Method for detecting aflatoxin in vegetable oil by fluorescent probe method

Abstract

A method for detecting aflatoxin in vegetable oil by a fluorescent probe method comprises; acquiring a time series of fluorescent intensities, segmenting the series according to a time scale to obtain a set of local time series of fluorescence intensities; arranging a local time series of each fluorescence intensity in a time dimension to obtain a set of local timing input vectors of the fluorescence intensities; separately performing feature extraction on a local timing input vector by a deep neural network based fluorescence intensity timing association feature extractor; performing feature extraction to obtain typical timing pattern features of the fluorescence intensities; and determining a concentration value of the aflatoxin based on the timing pattern of the fluorescence intensities.

Description

TECHNICAL FIELD The present application relates to the field of intelligent detection, and more particularly to a method for detecting aflatoxin in vegetable oil by a fluorescent probe method. BACKGROUND In the field of food safety, aflatoxin is a toxin produced by Aspergillus flavus, commonly found in foods such as vegetable oil. Aflatoxins have strong carcinogenicity and mutagenicity, and will cause serious harm to human health. Therefore, detection of the aflatoxin content in the vegetable oil is critical for food safety. However, traditional methods for detecting aflatoxin in vegetable oil typically employ high performance liquid chromatography (HPLC) and enzyme-linked immunosorbent assay (ELISA), and although these methods have high sensitivity and selectivity, they are complicated to operate, require professional operators to operate, and require a long analysis time, which cannot meet the needs of rapid detection, especially having low efficiency in the detection of large quantities of samples. In addition, conventional aflatoxin detection methods require expensive instruments and equipment and reagents, which makes the detection cost high and is not conducive to large-scale application and popularization. Therefore, a solution for detecting aflatoxin in vegetable oil by a fluorescent probe method is desired. SUMMARY In order to solve the above technical problems, the present application is proposed. An embodiment of the present application provides a method for detecting aflatoxin in vegetable oil by a fluorescent probe method, and the method performs time series analysis of time series data of fluorescence intensities by introducing a data processing and analysis algorithm at a back end, so as to capture a time series pattern and a change trend of the fluorescence intensities, so as to infer a concentration value of the aflatoxin based on the time series change characteristics of the fluorescence brightness, so as to detect the concentration value of the aflatoxin. According to one aspect of the present application, provided is a method for detecting aflatoxin in vegetable oil by a fluorescent probe method, including: acquiring a time series of fluorescence intensities; segmenting the time series of the fluorescence intensities according to a predetermined time scale to obtain a set of local time series of the fluorescence intensities; arranging a local time series of each fluorescence intensity in the set of the local time series of the fluorescence intensities in a time dimension to obtain a set of local timing input vectors of the fluorescence intensities; separately performing feature extraction on a local timing input vector of each fluorescence intensity in the set of the local timing input vectors of the fluorescence intensities by a deep neural network-based fluorescence intensity timing association feature extractor to obtain a set of local timing association feature vectors of the fluorescence intensities; performing typical feature extraction on the set of the local timing association feature vectors of the fluorescence intensities to obtain typical timing pattern features of the fluorescence intensities; and determining a concentration value of the aflatoxin based on the typical timing pattern features of the fluorescence intensities. Compared with the prior art, the method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to the present application performs time series analysis of time series data of fluorescence intensities by introducing a data processing and analysis algorithm at a back end, so as to capture a time series pattern and a change trend of the fluorescence intensities, so as to infer a concentration value of the aflatoxin based on the time series change characteristics of the fluorescence brightness, so as to detect the concentration value of the aflatoxin. BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects, features and advantages of the present application will become more apparent through the detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the description, and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps. FIG. 1 is a flow chart of a method for detecting aflatoxin in vegetable oil by a fluorescent probe method according to an embodiment of the present application; FIG. 2 is a system architecture diagram of the method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to the embodiment of the present application; FIG. 3 is a flow chart of a sub-step S5 of the method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to the embodiment of the present application; and FIG. 4 is a flow chart of a sub-step S51 of the method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to the embodiment of the present application. DETAILED DESCRIPTION Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments of the present application, and it should be understood that the present application is not limited by the exemplary embodiments described herein. As used in the present application and the claims, the words "a", "an", "an" and / or "the" do not refer specifically to the singular, and may also include the plural unless the context expressly indicates exceptions. In general, the terms "include" and "comprise" only indicate the inclusion of steps and elements which have been explicitly identified, and these steps and elements do not constitute an exclusive list, and a method or device may also contain other steps or elements. Although the present application makes various references to certain modules in a system according to the embodiments of the present application, any number of different modules may be used and run on a user terminal and / or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules. Flowcharts are used in the present application to illustrate operations performed by the system according to the embodiments of the present application. It should be understood that the foregoing or following operations are not necessarily performed precisely in order. On the contrary, various steps may be processed in a reverse order or simultaneously as needed. At the same time, other operations may be added to the processes, or a certain operation or operations may be removed from the processes. Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments of the present application, and it should be understood that the present application is not limited by the exemplary embodiments described herein. Traditional methods for detecting aflatoxin in vegetable oil typically employ high performance liquid chromatography (HPLC) and enzyme-linked immunosorbent assay (ELISA), and although these methods have high sensitivity and selectivity, they are complicated to operate, require professional operators to operate, and require a long analysis time, which cannot meet the needs of rapid detection, especially having low efficiency in the detection of large quantities of samples. In addition, conventional aflatoxin detection methods require expensive instruments and equipment and reagents, which makes the detection cost high and is not conducive to large-scale application and popularization. In view of the above technical problems, in the technical solution of the present application, a method for detecting aflatoxin in vegetable oil by a fluorescent probe method is proposed, which enables the preparation of carbon dots by using orange peel as a raw material, and constructs a method for accurate and rapid detection of aflatoxin Bl (AFB1) based on a time series change of the fluorescence intensity by using aptamer-modified carbon dots (DNA-CDs) as a fluorescent probe in combination with graphene oxide as a quencher. This method utilizes carbohydrates abundant in the orange peel, including resources such as fructose, glucose, sucrose and cellulose to prepare carbon dots with high oxygen functional groups by a hydrothermal method, which not only improves the performance of the probe, but also realizes the recycling of waste. In particular, in the detection process of aflatoxin Bl, first, an aqueous graphene oxide solution is added into a dispersed solution of DNA-CDs, and then an AFB1 standard solution is added into the above solution system while recording a time series of fluorescence intensities. A data processing and analysis algorithm is then introduced at a back end to perform time series analysis of time series data of these fluorescence intensities, so as to capture a time series pattern and a change trend of the fluorescence intensities, so as to infer a concentration value of the aflatoxin based on the time series change characteristics of the fluorescence brightness, so as to detect the concentration value of the aflatoxin. A principle of this method is that carbon dots (CDs) with carboxyl are first modified with an aptamer composed of DNA with amino to achieve functionalization of the carbon dots. After graphene oxide is added, the aptamer-modified carbon dots will interact with graphene oxide, resulting in efficient quenching of the fluorescence of the carbon dots. When AFB1 is present or introduced in a system, the aptamer on graphene oxide will recognize a target and react with AFB 1. forming a well-folded AFBl-aptamer carbon dot complex. At this time, the aptamer-modified carbon dots are no longer attached to graphene oxide, but desorbed from graphene oxide, causing the fluorescence of the aptamer-modified carbon dots to be restored. In this way, the concentration of AFB1 can be detected by recording time series data of the fluorescence intensities and analyzing the change characteristics based on the principle described above. In the technical solution of the present application, a method for detecting aflatoxin in vegetable oil by a fluorescent probe method is proposed. FIG. 1 is a flow chart of a method for detecting aflatoxin in vegetable oil by a fluorescent probe method according to an embodiment of the present application. FIG. 2 is a system architecture diagram of the method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to the embodiment of the present application. As shown in FIG. 1 and FIG. 2, the method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to the embodiment of the present application includes the steps of: SI, acquiring a time series of fluorescence intensities; S2, segmenting the time series of the fluorescence intensities according to a predetermined time scale to obtain a set of local time series of the fluorescence intensities; S3, arranging a local time series of each fluorescence intensity in the set