Multi-component substance fluorescence spectrum analysis method, device, equipment and storage medium

Abstract

The invention provides a multi-component substance fluorescence spectrum analysis method, device and equipment and a storage medium. The method comprises the following steps: acquiring a three-dimensional fluorescence fingerprint spectrum of a to-be-detected multi-component substance; inputting the three-dimensional fluorescence fingerprint spectrum of the multi-component substance into a pre-trained multi-component substance fluorescence spectrum model to obtain detection information; the detection information comprises the number of fluorescent components, the concentration of each fluorescent component and a three-dimensional fluorescent fingerprint spectrum corresponding to each fluorescent component; the multi-component substance fluorescence spectrum model comprises a multi-task prediction network and a component fitting network, and the multi-task prediction network predicts the number of fluorescence components of the to-be-detected multi-component substance and the concentration of each fluorescence component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance; and the component fitting network outputs the three-dimensional fluorescent fingerprint spectrum of each fluorescent component based on the three-dimensional fluorescent fingerprint spectrum of the multi-component substance and the prediction result of the number of the fluorescent components. According to the invention, the detection efficiency of multi-component substances can be greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of fluorescence detection technology, and in particular to a method, device, equipment and storage medium for fluorescence spectrum analysis of multi-component substances. Background Art

[0002] Three-dimensional fluorescence spectroscopy detection has the advantages of strong selectivity, high sensitivity, simple sample pretreatment, fast measurement speed, low measurement cost, and easy automation. It is widely used in environmental monitoring, food monitoring and other fields.

[0003] In the detection of three-dimensional fluorescence spectra, the analytes usually coexist in the form of multiple components, and the fluorescence spectra of homologues are highly similar, resulting in spectral aliasing, which makes it impossible to directly and effectively extract and analyze the fluorescence information of multiple components. At present, the commonly used analytical method for analyzing multi-component aliasing spectra is the chemometric method, which replaces the traditional "physical and chemical separation" with "mathematical separation" to effectively solve the problem that the target analytes cannot be qualitatively identified and quantitatively analyzed due to overlapping fluorescence signals.

[0004] However, the chemometric analysis procedure is complex and time-consuming, and has strict requirements on data. At the same time, in practical application, this method requires the collection of three-dimensional spectral data of multiple sets of samples to ensure the accuracy of component decomposition. Summary of the invention

[0005] The embodiments of the present invention provide a method, device, equipment and storage medium for fluorescence spectrum analysis of multi-component substances to solve the problem that current detection methods are complicated and time-consuming.

[0006] In a first aspect, an embodiment of the present invention provides a method for fluorescence spectrum analysis of a multi-component substance, comprising: Obtaining a three-dimensional fluorescence fingerprint of the multi-component substance to be tested; Inputting the three-dimensional fluorescence fingerprint of the multi-component substance into a pre-trained fluorescence spectrum model of the multi-component substance to obtain detection information of the multi-component substance to be tested; wherein the detection information includes the number of fluorescent components, the concentration of each fluorescent component, and the three-dimensional fluorescence fingerprint corresponding to each fluorescent component; Among them, the fluorescence spectrum model of multi-component substances includes a multi-task prediction network and a component fitting network. The multi-task prediction network is used to simultaneously predict the number of fluorescent components of the multi-component substance to be tested and the concentration of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance; the component fitting network is used to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance and the prediction results of the number of fluorescent components. In a possible implementation, the multi-task prediction network includes a feature extraction subnetwork, a fluorescent component number prediction subnetwork, and a fluorescent component concentration prediction subnetwork; The feature extraction subnetwork is used to extract feature information from the three-dimensional fluorescence fingerprint of multi-component substances; The fluorescence component number prediction subnetwork is used to classify the three-dimensional fluorescence fingerprint of multi-component substances based on feature information to obtain the number of fluorescent components; The fluorescence component concentration prediction subnetwork is used to predict the concentration of each fluorescent component in the three-dimensional fluorescence fingerprint of a multi-component substance based on the characteristic information to obtain the concentration of each fluorescent component. In one possible implementation, the fluorescent component number prediction subnetwork uses a cross entropy loss function, and the fluorescent component concentration prediction subnetwork uses a mean square error loss function.

