CN113672662A

CN113672662A - Food multi-source information fusion method and device and storage medium

Info

Publication number: CN113672662A
Application number: CN202110895244.0A
Authority: CN
Inventors: 赵思明; 陈燕雨; 熊善柏; 张宾佳
Original assignee: Huazhong Agricultural University
Current assignee: Huazhong Agricultural University
Priority date: 2021-08-04
Filing date: 2021-08-04
Publication date: 2021-11-19

Abstract

The invention relates to a food multi-source information fusion method, a device and a storage medium, wherein the method comprises the following steps: acquiring various food information of food types, and establishing a standard database of the food types; establishing a contribution type model and/or a discriminant type model according to data in a standard database; and collecting food indexes of the food to be detected, and determining the quality index based on the contribution model and the food indexes and/or identifying whether the food belongs to a preset category based on the discriminant model and the food indexes. The method adopts multi-source information fusion, is favorable for digging out hidden information and deeply and finely describing objects in multiple dimensions, thereby increasing the reliability, the accuracy and the comprehensiveness of the aspects of food identification, attribute judgment and the like, increasing the credibility of data or information, establishing a contribution type model, accurately and quickly determining the quality index of the food to be detected, establishing a discriminant type model, efficiently judging whether the food to be detected belongs to a preset variety or not, and ensuring efficient, accurate and comprehensive monitoring of the food.

Description

Food multi-source information fusion method and device and storage medium

Technical Field

The invention relates to the technical field of food information, in particular to a food multi-source information fusion method, a device and a storage medium.

Background

Multi-source Information Fusion (MIF) is a technique that takes Information obtained in multiple ways, at any time and in space as an overall comprehensive analysis, integrates Information collected by multiple channels, eliminates redundant Information among different Information, and forms a relatively complete description process for overall characteristics. The multi-source information fusion technology utilizes computer analysis, integrates the redundant, competitive, complementary and cooperative information of the whole system in time and space with the optimal energy efficiency ratio according to the relevant criteria of domain knowledge, reasonably utilizes the different attribute information of the whole system, improves the reliability of the system and the efficiency of solving problems compared with the subsets of all components of the system, is favorable for digging out implicit information, increases the deep understanding of the essence of the object, can obtain the information of the target to be measured to a greater extent through the deep and fine multi-dimensional description of the object, thereby increasing the reliability and the accuracy of the aspects of food identification, attribute judgment and the like, and increasing the credibility of data or information.

The earliest application field of the multi-source information fusion technology is multi-sensor data fusion, and with the development of research fields of various subjects, the definition of a sensor in the information fusion technology extends to information of various sources, such as images, temperature, hyperspectrum, electric quantity and the like. At present, the multi-source information fusion technology has been applied to the fields of industrial monitoring systems, fault diagnosis systems and artificial intelligence, such as equipment fault diagnosis, safety detection, image fusion and the like. In the field of rice identification, a rice production place identification model established by fusing rice infrared spectrum fingerprint information and volatile component information obtains a better effect. In the field of food analysis, multi-source information fusion technology has been used for nondestructive testing of potatoes, freshness detection of meat, cold chain transportation of agricultural products and the like. At present, the application of a food characteristic portrait technology established by multi-source information fusion in food quality classification, characteristic identification (including brands and production places), food safety risk assessment and food safety hazard identification is very important. Therefore, how to utilize multi-source information to perform efficient monitoring of food information is an urgent problem to be solved.

Disclosure of Invention

In view of the above, there is a need to provide a food multi-source information fusion method, device and storage medium, so as to overcome the problem in the prior art that the food safety monitoring process does not effectively combine multi-source information.

The invention provides a food multi-source information fusion method, which comprises the following steps:

acquiring multiple kinds of food information of at least one kind of food, and establishing a standard database of corresponding food kinds according to the multiple kinds of food information;

establishing a contribution type model and/or a discriminant type model according to the data in the standard database;

collecting food indexes of food to be detected, determining the quality index of the food to be detected based on the contribution model and the food indexes and/or identifying whether the food to be detected belongs to a preset category or not based on the discriminant model and the food indexes.