of the local time series of the fluorescence intensities in a time dimension to obtain a set of local timing input vectors of the fluorescence intensities; S4, separately performing feature extraction on a local timing input vector of each fluorescence intensity in the set of the local timing input vectors of the fluorescence intensities by a deep neural network-based fluorescence intensity timing association feature extractor to obtain a set of local timing association feature vectors of the fluorescence intensities; S5, performing typical feature extraction on the set of the local timing association feature vectors of the fluorescence intensities to obtain typical timing pattern features of the fluorescence intensities; and S6, determining a concentration value of the aflatoxin based on the typical timing pattern features of the fluorescence intensities. In particular, in the SI and the S2, the time series of the fluorescence intensities is acquired; and the time series of the fluorescence intensities is segmented according to the predetermined time scale to obtain the set of the local time series of the fluorescence intensities. The fluorescence intensity will have different intensity changes when different solutions are added, that is, the fluorescence intensity will change continuously over time in a time dimension, and a data volume may be relatively large during monitoring of fluorescence intensity data, small timing fluctuation information of the fluorescence intensity can be easily ignored. Therefore, in order to be able to more fully and accurately perform time series feature analysis of the fluorescence intensity, so as to infer a concentration value of aflatoxin, in the technical solution of the present application, it is necessary to segment the time series of the fluorescence intensities according to the predetermined time scale to obtain the set of the local time series of the fluorescence intensities, and arrange the local time series of each fluorescence intensity in the set of the local time series of the fluorescence intensities in the time dimension to obtain the set of the local timing input vectors of the fluorescence intensities. By segmenting the time series of the fluorescence intensities, the entire time series of the fluorescence intensities can be segmented into a plurality of small time periods for subsequent more detailed analysis of a time series change and pattern of the fluorescence intensities within local time periods. In particular, in the S3, the local time series of each fluorescence intensity in the set of the local time series of the fluorescence intensities is arranged in the time dimension to obtain the set of the local timing input vectors of the fluorescence intensities. It should be understood that by arranging in the time dimension, time series distribution information of the fluorescence intensity in each local time period can be retained, providing a data basis for subsequent time series feature extraction and change analysis of the fluorescence intensity. In particular, in the S4, feature extraction is separately performed on the local timing input vector of each fluorescence intensity in the set of the local timing input vectors of the fluorescence intensities by the deep neural network-based fluorescence intensity timing association feature extractor to obtain the set of the local timing association feature vectors of the fluorescence intensities. In particular, in one specific example of the present application, the deep neural network-based fluorescence intensity timing association feature extractor is a one-dimensional convolutional layer-based fluorescence intensity timing association feature extractor. In the technical solution of the present application, in order to be able to explore a local time series pattern and a change trend of the fluorescence intensities in each local time period, thereby performing efficient expression of local time series features of the fluorescence intensities, in the technical solution of the present application, feature mining is further performed on a local timing input vector of each fluorescence intensity in the set of the local timing input vectors of the fluorescence intensities by the one-dimensional convolutional layer-based fluorescence intensity timing association feature extractor, so as to separately extract local timing dynamic feature information of the fluorescence intensities in each local time period, thereby obtaining the set of the local timing association feature vectors of the fluorescence intensities. It is worth noting that a one-dimensional convolutional layer is a commonly used layer in deep learning for processing data with time series association. In the one-dimensional convolutional layer, convolution operations are conducted along one dimension (usually the time dimension) of input data to extract the time series features. In particular, in the S5, typical feature extraction is performed on the set of the local timing association feature vectors of the fluorescence intensities to obtain the typical timing pattern features of the fluorescence intensities. In particular, in a specific example of the present application, as shown in FIG. 3, the S5 includes: S51, feature optimization is performed on the set of the local timing association feature vectors of the fluorescence intensities to obtain a set of optimized local timing association feature vectors of the fluorescence intensities; and S52, the set of the optimized local timing association feature vectors of the fluorescence intensities is allowed to pass through a typical feature extraction network to obtain typical timing pattern feature vectors of the fluorescence