[0007] In a possible implementation, the component fitting network includes a plurality of sub-fitting networks, each of which includes a different number of single fluorescence component fitting networks; The component fitting network is used to input the three-dimensional fluorescence fingerprint spectrum of the multi-component substance into the target sub-fitting network based on the prediction result of the number of fluorescent components, and output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component; wherein the number of single fluorescent component fitting networks contained in the target sub-fitting network is the same as the prediction result of the number of fluorescent components, and the target sub-fitting network is any sub-fitting network among the multiple sub-fitting networks.

[0008] In a possible implementation, the component fitting network includes a plurality of single fluorescence component fitting networks; The component fitting network is used to output the three-dimensional fluorescence fingerprint of each fluorescent component based on the prediction results of the number of fluorescent components and the three-dimensional fluorescence fingerprint of the multi-component substance.

[0009] In one possible implementation, both the multi-task prediction network and the component fitting network adopt the MobileNetV4 network architecture.

[0010] In a possible implementation, the training process of the multi-component substance fluorescence spectrum model is: Collecting a training sample set; wherein the training sample set includes a plurality of training samples, each training sample is a multi-component substance, and each training sample includes a three-dimensional fluorescence fingerprint spectrum of the multi-component substance, the number of fluorescent components corresponding to the three-dimensional fluorescence fingerprint spectrum of the multi-component substance, the concentration of each fluorescent component, and the three-dimensional fluorescence fingerprint spectrum of each fluorescent component; The training sample set is input into a pre-built multi-component material fluorescence spectrum model to train the multi-component material fluorescence spectrum model.

[0011] In a second aspect, an embodiment of the present invention provides a device for fluorescence spectrum analysis of a multi-component substance, comprising: An acquisition module, used to acquire a three-dimensional fluorescence fingerprint of a multi-component substance to be tested; A detection module, used to input the three-dimensional fluorescence fingerprint of the multi-component substance into a pre-trained fluorescence spectrum model of the multi-component substance to obtain detection information of the multi-component substance to be tested; wherein the detection information includes the number of fluorescent components, the concentration of each fluorescent component and the three-dimensional fluorescence fingerprint corresponding to each fluorescent component; Among them, the fluorescence spectrum model of multi-component substances includes a multi-task prediction network and a component fitting network. The multi-task prediction network is used to simultaneously predict the number of fluorescent components of the multi-component substance to be tested and the concentration of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance; the component fitting network is used to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance and the prediction results of the number of fluorescent components.

[0012] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the processor implements the steps of the method described in the first aspect or any possible implementation manner of the first aspect.

[0013] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method described in the first aspect or any possible implementation method of the first aspect are implemented.