Further, the establishing a standard database of corresponding food types according to the information of the plurality of types of food comprises:

classifying the various food information to determine structured data and unstructured data;

carrying out data cleaning and data completion on the structured data to form numerical information;

carrying out structuring processing on the unstructured data and extracting corresponding characteristic information;

and taking the numerical information and the characteristic information as database entry data, and establishing the corresponding standard database.

Further, the database entry data comprises a plurality of characteristic parameters, and the establishment process of the contribution type model comprises the following steps:

determining a contribution index of each characteristic parameter to the food index by adopting a principal component analysis method and a principal component regression method according to a plurality of characteristic parameters and the food index corresponding to at least one food type;

and determining the contribution type model according to the plurality of characteristic parameters and the corresponding contribution indexes.

Further, the food index includes at least one of a nutrition index, a flavor index, a food safety index, a place of origin index, a breed index, and a brand index, and the contribution model is represented by the following formula:

wherein f (x) represents a model function of the contribution model, the corresponding function value is the food index, a₀Denotes a predetermined coefficient, x_nRepresenting the n-th said characteristic parameter, a_nAnd representing the contribution coefficient corresponding to the nth characteristic parameter.

Further, the database entry data includes a plurality of characteristic parameters, and the establishment process of the discriminant model includes:

determining a contribution index of each characteristic parameter to the food index according to a plurality of characteristic parameters and the food index corresponding to the at least one food category;

determining a corresponding discrimination function value according to the plurality of characteristic parameters and the corresponding contribution indexes;

and calculating the corresponding discrimination function value according to the food index of the at least one food type under the preset type, and determining the numerical value preset range corresponding to the food index.

Further, the determining the quality index of the food to be tested based on the contribution model and the food index comprises:

determining the characteristic parameters and the corresponding contribution indexes according to the food indexes;

and multiplying the food indexes by the corresponding contribution indexes, accumulating, and adding the product indexes and the preset coefficient to determine the quality index corresponding to the food to be detected.

Further, the identifying whether the food to be detected belongs to a preset category based on the discriminant model and the food index comprises:

multiplying the food indexes by the corresponding contribution indexes, accumulating, and adding the product indexes and the preset coefficients to determine the corresponding quality indexes;

and judging whether the quality index belongs to the numerical value preset range, if so, determining that the food to be detected belongs to the preset category.

Further, the quality index includes at least one of a nutritional index, a flavor index, a food safety index, a production place index, a variety index and a brand index, and the determining whether the quality index belongs to the preset range of values, if so, the determining that the food to be tested belongs to the preset category includes:

judging whether the food to be detected belongs to the numerical value preset range of different preset types or not according to the nutritional index, the flavor index, the food safety index, the origin index, the variety index and the brand index of the food to be detected;

and if so, the food to be detected belongs to the preset category.

The invention also provides a food multi-source information fusion device, which comprises:

the system comprises an acquisition unit, a storage unit and a processing unit, wherein the acquisition unit is used for acquiring various food information of at least one food type and establishing a standard database of the corresponding food type according to the various food information;

the processing unit is used for establishing a contribution type model and/or a discriminant type model according to the data in the standard database;

the application unit is used for acquiring food indexes of food to be detected, determining the quality index of the food to be detected based on the contribution model and the food indexes and/or identifying whether the food to be detected belongs to a preset category or not based on the discriminant model and the food indexes.

The invention also provides a computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to realize the food multi-source information fusion method.