intensities as the typical timing pattern features of the fluorescence intensities. In particular, in the S51, feature optimization is performed on the set of the local timing association feature vectors of the fluorescence intensities to obtain the set of the optimized local timing association feature vectors of the fluorescence intensities. In particular, in a specific example of the present application, as shown in FIG. 4, the S51 includes: S511, a weighting coefficient for a local timing association feature vector of each fluorescence intensity in the set of the local timing association feature vectors of the fluorescence intensities is separately calculated to obtain a weighting coefficient vector consisting of a plurality of weighting coefficients; and S512, weighted optimization is performed on the set of the local timing association feature vectors of the fluorescence intensities with each weighting coefficient in the weighting coefficient vector as a weighting factor to obtain the set of the optimized local timing association feature vectors of the fluorescence intensities. More specifically, in the S511, the weighting coefficient for the local timing association feature vector of each fluorescence intensity in the set of the local timing association feature vectors of the fluorescence intensities is separately calculated to obtain the weighting coefficient vector consisting of the plurality of the weighting coefficients. It should be understood that in the technical solution described above, the local timing association feature vector of each fluorescence intensity in the set of the local timing association feature vectors of the fluorescence intensities expresses a timing association feature of the fluorescence intensity in a local time domain. Therefore, when the set of the local timing association feature vectors of the fluorescence intensities is subjected to fluorescence intensity timing typical pattern feature extraction by the typical feature extraction network, the typical timing pattern feature vectors of the fluorescence intensities are expected to improve the decoding probability certainty of each of the local timing association feature vectors of the fluorescence intensities to a global timedomain timing feature domain under local time-domain distributive discrimination, while maintaining timing association feature profile discriminability of the local timing association feature vectors of the each fluorescence intensities with respect to the local time domain, thereby improving the effect of global time-domain timing feature expression of the typical timing pattern feature vectors of the fluorescence intensities on a predetermined decoded value of a decoder-based aflatoxin concentration inference device. Based on this, the applicant of the present application separately calculates the weighting coefficient for the local timing association feature vector of each fluorescence intensity in the set of the local timing association feature vectors of the fluorescence intensities, expressed as: L 1 V-1 2 = exp[log(yt_max)] + ax J 7=1 wherein Vj_max and Vj_j are a maximum feature value and a jth feature value of a local timing correlation feature vector Vj of an ith fluorescence intensity in the set of the local timing association feature vectors of the fluorescence intensities, respectively, Lisa length of a feature vector, log represents a logarithm with a base of 2, and a is a weight hyperparameter. Specifically, a majority voting mechanism under a crowdsourcing form of timing features distributed by local time domain timing associations of the local timing association feature vectors Vj of the fluorescence intensities is respectively used to seek individual information expectation maximization of the local timing association feature vectors of the fluorescence intensities under a distribution dependent probability model with respect to a global time domain decoding probability space, thus, the local timing association feature vectors Vj of the fluorescence intensities are then weighted by a factor co. to optimize the local timing association feature vectors Vj of the fluorescence intensities, a joint estimation of a distribution with respect to the global time domain decoding probability space can be performed based on a local time domain average information confidence of respective feature values of the local timing association feature vectors Vj of the fluorescence intensities, to improve its decoding probability certainty with respect to a decoding probability space while maintaining timing feature distribution differentiability of the local timing association feature vectors Vj of the fluorescence intensities, thereby improving the effect of global time-domain timing feature expression of the typical timing pattern feature vectors of the fluorescence intensities on the predetermined decoded value of the decoder-based aflatoxin concentration inference device, improving the accuracy of the inference result. In this way, the detection of the concentration value of the aflatoxin in the vegetable oil can be performed more accurately based on the time series change characteristics of the fluorescence brightness. More specifically, in the S512, weighted optimization is performed on the set of the local timing association feature vectors of the fluorescence intensities with each weighting coefficient in the weighting coefficient vector as the weighting factor to obtain the set of the optimized local timing association feature vectors of the fluorescence intensities. That is, each weighting coefficient in the weighting coefficient vector is applied to the set of the local timing association feature vectors of the fluorescence