[0014] The embodiments of the present invention provide a method, device, equipment and storage medium for fluorescence spectrum analysis of multi-component substances. In order to improve the operability of detection, the three-dimensional fluorescence fingerprint spectrum of the multi-component substance to be tested can be directly input into a pre-trained fluorescence spectrum model of the multi-component substance, and the detection information of the multi-component substance to be tested can be obtained. Among them, the detection information includes not only the number of fluorescent components and the concentration of each fluorescent component, but also the three-dimensional fluorescence fingerprint spectrum corresponding to each fluorescent component. Once the model is established, in actual applications, only the three-dimensional fluorescence spectrum of a single sample to be tested needs to be input to simultaneously identify the spectrum and quantify the concentration of multiple fluorescent components, thereby greatly improving the detection efficiency of three-dimensional fluorescence spectrum for multi-component substances. BRIEF DESCRIPTION OF THE DRAWINGS In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0015] Figure 1It is a flow chart of the implementation of the method for fluorescence spectrum analysis of multi-component substances provided by an embodiment of the present invention; Figure 2 is a block diagram of a method for fluorescence spectrum analysis of multi-component substances provided by an embodiment of the present invention; Figure 3 is a schematic diagram of a fluorescence spectrum model of a multi-component substance provided by an embodiment of the present invention; Figure 4 is a schematic diagram of a device for fluorescence spectrum analysis of multi-component substances provided in an embodiment of the present invention; Figure 5 is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0016] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0018] As introduced in the background technology, currently in the three-dimensional fluorescence spectrum detection, the chemometric method is mainly used for analysis. However, the chemometric analysis procedure is complicated and time-consuming, and the data needs to be processed multiple times, the steps are cumbersome, and there are strict requirements on the data. At the same time, in practical applications, it is necessary to collect three-dimensional spectral data of multiple groups of samples to ensure the accuracy of component decomposition, and the number of samples is greater than or at least equal to the number of pure components. The traditional PARAFAC method can be analyzed by at least 21 groups of data, so that several substances and related concentration relationships can be analyzed in the 21 groups of data, but the above method needs to process the data multiple times, and the steps are cumbersome. These problems limit the rapid analysis and identification of multi-component mixtures by chemometric methods such as PARAFAC, resulting in only staying in the laboratory sample analysis stage. In order to solve the problems of the prior art, the embodiments of the present invention provide a method, device, equipment and storage medium for fluorescence spectrum analysis of multi-component substances. The method for fluorescence spectrum analysis of multi-component substances provided by the embodiments of the present invention is first introduced below.

[0019] See also Figure 1 and Figure 2 , which shows a flowchart of the method for fluorescence spectrum analysis of multi-component substances provided by an embodiment of the present invention, and is described in detail as follows: S110, obtaining a three-dimensional fluorescence fingerprint of the multi-component substance to be tested. Three-dimensional fluorescence fingerprint is a spectral technology that expresses fluorescence intensity as the three-dimensional coordinates of excitation wavelength, emission wavelength and fluorescence intensity. In the three-dimensional fluorescence fingerprint, the excitation wavelength and emission wavelength are used as two independent variables, and the fluorescence intensity measured at each pair of excitation-emission wavelengths is used as the third variable, thus constructing a three-dimensional spectral characteristic diagram.

[0020] In order to detect the object to be detected, it is necessary to obtain a three-dimensional fluorescence fingerprint of the multi-component substance to be detected before the detection.

[0021] S120, inputting the three-dimensional fluorescence fingerprint of the multi-component substance into a pre-trained fluorescence spectrum model of the multi-component substance to obtain detection information of the multi-component substance to be tested.

[0022] The detection information includes the number of fluorescent components, the concentration of each fluorescent component and the three-dimensional fluorescent fingerprint map corresponding to each fluorescent component.

[0023] The multi-component material fluorescence spectrum model includes a multi-task prediction network and a component fitting network. The multi-task prediction network is used to predict the number of fluorescent components of the multi-component material to be tested and the concentration of each fluorescent component based on the three-dimensional fluorescence fingerprint of the multi-component material. The component fitting network is used to output the three-dimensional fluorescence fingerprint of each fluorescent component based on the three-dimensional fluorescence fingerprint of the multi-component material and the prediction results of the number of fluorescent components.

[0024] In some embodiments, the training process of the multi-component substance fluorescence spectrum model is: First, a training sample set is collected. The training sample set includes multiple training samples, each of which is a multi-component substance, and each training sample includes a three-dimensional fluorescence fingerprint of the multi-component substance, the number of fluorescent components corresponding to the three-dimensional fluorescence fingerprint of the multi-component substance, the concentration of each fluorescent component, and the three-dimensional fluorescence fingerprint of each fluorescent component. It should be noted that the number of fluorescent components corresponding to the three-dimensional fluorescence fingerprint of the multi-component substance, the concentration of each fluorescent component, and the three-dimensional fluorescence fingerprint of each fluorescent component can be obtained directly from the experiment, or can be obtained by performing parallel factor analysis on the multi-component substance. The three-dimensional fluorescence fingerprint of each fluorescent component here refers to the three-dimensional fluorescence fingerprint of each single-component substance.