Compared with the prior art, the invention has the beneficial effects that: firstly, effectively acquiring food information of different food types, and establishing a corresponding standard database; then, performing corresponding processing based on the input data of the standard database, and establishing a contribution type model and/or a discrimination type model; and finally, inputting the food indexes of the food to be detected into the contribution type model to obtain the quality index which feeds back the quality of each aspect of the food to be detected, and inputting the quality index into the discrimination type model to determine whether the food to be detected belongs to the preset type. In conclusion, the method adopts multi-source information fusion, is favorable for excavating hidden information, increases the deep understanding of the essence of the object and the deep and refined multi-dimensional description of the object, thereby increasing the reliability, the accuracy and the comprehensiveness of the aspects of food identification, attribute judgment and the like, establishing a contribution type model, accurately and quickly determining the quality index of the food to be detected, establishing a discriminant type model, efficiently judging whether the food to be detected belongs to the preset variety or not, and ensuring the efficient, accurate and comprehensive monitoring of the food.

Drawings

FIG. 1 is a schematic view of an embodiment of an application system of a multi-source food information fusion method according to the present invention;

FIG. 2 is a schematic flow chart of an embodiment of a multi-source food information fusion method according to the present invention;

FIG. 3 is a flowchart illustrating an embodiment of step S1 in FIG. 2 according to the present invention;

FIG. 4 is a schematic flowchart of an embodiment of the method for establishing the contribution model in step S2 in FIG. 2 according to the present invention;

FIG. 5 is a flowchart illustrating an embodiment of the discriminant model building process of step S2 in FIG. 2 according to the present invention;

fig. 6 is a schematic flowchart of an embodiment of determining the quality index of the food to be tested in step S3 in fig. 2 according to the present invention;

fig. 7 is a flowchart illustrating an embodiment of identifying whether the food to be tested belongs to the preset category in step S3 in fig. 2 according to the present invention;

fig. 8 is a flowchart illustrating an embodiment of the step S703 in fig. 7 according to the present invention;

fig. 9 is a schematic structural diagram of an embodiment of a food multi-source information fusion device provided by the invention.

Detailed Description

The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.

In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implying any number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. Further, "plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.

Reference throughout this specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present invention. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the described embodiments can be combined with other embodiments.

The invention provides a food multi-source information fusion method, a device and a storage medium, which are applied to the technical field of food information. The following are detailed below:

an embodiment of the present invention provides an application system of a food multi-source information fusion method, and fig. 1 is a schematic view of a scene of an embodiment of an application system of a food multi-source information fusion method provided by the present invention, where the system may include a server 100, and a food multi-source information fusion device, such as the server in fig. 1, is integrated in the server 100.

The server 100 in the embodiment of the present invention is mainly used for:

In this embodiment of the present invention, the server 100 may be an independent server, or may be a server network or a server cluster composed of servers, for example, the server 100 described in this embodiment of the present invention includes, but is not limited to, a computer, a network host, a single network server, a plurality of network server sets, or a cloud server composed of a plurality of servers. Among them, the Cloud server is constituted by a large number of computers or web servers based on Cloud Computing (Cloud Computing).

It is to be understood that the terminal 200 used in the embodiments of the present invention may be a device that includes both receiving and transmitting hardware, i.e., a device having receiving and transmitting hardware capable of performing two-way communication over a two-way communication link. Such a device may include: a cellular or other communication device having a single line display or a multi-line display or a cellular or other communication device without a multi-line display. The specific terminal 200 may be a desktop, a laptop, a web server, a Personal Digital Assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, an embedded device, and the like, and the type of the terminal 200 is not limited in this embodiment.

Those skilled in the art can understand that the application environment shown in fig. 1 is only one application scenario of the present invention, and does not constitute a limitation on the application scenario of the present invention, and that other application environments may further include more or fewer terminals than those shown in fig. 1, for example, only 2 terminals are shown in fig. 1, and it can be understood that the application system of the food multi-source information fusion method may further include one or more other terminals, which is not limited herein.

In addition, as shown in fig. 1, the application system of the food multi-source information fusion method may further include a memory 200 for storing data, such as various food information, a standard database, a contribution type model, a discriminant type model, and the like.