intensities as the weighting factor to achieve the effect of weighted optimization. It is worth mentioning that in other embodiments of the present application, feature optimization can also be performed on the set of the local timing association feature vectors of the fluorescence intensities in other ways to obtain the set of the optimized local timing association feature vectors of the fluorescence intensities, such as: inputting the set of the local timing correlation feature vectors of the fluorescence intensities; transforming the feature vectors to extract more informative features; applying a smoothing technique (e.g., moving average, Gaussian filtering, etc.) to smooth the feature vectors; normalizing the feature vectors to ensure that the numerical ranges of different features are consistent and avoid the model from being affected by the difference in the range of feature values, so as to obtain the set of the optimized local timing association feature vectors of the fluorescence intensities. Specifically, in the S52, the set of the optimized local timing association feature vectors of the fluorescence intensities is allowed to pass through the typical feature extraction network to obtain the typical timing pattern feature vectors of the fluorescence intensities as the typical timing pattern features of the fluorescence intensities. It should be understood that since a local timing association feature vector of each fluorescence intensity in the set of the local timing association feature vectors of the fluorescence intensities represents a local timing feature of the fluorescence intensity in each local time period, in order to be able to enhance the expression ability of the local timing features of the fluorescence intensities in the local time period after the addition of the aqueous graphene oxide solution and the AFB1 standard solution, thereby highlighting local timing feature semantics of the fluorescence intensities related to the detection of aflatoxin concentration, and in the technical solution of the present application, the set of the local timing association feature vectors of the fluorescence intensities is further allowed to pass through the typical feature extraction network to obtain the typical timing pattern feature vectors of the fluorescence intensities. It should be understood that the typical feature extraction network is capable of learning and capturing the association relationship and interaction between the local timing association feature vectors of the fluorescence intensities to identify and highlight and display the local timing association features of the fluorescence intensities of key nodes in the set of the local timing association feature vectors of the fluorescence intensities. In this way, it is possible to capture and highlight typical patterns and variation features between local timing features of the fluorescence intensities within each local time period, and capture the association relationship and interaction between the local timing characteristics of the fluorescence intensities, helping to better describe a time series pattern and a change trend of the fluorescence intensities, thereby facilitating more accurate inference of the concentration value of the aflatoxin based on the timing change characteristics of the fluorescence intensities. More specifically, allowing the set of the optimized local timing association feature vectors of the fluorescence intensities to pass through the typical feature extraction network to obtain the typical timing pattern feature vectors of the fluorescence intensities as the typical timing pattern features of the fluorescence intensities includes processing the set of the optimized local timing association feature vectors of the fluorescence intensities through the typical feature extraction network by using the following typical feature extraction formula to obtain the typical timing pattern feature vectors of the fluorescence intensities, wherein the typical feature extraction formula is as follows: L Dt=^ ^Jv^log O / x))] (vi,Vj)EVk x=l y- exp(-Dt') V = >----------Vi wherein and Vj^ are feature values for each position of ith and jth optimized local timing association feature vectors of the fluorescence intensities in the set of the optimized local timing association feature vectors of the fluorescence intensities, respectively, V[ and Vj are ith and Jth optimized local timing association feature vectors of the fluorescence intensities in the set of the optimized local timing association feature vectors of the fluorescence intensities, respectively, Vk is the set of the optimized local timing association feature vectors of the fluorescence intensities, log represents a logarithmic function value with a base of 2, L is a length of each of the optimized local timing association feature vectors of the fluorescence intensities, M is the number of vectors in the set of the optimized local timing association feature vectors of the fluorescence intensities minus one, Dt is a feature value for each position in a local timing semantic fluctuation feature vector of the fluorescence intensities, Ns is a length of the local timing semantic fluctuation feature vector of the fluorescence intensities, expQ) is an exponential operation, and V is atypical timing pattern feature vector of each fluorescence intensity. It is worth mentioning that in other specific examples of the present application, typical feature extraction can also be performed on the set of the local timing association feature vectors of the fluorescence intensities in other ways to obtain the typical timing pattern features of the fluorescence intensities, such as: inputting the set of the local