[0025] Then, the training sample set is input into a pre-built multi-component material fluorescence spectrum model to train the multi-component material fluorescence spectrum model.

[0026] In some embodiments, the multi-task prediction network includes a feature extraction subnetwork, a fluorescent component number prediction subnetwork, and a fluorescent component concentration prediction subnetwork. The feature extraction subnetwork is used to extract feature information from the three-dimensional fluorescent fingerprint spectrum of the multi-component substance, and the fluorescent component number prediction subnetwork is used to classify the three-dimensional fluorescent fingerprint spectrum of the multi-component substance based on the feature information to obtain the number of each individual fluorescent component. The fluorescent component concentration prediction subnetwork is used to predict the concentration of each individual fluorescent component in the three-dimensional fluorescent fingerprint spectrum of the multi-component substance based on the feature information to obtain the concentration of each individual fluorescent component.

[0027] In this embodiment, the multi-task prediction network and the component fitting network are both built based on the MobileNetV4 network architecture.

[0028] After the three-dimensional fluorescence fingerprint of the multi-component substance is input into the multi-task prediction network constructed by the MobileNetV4 network architecture, it first undergoes feature extraction, and then the extracted features are input into two branches of the MobileNetV4 network architecture, one branch is used to predict the number of each single fluorescent component, and the other is used to predict the concentration of each single fluorescent component, so that the number of fluorescent components and the concentration of each fluorescent component of the multi-component substance to be tested can be determined at the same time.

[0029] In some embodiments, the fluorescent component number prediction subnetwork is used to classify the three-dimensional fluorescent fingerprint of a multi-component substance based on feature information to obtain the number of fluorescent components.

[0030] In this embodiment, the fluorescent component number prediction subnetwork is a branch of the multi-task prediction network, and the cross entropy loss function is used. The cross entropy loss function is suitable for the fluorescent component number prediction subnetwork because it can handle classification problems well. When predicting the number of fluorescent components, it is actually performing a classification task, such as determining whether there are 1, 2 or more components, etc. The cross entropy loss function can effectively measure the difference between the predicted value and the true value, thereby prompting the network to learn the correct classification mode.

[0031] In some embodiments, the fluorescent component concentration prediction subnetwork is a branch of the multi-task prediction network, which uses a mean square error loss function to predict the concentration of each fluorescent component in the three-dimensional fluorescence fingerprint spectrum of the multi-component substance based on the feature information to obtain the concentration of each fluorescent component.

[0032] In this embodiment, the task of the fluorescent component concentration prediction subnetwork is to predict the concentration as a continuous variable. The mean square error loss function calculates the average of the squared difference between the predicted value and the true value, which is very suitable for evaluating the accuracy of continuous variable prediction. By minimizing the mean square error, the fluorescent component concentration prediction subnetwork can continuously adjust its own parameters to more accurately predict the concentration value.

[0033] In some embodiments, the component fitting network includes a plurality of sub-fitting networks, each sub-fitting network including a different number of single fluorescent component fitting networks.

[0034] Exemplarily, the component fitting network includes one single fluorescence component fitting network, two single fluorescence component fitting networks, three single fluorescence component fitting networks...N single fluorescence component fitting networks.

[0035] The component fitting network is used to input the three-dimensional fluorescence fingerprint of the multi-component substance into the target sub-fitting network based on the prediction result of the number of fluorescent components to output the three-dimensional fluorescence fingerprint of each fluorescent component. Among them, the number of single fluorescent component fitting networks contained in the target sub-fitting network is the same as the prediction result of the number of fluorescent components, and the target sub-fitting network is any sub-fitting network among the multiple sub-fitting networks.

[0036] Exemplarily, if the predicted result of the number of fluorescent components is 4 components, the three-dimensional fluorescence fingerprint spectrum of the multi-component substance is input into a fitting network containing at least 4 single fluorescent components to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component.

[0037] In some embodiments, the component fitting network may also include multiple single fluorescence component fitting networks, and the component fitting network is used to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component based on the prediction results of the number of fluorescent components and the three-dimensional fluorescence fingerprint spectrum of the multi-component substance.

