacterial strain samples, and the spectrum data of the target samples at the characteristic wavelength band is used to train the bacterial strain content detection model; The first The aliasing coefficient of the first waveband The calculation formula is: ； wherein, is the overlapping coefficient of the th wavelength band, is a sequence of spectral absorbance of the th wavelength band in the spectral data of the plurality of blank matrix samples, is a sequence of content of the th chemical substance in the plurality of blank matrix samples, is a sequence of Pearson correlation coefficient of the th wavelength band and all chemical substances, is a sequence of maximum value of Pearson correlation coefficient of the th wavelength band and all chemical substances, to prevent the denominator from being 0; No. Under the first band, the first For the first strain of bacteria The influence coefficient of absorbance of the strain for: ； wherein, indicates the absorbance influence coefficient of the first bacteria on the second bacteria in the first wave band, indicates the sample set of the first bacteria cultured in the single bacteria sample, indicates the size of the set of the first bacteria cultured in the single bacteria sample, that is, the number of samples in the set, indicates the sequence number of the first single bacteria sample in the set of indicates the absolute content of the first bacteria in the first mixed bacteria sample, in units of CFU / g, the absolute content indicating the actual number of bacteria in the bacteria sample, indicates the absorbance of the first wave band in the single bacteria spectrum data of the first single bacteria sample, indicates the absorbance of the first wave band in the mixed bacteria spectrum data of the first mixed bacteria sample;​​​​​​​​​​​​​​​​​​​​​​​​​​​​ Characterization in the In the mixed bacterial sample, the first The theoretical absorbance of each wavelength band This represents the difference between the theoretical absorbance and the actual absorbance, that is, it represents the first... The first bacterial species for the first The magnitude of the influence of each bacterial species This is to measure the unit number The difference in the content of each bacterial species is important for the first... The magnitude of the influence of each bacterial species; The spectral characterization intensity of the first bacterial species in the first waveband is : ； wherein, indicates the spectral characterization intensity of the th species at the th waveband, the th species being a first species and the th species being a second species, indicates the number of species samples, indicates the aliasing coefficient of the th waveband, indicates the absorbance influence coefficient of the th species on the th species at the th waveband, denotes a normalization function; No. Representative coefficient of bacterial community in each band The calculation formula is: ,in, Instruction No. The bacterial species in the first Spectral characterization intensity in each band Instructions from the first The multiple bands corresponding to the first Among the spectral characterization intensities of each bacterial species, the one with the largest spectral characterization intensity was selected. The screening of the characteristic wavelength band from each wavelength band according to the bacterial community representative coefficient of each wavelength band and the absorbance sequence of the corresponding wavelength band in the spectrum data of the plurality of target samples includes: a. selecting the wavelength band corresponding to the maximum bacterial community representative coefficient from the plurality of bacterial community representative coefficients of the plurality of wavelength bands as the first characteristic wavelength band in the selected characteristic wavelength band set; b. calculating the gain coefficient of each candidate wavelength band according to the bacterial community representative coefficient of each candidate wavelength band, the absorbance sequence of the corresponding candidate wavelength band in the spectrum data of the plurality of target samples, and the absorbance sequence of each characteristic wavelength band in the selected characteristic wavelength band set in the spectrum data of the plurality of target samples; each candidate wavelength band is a wavelength band other than the characteristic wavelength band, each candidate wavelength band corresponds to an absorbance sequence, and each characteristic wavelength band corresponds to an absorbance sequence; c. selecting the candidate wavelength band with the maximum gain coefficient from the plurality of gain coefficients of all candidate wavelength bands, and adding the candidate wavelength band to the selected characteristic wavelength band set as a new characteristic wavelength band; d. repeating steps b and c until the gain coefficient of each candidate wavelength band is less than a preset threshold. to-be-selected wave band gain coefficient The calculation formula is: ； in, Candidate bands Gain coefficient, This represents the non-redundancy weight, which can be set to 0.

7. Instruction No. The representative coefficient of the bacterial community in each band, Multiple spectral data indicating multiple target samples in the first The absorbance sequences of each wavelength band were used to analyze the target samples, which included multiple blank matrix samples and multiple bacterial samples. Indicates the first in the selected feature band set One characteristic band, Multiple spectral data indicating multiple target samples in the first Absorbance sequence of each characteristic band, instruct and The Pearson correlation coefficient between them Indicating the spectral data of multiple target samples in the first... For each absorbance band, calculate multiple Pearson correlation coefficients between the absorbance sequences of the target samples and the absorbance sequences of each characteristic band. Each characteristic band corresponds to one absorbance sequence. Calculate multiple absolute values ​​of the multiple Pearson correlation coefficients, and subtract 1 from each of these absolute values ​​to confirm the results. The largest value in the middle.

Images