timing association feature vectors of the fluorescence intensities; performing a dimensionality reduction process on the local timing association feature vectors of the fluorescence intensities; clustering the dimensionality-reduced feature vectors by using a clustering algorithm, classifying similar feature vectors into the same category; analyzing feature vectors in each cluster to find representative patterns or center points representing typical features in the cluster to obtain the typical timing pattern features of the fluorescence intensities. In particular, in the S6, the concentration value of the aflatoxin is determined based on the typical timing pattern features of the fluorescence intensities. In particular, in one specific example of the present application, the typical timing pattern feature vectors of the fluorescence intensities are allowed to pass through a decoder-based aflatoxin concentration inference device to obtain an inference result, the inference result being the concentration value of the aflatoxin. That is, decoding regression is performed by using the typical timing pattern features of the fluorescence intensities, so that a concentration value of the aflatoxin is inferred based on the time series change characteristics of the fluorescence brightness, so as to detect the concentration value of the aflatoxin in the vegetable oil by the fluorescent probe method. In particular, allowing the typical timing pattern feature vectors of the fluorescence intensities to pass through the decoder-based aflatoxin concentration inference device to obtain the inference result, the inference result being the concentration value of the aflatoxin includes performing decoding regression on the typical timing pattern feature vectors of the fluorescence intensities with the following formula by using the decoder to obtain a decoded value representing the concentration value of the aflatoxin; wherein the formula is: Y = £ W 0 X, wherein X represents a typical timing pattern feature vector of each fluorescence intensity, Y is the decoded value, W is a weight matrix, and 0 represents matrix multiplication. In summary, the method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to the embodiments of the present application is elucidated, and performs time series analysis of time series data of fluorescence intensities by introducing a data processing and analysis algorithm at a back end, so as to capture a time series pattern and a change trend of the fluorescence intensities, so as to infer a concentration value of the aflatoxin based on the time series change characteristics of the fluorescence brightness, so as to detect the concentration value of the aflatoxin. In the technical solution of the present application, preparation of carbon dots is carried out by using orange peel as a raw material, and by using aptamer-modified carbon dots (DNA-CDs) as a fluorescent probe, and graphene oxide as a quencher, a new method for accurately and rapidly detecting aflatoxin Bl (AFB1) is constructed. The orange peel, as a common household waste, contains a large amount of carbohydrates, including fructose, glucose, sucrose, cellulose, and the like. The richness of these raw materials will result in the formation of a large number of oxygen functional groups during the preparation of the carbon dots by a hydrothermal method, thereby improving the performance of the carbon dots, and the orange peel has a wide range of sources and low cost, and waste utilization can be achieved. 1. Preparation of carbon dots from orange peel Waste orange peel was collected and washed in water and dried in sunlight. Then, 2 g of pretreated orange peel was washed in 100 mL of a 0.1 M aqueous H2SO4 solution, and rinsed with water, filtered and dried in an oven at 150°C for 2 h. The treated orange peel was soaked in a sodium hypochlorite solution for 4 h for oxidation treatment, and then rinsed in water for multiple times until a pH of washing water reached 7. 1 g of urea was weighed to be dissolved into 25 ml of ultrapure water, and the obtained solution was sonicated for 15 min. The oxidized orange peel was placed in the above solution to be placed in a Teflon-lined autoclave, a reaction was carried out at 230°C for 4 h, the autoclave was allowed to cool naturally, and the obtained brown solution was washed with dichloromethane to remove an unreacted organic portion. The resulting carbon dot solution was centrifuged at 4000 rpm for 15 min and filtered through a 0.22 pm filter membrane to remove residues to obtain a brown carbon dot solution which was kept in a dark place at 4°C. The reaction temperature and time were optimized during the preparation of the carbon dots, and the preparation was carried out at 180°C, 190°C, 200°C, 210°C, 220°C, and 230°C for 4h, 6h, 8h, and lOh, respectively, and finally the carbon dots obtained at 230°C for 4h were found to have the highest quantum yield. 2. Aptamer modification First, 30 uL of EDC (1 mg / uL) powder and 30 uL of NHS (0.25 mg / uL) particles were weighed into a beaker, 1 mL of a PBS buffer solution with a pH of 7.4 was added, and 100 pL of the prepared solution of CDs was added, and the mixed solution was ultrasonically stirred for 30 min until the solution was uniformly mixed. Next, 20 pL of an AFB1 aptamer (10 pM) was added to 10 pL of the mixed solution, and sonicating was performed for 2 h. To ensure successful modification of the aptamer onto the carbon dots. 3. Selection of quenchers Quenching experiments were performed on the resulting carbon dot material with five different quenchers of equal concentrations (1 mg / ml) of a solution of graphene oxide in ethanol, a solution of graphene oxide in water, a gold nanocolloid, and a solution of tungsten disulfide nanoflakes (WS2) in ethanol. The results showed the quenching efficiencies were 88.79%, 94.51%, 69.46%, and 45.19%, respectively. The aqueous graphene oxide solution with the highest quenching efficiency was therefore selected for subsequent experiments. 