[0038] In this embodiment, if the predicted result of the number of fluorescent components is 4 components, the component fitting network is used to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component based on the 4 components and the three-dimensional fluorescence fingerprint spectrum of the multi-component substance based on the number of fluorescent components. The first 4 single fluorescent component fitting networks in the component fitting network are used to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component.

[0039] Exemplarily, the component fitting network includes 4 single fluorescent component fitting networks, each of which is composed of a MobileNetV4 network architecture. For example, if the prediction result of the number of fluorescent components is 3 components, the component fitting network will use 3 of the single fluorescent component fitting networks to output the 3D fluorescent fingerprints of the 3 fluorescent components respectively.

[0040] The method for fluorescence spectrum analysis of multi-component substances provided by the present invention can directly input the three-dimensional fluorescence fingerprint spectrum of the multi-component substance to be tested into a pre-trained fluorescence spectrum model of the multi-component substance in order to improve the operability of the detection, so as to obtain the detection information of the multi-component substance to be tested. Among them, the detection information includes not only the number of fluorescent components and the concentration of each fluorescent component, but also the three-dimensional fluorescence fingerprint spectrum corresponding to each fluorescent component. Once the model is established, in actual applications, only the three-dimensional fluorescence spectrum of a single sample to be tested needs to be input to obtain the detection information of the multi-component substance to be tested, which includes not only the number of fluorescent components and the concentration of each fluorescent component, but also the three-dimensional fluorescence fingerprint spectrum corresponding to each fluorescent component. Thereby improving the detection efficiency of three-dimensional fluorescence.

[0041] In order to facilitate understanding of the three-dimensional fluorescence detection method provided by the present invention, Figure 3 As shown, the detection method is applied to the detection of pollutants in tap water as an example for explanation. Figure 3 It includes a multi-task network and a component fitting network. The architecture of the multi-task network is MobileNetV4. The multi-task network is used to predict the number of components and the concentration of each component. Figure 3 The model in can predict up to 4 component numbers, corresponding to Figure 3 1, 2, 3 and 4 in , C1 corresponds to the predicted concentration of 1 group, C2 corresponds to the predicted concentration of 2 groups, C3 corresponds to the predicted concentration of 3 groups, and C4 corresponds to the predicted concentration of 4 groups. The component fitting network includes 4 single fluorescent component fitting networks, each of which is composed of the MobileNetV4 network architecture, namely MobileNetV4_1, MobileNetV4_2, MobileNetV4_3 and MobileNetV4_4.

[0042] Pollutants in tap water often coexist in the form of multiple components, and the fluorescence spectra of homologues are highly similar, and there is a spectral aliasing phenomenon, which makes it impossible to directly and effectively extract and analyze the fluorescence information of multiple components. In order to quickly and effectively analyze the components of pollutants, a training sample set of the fluorescence spectrum model of multi-component substances is first constructed. Under laboratory conditions, a three-dimensional fluorescence experiment of a common three-component pollutant mixture in tap water is designed, including phenol, resorcinol, and salicylic acid, which are common in tap water, to construct a fluorescence fingerprint spectrum data set of multi-component pollutants in tap water. The training sample set includes multiple training samples, each of which is a three-dimensional fluorescence fingerprint spectrum of a multi-component substance, and the analytical data corresponding to the three-dimensional fluorescence fingerprint spectrum of the multi-component substance. The analytical data is obtained by performing parallel factor analysis on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance, and the analytical data at least includes the number of fluorescent components corresponding to the three-dimensional fluorescence fingerprint spectrum of the multi-component substance, the concentration of each fluorescent component, and the three-dimensional fluorescence fingerprint spectrum of each fluorescent component. It should also be noted here that the number of fluorescent components corresponding to the three-dimensional fluorescence fingerprint of the multi-component substance, the concentration of each fluorescent component and the three-dimensional fluorescence fingerprint of each fluorescent component can also be experimental data obtained in the experiment. If there are experimental data, there is no need to perform parallel factor analysis.