4. Detection of aflatoxin Bl First, a dispersed solution of DNA-CDs was added to an aqueous graphene oxide solution, and the fluorescence intensity before the addition of graphene oxide and 3 minutes after the addition of graphene oxide was accurately recorded. Different concentrations of AFB1 standard solutions were added to the above solution system, and the fluorescence intensity was recorded after incubation at room temperature for 30 minutes. First, carbon dots (CDs) with carboxyl are modified with an aptamer composed of DNA with amino to functionalize the carbon dots. At this time, when graphene oxide is added, the aptamer-modified carbon dots interact with graphene oxide due to u-u stacking, causing the aptamer-modified carbon dots being adsorbed on the surface of graphene oxide, thereby effectively quenching the fluorescence of the carbon dots; and when AFB1 is present in the system or after AFB1 is introduced into the system, the adapter attached to graphene oxide recognizes a target and reacts with AFB1 to form a well-folded AFB1 -aptamer carbon dot complex. Bases participating in the tt-tt stacking interaction have been occupied by AFB1, so the aptamer-modified carbon dots cannot be attached to graphene oxide any more, but will be desorbed from graphene oxide, thereby restoring the fluorescence of the aptamer-modified carbon dots. This part of experiments detected AFB 1 based on the principles described above. 5. Detection of AFB 1 in vegetable oil 50 pL of aflatoxin Bl solutions of different concentrations were added to 200 pL of a household peanut oil sample, then a pH was adjusted to be neutral, an aqueous graphene oxide solution was added to a dispersed solution of DNA-CDs, and the fluorescence intensity before the addition of graphene oxide and 3 minutes after the addition of graphene oxide was accurately recorded. The above mixed solution and a 0.01 M phosphate buffer (PB, pH 7.0) containing different concentrations of aflatoxin Bl were then incubated for 30 minutes at room temperature, and the fluorescence intensity was recorded. The above experiment was repeated for five times, and a relative standard deviation was calculated. 6. Selectivity and stability of a fluorescence sensor To evaluate the selectivity of an orange peel-aptamer functionalized carbon dot system for AFB 1 detection, the effects of possible interferents including aflatoxin B2 (AFB2), aflatoxin Gl (AFG1), aflatoxin G2 (AFG2), ochratoxin A (OTA), expansin (PTL), deoxynivalenol (DON), zearalenone (ZEN), and the like on the system were selected to be determined. The carbon dot solution was stored at 4°C in the dark and its fluorescence intensity was recorded and detected after 15 days, 30 days and 45 days to determine the stability of the sensor. The reproducibility of this sensor was investigated by preparing five batches of aptamer-modified carbon dots and preparing AFB 1 sensors with different batches of carbon dot solutions, respectively. 7. Calculation of system fluorescence recovery by AFB1 A calculation equation for the system fluorescence recovery (1%) by AFB1 is shown below, wherein Fo is the fluorescence intensity in the carbon dot solution, Fhas represents the fluorescence intensity three minutes after the addition of the aqueous graphene oxide solution, and Fhas / afbi represents the fluorescence intensity after incubation by adding the graphene oxide solution and the aflatoxin Bl solution. / % = (FHAs / AFBl - FHAs) / (F0 - FHAs) The embodiments of the present application have been described above and the above description is exemplary not exhaustive and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terms used herein is selected to best explain the principles of the embodiments, the practical application or improvements to the technology in the market, or to enable those of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for detecting aflatoxin in vegetable oil by a fluorescence probe method, characterized by comprising:acquiring a time series of fluorescence intensities;segmenting the time series of the fluorescence intensities according to a predetermined time scale to obtain a set of local time series of the fluorescence intensities;arranging a local time series of each fluorescence intensity in the set of the local time series of the fluorescence intensities in a time dimension to obtain a set of local timing input vectors of the fluorescence intensities;separately performing feature extraction on a local timing input vector of each fluorescence intensity in the set of the local timing input vectors of the fluorescence intensities by a deep neural network-based fluorescence intensity timing association feature extractor to obtain a set of local timing association feature vectors of the fluorescence intensities;performing typical feature extraction on the set of the local timing association feature vectors of the fluorescence intensities to obtain typical timing pattern features of the fluorescence intensities; anddetermining a concentration value of the aflatoxin based on the typical timing pattern features of the fluorescence intensities.

2. The method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to claim 1, characterized in that the deep neural network-based fluorescence intensity timing association feature extractor is a one-dimensional convolutional layer-based fluorescence intensity timing association feature extractor.