[0043] Then, a multi-component fluorescence spectrum model is constructed. The multi-component fluorescence spectrum model includes a multi-task prediction network and a component fitting network. Both the multi-task prediction network and the component fitting network are built based on the MobileNetV4 network architecture. The component fitting network includes 4 single fluorescence component fitting networks. Each single fluorescence component fitting network is also built based on the MobileNetV4 architecture. Therefore, it can be used to detect pollutants with less than or equal to 4 components.

[0044] Next, the constructed multi-component substance fluorescence spectrum model is trained using the training sample set to obtain a trained multi-component substance fluorescence spectrum model.

[0045] Finally, the three-dimensional fluorescence fingerprint of the multi-component substance to be tested is input into the trained multi-component substance fluorescence spectrum model to obtain the number of fluorescent components in the substance to be tested, the concentration of each fluorescent component and the three-dimensional fluorescence fingerprint corresponding to each fluorescent component.

[0046] In addition, when performing component fitting, the three-dimensional fluorescence fingerprint spectrum of the multi-component substance to be tested needs to be gray-processed, and the gray-processed spectrum is input into the component fitting network to output the three-dimensional fluorescence fingerprint spectrum corresponding to each fluorescent component.

[0047] It should be understood that the order of execution of the steps in the above embodiment does not necessarily 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 embodiment of the present invention.

[0048] Based on the method for fluorescence spectrum analysis of multi-component substances provided in the above embodiment, the present invention also provides a specific implementation of a device for fluorescence spectrum analysis of multi-component substances applied to the method for fluorescence spectrum analysis of multi-component substances. Please refer to the following embodiment.

[0049] like Figure 4 As shown, a device 400 for fluorescence spectrum analysis of multi-component substances is provided, the device comprising: An acquisition module 410 is used to acquire a three-dimensional fluorescence fingerprint of a multi-component substance to be tested; The detection module 420 is used to input the three-dimensional fluorescence fingerprint of the multi-component substance into a pre-trained fluorescence spectrum model of the multi-component substance to obtain the detection information of the multi-component substance to be tested; wherein the detection information includes the number of fluorescent components, the concentration of each fluorescent component and the three-dimensional fluorescence fingerprint corresponding to each fluorescent component; Among them, the fluorescence spectrum model of multi-component substances includes a multi-task prediction network and a component fitting network. The multi-task prediction network is used to predict the number of fluorescent components of the multi-component substance to be tested and the concentration of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance; the component fitting network is used to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance and the prediction results of the number of fluorescent components.

[0050] In a possible implementation, the multi-task prediction network includes a feature extraction subnetwork, a fluorescent component number prediction subnetwork, and a fluorescent component concentration prediction subnetwork; The feature extraction subnetwork is used to extract feature information from the three-dimensional fluorescence fingerprint of multi-component substances; The fluorescence component number prediction subnetwork is used to classify the three-dimensional fluorescence fingerprint of multi-component substances based on feature information to obtain the number of fluorescent components; The fluorescence component concentration prediction subnetwork is used to predict the concentration of each fluorescent component in the three-dimensional fluorescence fingerprint of a multi-component substance based on the characteristic information to obtain the concentration of each fluorescent component.

[0051] In one possible implementation, the fluorescent component number prediction subnetwork uses a cross entropy loss function, and the fluorescent component concentration prediction subnetwork uses a mean square error loss function.

[0052] In a possible implementation, the component fitting network includes a plurality of sub-fitting networks, each of which includes a different number of single fluorescence component fitting networks; The component fitting network is used to input the three-dimensional fluorescence fingerprint spectrum of the multi-component substance into the first sub-fitting network based on the prediction result of the number of fluorescent components, and output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component; wherein the number of single fluorescent component fitting networks contained in the first sub-fitting network is the same as the prediction result of the number of fluorescent components.