3. The method for detecting the aflatoxin in the vegetable oil by the fluorescent probemethod according to claim 2, characterized in that separately performing feature extraction on the local timing input vector of each fluorescence intensity in the set of the local timing input vectors of the fluorescence intensities by the deep neural networkbased fluorescence intensity timing association feature extractor to obtain the set of the local timing association feature vectors of the fluorescence intensities comprises separately allowing the local timing input vector of each fluorescence intensity in the set of the local timing input vectors of the fluorescence intensities to pass through the one-dimensional convolutional layer-based fluorescence intensity timing association feature extractor to obtain the set of the local timing association feature vectors of the fluorescence intensities.

4. The method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to claim 3, characterized in that performing typical feature extraction on the set of the local timing association feature vectors of the fluorescence intensities to obtain the typical timing pattern features of the fluorescence intensities comprises:performing feature optimization on the set of the local timing association feature vectors of the fluorescence intensities to obtain a set of optimized local timing association feature vectors of the fluorescence intensities; andallowing the set of the optimized local timing association feature vectors of the fluorescence intensities to pass through a typical feature extraction network to obtain typical timing pattern feature vectors of the fluorescence intensities as the typical timing pattern features of the fluorescence intensities.

5. The method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to claim 4, characterized in that performing feature optimization on the set of the local timing association feature vectors of the fluorescence intensities to obtain the set of the optimized local timing association feature vectors of thefluorescence intensities comprises:separately calculating a weighting coefficient for a local timing association feature vector of each fluorescence intensity in the set of the local timing association feature vectors of the fluorescence intensities to obtain a weighting coefficient vector consisting of a plurality of weighting coefficients; andperforming weighted optimization on the set of the local timing association feature vectors of the fluorescence intensities with each weighting coefficient in the weighting coefficient vector as a weighting factor to obtain the set of the optimized local timing association feature vectors of the fluorescence intensities.

6. The method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to claim 5, characterized in that allowing the set of the optimized local timing association feature vectors of the fluorescence intensities to pass through the typical feature extraction network to obtain the typical timing pattern feature vectors of the fluorescence intensities as the typical timing pattern features of the fluorescence intensities comprises processing the set of the optimized local timing association feature vectors of the fluorescence intensities through the typical feature extraction network by using the following typical feature extraction formula to obtain the typical timing pattern feature vectors of the fluorescence intensities,wherein the typical feature extraction formula is as follows:L(vi,vj)ev& x=lV-1 exp(—D,)v= Z(vi,VJ)EvkM=lexp^wherein and Vj^ are feature values for each position of ith and jth optimized local timing association feature vectors of the fluorescence intensities in the set of the optimized local timing association feature vectors of the fluorescence intensities, respectively, Vt and Vj are ith and jth optimized local timing associationfeature vectors of the fluorescence intensities in the set of the optimized local timing association feature vectors of the fluorescence intensities, respectively, Vk is the set of the optimized local timing association feature vectors of the fluorescence intensities, log represents a logarithmic function value with a base of 2, L is a length of each of the optimized local timing association feature vectors of the fluorescence intensities, M is the number of vectors in the set of the optimized local timing association feature vectors of the fluorescence intensities minus one, Dt is a feature value for each position in a local timing semantic fluctuation feature vector of the fluorescence intensities, Ns is a length of the local timing semantic fluctuation feature vector of the fluorescence intensities, expQ) is an exponential operation, and V is a typical timing pattern feature vector of each fluorescence intensity.

7. The method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to claim 6, characterized in that determining the concentration value of the aflatoxin based on the typical timing pattern features of the fluorescence intensities comprises allowing the typical timing pattern feature vectors of the fluorescence intensities to pass through a decoder-based aflatoxin concentration inference device to obtain an inference result, the inference result being the concentration value of the aflatoxin.

8. The method for detecting the aflatoxin in the vegetable oil by the fluorescent probe method according to claim 7, characterized in that allowing the typical timing pattern feature vectors of the fluorescence intensities to pass through the decoder-based aflatoxin concentration inference device to obtain the inference result, the inference result being the concentration value of the aflatoxin comprises performing decoding regression on the typical timing pattern feature vectors of the fluorescence intensities with the following formula by using the decoder to obtain a decoded value representingthe concentration value of the aflatoxin,wherein the formula is: Y = ® X, wherein X represents a typical timingpattern feature vector of each fluorescence intensity, Y is the decoded value, PF is a weight matrix, and ® represents matrix multiplication.

Images