[0053] In a possible implementation, the component fitting network includes a plurality of single fluorescence component fitting networks; The component fitting network is used to output the three-dimensional fluorescence fingerprint of each fluorescent component based on the prediction results of the number of fluorescent components and the three-dimensional fluorescence fingerprint of the multi-component substance.

[0054] In one possible implementation, both the multi-task prediction network and the component fitting network adopt the MobileNetV4 network architecture.

[0055] In a possible implementation, the training process of the multi-component substance fluorescence spectrum model is: Collect a training sample set; wherein the training sample set includes multiple training samples, each of which is a three-dimensional fluorescence fingerprint of a multi-component substance, and analytical data corresponding to the three-dimensional fluorescence fingerprint of the multi-component substance, the analytical data being obtained by performing parallel factor analysis on the three-dimensional fluorescence fingerprint of the multi-component substance, and the analytical data at least includes the number of fluorescent components corresponding to the three-dimensional fluorescence fingerprint of the multi-component substance, the concentration of each fluorescent component and the three-dimensional fluorescence fingerprint of each fluorescent component; it should also be noted here that the number of fluorescent components corresponding to the three-dimensional fluorescence fingerprint of the multi-component substance, the concentration of each fluorescent component and the three-dimensional fluorescence fingerprint of each fluorescent component may also be experimental data obtained in the experiment. If there is experimental data, there is no need to perform parallel factor decomposition.

[0056] The training sample set is input into a pre-built multi-component material fluorescence spectrum model to train the multi-component material fluorescence spectrum model.

[0057] Figure 5 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 5 As shown, the electronic device 5 of this embodiment includes: a processor 50, a memory 51, and a computer program 52 stored in the memory 51 and executable on the processor 50. When the processor 50 executes the computer program 52, the steps in the above-mentioned method embodiments for fluorescence spectrum analysis of multi-component substances are implemented, for example Figure 1 Alternatively, when the processor 50 executes the computer program 52, the functions of each module in the above-mentioned device embodiments are realized, for example, Figure 4 The functions of modules 410 to 420 are shown.

[0058] Exemplarily, the computer program 52 may be divided into one or more modules, which are stored in the memory 51 and executed by the processor 50 to implement the present invention. The one or more modules may be a series of computer program instruction segments that can implement specific functions, and the instruction segments are used to describe the execution process of the computer program 52 in the electronic device 5. For example, the computer program 52 may be divided into Figure 4 Modules 410 to 420 are shown.

[0059] The electronic device 5 may include, but is not limited to, a processor 50 and a memory 51. Those skilled in the art will appreciate that Figure 5 It is only an example of the electronic device 5 and does not constitute a limitation of the electronic device 5. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0060] The processor 50 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0061] The memory 51 may be an internal storage unit of the electronic device 5, such as a hard disk or memory of the electronic device 5. The memory 51 may also be an external storage device of the electronic device 5, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 5. Further, the memory 51 may also include both an internal storage unit of the electronic device 5 and an external storage device. The memory 51 is used to store the computer program and other programs and data required by the electronic device. The memory 51 may also be used to temporarily store data that has been output or is to be output.

[0062] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0063] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0064] Those skilled 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 the present invention.

[0065] In the embodiments provided by the present invention, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0066] 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.

[0067] In addition, each functional unit in each embodiment of the present invention 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. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0068] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned method embodiments for fluorescence spectrum analysis of each multi-component substance can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium.

[0069] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for fluorescence spectrum analysis of multi-component substances, characterized in that: include: Obtaining a three-dimensional fluorescence fingerprint of the multi-component substance to be tested; Inputting the three-dimensional fluorescence fingerprint of the multi-component substance into a pre-trained fluorescence spectrum model of the multi-component substance to obtain detection information of the multi-component substance to be tested; wherein the detection information includes the number of fluorescent components, the concentration of each fluorescent component and the three-dimensional fluorescence fingerprint corresponding to each fluorescent component; Among them, the multi-component substance fluorescence spectrum model includes a multi-task prediction network and a component fitting network. The multi-task prediction network is used to simultaneously predict the number of fluorescent components of the multi-component substance to be tested and the concentration of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance; the component fitting network is used to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance and the prediction results of the number of the fluorescent components.

2. The method for fluorescence spectrum analysis of multi-component substances according to claim 1, characterized in that: The multi-task prediction network includes a feature extraction subnetwork, a fluorescent component number prediction subnetwork and a fluorescent component concentration prediction subnetwork; The feature extraction subnetwork is used to extract feature information from the three-dimensional fluorescence fingerprint of the multi-component substance; The fluorescent component number prediction subnetwork is used to classify the three-dimensional fluorescent fingerprint spectrum of the multi-component substance based on the characteristic information to obtain the number of the fluorescent components; The fluorescent component concentration prediction subnetwork is used to predict the concentration of each fluorescent component in the three-dimensional fluorescent fingerprint of the multi-component substance based on the characteristic information to obtain the concentration of each fluorescent component.

3. The method for fluorescence spectrum analysis of multi-component substances according to claim 2, characterized in that: The fluorescent component number prediction subnetwork uses a cross entropy loss function, and the fluorescent component concentration prediction subnetwork uses a mean square error loss function.

4. The method for fluorescence spectrum analysis of multi-component substances according to claim 1, characterized in that: The component fitting network includes a plurality of sub-fitting networks, each of which includes a different number of single fluorescent component fitting networks; The component fitting network is used to input the three-dimensional fluorescence fingerprint spectrum of the multi-component substance into the target sub-fitting network based on the prediction result of the number of the fluorescent components, and output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component; wherein the number of single fluorescent component fitting networks contained in the target sub-fitting network is the same as the prediction result of the number of fluorescent components in the target sub-fitting network, and the target sub-fitting network is any one of the multiple sub-fitting networks.

5. The method for fluorescence spectrum analysis of multi-component substances according to claim 1, characterized in that: The component fitting network includes a plurality of single fluorescence component fitting networks; The component fitting network is used to output the three-dimensional fluorescence fingerprint of each fluorescent component based on the prediction result of the number of the fluorescent components and the three-dimensional fluorescence fingerprint of the multi-component substance.

6. The method for fluorescence spectrum analysis of multi-component substances according to any one of claims 1 to 5, characterized in that: The multi-task prediction network and component fitting network both adopt the MobileNetV4 network architecture.

7. The method for fluorescence spectrum analysis of multi-component substances according to any one of claims 1 to 5, characterized in that: The training process of the multi-component substance fluorescence spectrum model is as follows: Collecting a training sample set; wherein the training sample set includes a plurality of training samples, each training sample is a multi-component substance, and each training sample includes a three-dimensional fluorescence fingerprint spectrum of the multi-component substance, the number of fluorescent components corresponding to the three-dimensional fluorescence fingerprint spectrum of the multi-component substance, the concentration of each fluorescent component, and the three-dimensional fluorescence fingerprint spectrum of each fluorescent component; The training sample set is input into a pre-constructed multi-component substance fluorescence spectrum model to train the multi-component substance fluorescence spectrum model.

8. A device for fluorescence spectrum analysis of multi-component substances, characterized in that: include: An acquisition module, used to acquire a three-dimensional fluorescence fingerprint of a multi-component substance to be tested; A detection module, used for inputting the three-dimensional fluorescence fingerprint of the multi-component substance into a pre-trained fluorescence spectrum model of the multi-component substance to obtain detection information of the multi-component substance to be tested; wherein the detection information includes the number of fluorescent components, the concentration of each fluorescent component and the three-dimensional fluorescence fingerprint corresponding to each fluorescent component; Among them, the multi-component substance fluorescence spectrum model includes a multi-task prediction network and a component fitting network. The multi-task prediction network is used to simultaneously predict the number of fluorescent components of the multi-component substance to be tested and the concentration of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance; the component fitting network is used to output the three-dimensional fluorescence fingerprint spectrum of each fluorescent component based on the three-dimensional fluorescence fingerprint spectrum of the multi-component substance and the prediction results of the number of the fluorescent components.

9. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images