It should be noted that the scene diagram of the application system of the food multi-source information fusion method shown in fig. 1 is only an example, and the application system and the scene of the food multi-source information fusion method described in the embodiment of the present invention are for more clearly illustrating the technical solution of the embodiment of the present invention, and do not form a limitation on the technical solution provided in the embodiment of the present invention.

An embodiment of the present invention provides a food multi-source information fusion method, and referring to fig. 2, fig. 2 is a schematic flow chart of an embodiment of the food multi-source information fusion method provided by the present invention, including steps S1 to S3, where:

in step S1, acquiring a plurality of food information of at least one food type, and establishing a standard database of corresponding food types according to the plurality of food information;

in step S2, a contribution type model and/or a discriminant type model is/are established according to the data in the standard database;

in step S3, a food index of a food to be tested is collected, and a quality index of the food to be tested is determined based on the contribution model and the food index and/or whether the food to be tested belongs to a preset category is identified based on the discriminant model and the food index.

In the embodiment of the invention, firstly, food information of different food types is effectively acquired, and a corresponding standard database is established; then, performing corresponding processing based on the input data of the standard database, and establishing a contribution type model and/or a discrimination type model; and finally, inputting the food indexes of the food to be detected into the contribution type model to obtain the quality index which feeds back the quality of each aspect of the food to be detected, and inputting the quality index into the discrimination type model to determine whether the food to be detected belongs to the preset type.

It should be noted that the invention establishes a food characteristic standard database by collecting relevant data of various foods and performing information classification, data acquisition and information arrangement, establishes a corresponding mathematical model based on characteristic data extracted from multi-source information of materials or foods, and quantitatively describes comprehensive characteristics including nutrition index, flavor index, food safety index, production area, variety and brand.

It should be noted that, the invention firstly establishes a food characteristic standard database, including information classification, data acquisition, information arrangement (structured processing, characteristic information extraction) and establishment of an original database; then, carrying out comprehensive information description, including establishing a multi-source information fusion algorithm, outputting comprehensive information, finally describing (portraying) food characteristics and establishing a food standard portrayal; the application comprises the steps of collecting (detecting) food information, obtaining the portrait of the food according to the method, and comparing the portrait with the established standard food portrait, so that the applications in the aspects of food quality classification, feature identification (including brands and production places), food safety risk assessment and food safety hazard identification can be realized.

It should be noted that, in a specific application scenario, collected information of rice of different varieties or production places, different meats such as pork, beef or meat on different parts of livestock and poultry, white spirits of different brands or other foods is subjected to comprehensive information description, characteristic information is extracted and input into an original standard database, the multi-source information with characteristic identification effect is fused through a mathematical algorithm and the like to establish an identification model, and the model can obtain characteristic images of the foods, including varieties, genes, nutritional characteristics, flavor characteristics, processing adaptability, health effects and food safety risks.

As a preferred embodiment, referring to fig. 3, fig. 3 is a schematic flowchart of an embodiment of step S1 in fig. 2 provided by the present invention, and step S1 specifically includes steps S11 to S14, where:

in step S11, the plurality of food information are classified into information, and structured data and unstructured data are determined; wherein, the information classification comprises food classification, including fresh food, pre-packaged food, food and beverage, food information classification, including physical, chemical and biological information, including nutriology (including nutrient components, such as energy, sugar, protein, lipid, vitamin and mineral substances), spectroscopy, bopagt, chromatography (including gas phase and liquid phase), electrochemistry (electronic nose, electronic tongue and electric potential), mass spectroscopy, mechanics, thermal, optical information and biological information (including food-borne microorganism information and genetic information of biological variety genes);

in step S12, performing data cleaning and data completion on the structured data to form numerical information;

in step S13, structuring the unstructured data and extracting corresponding feature information;

in step S14, the numerical information and the characteristic information are used as database entry data, and the corresponding standard database is established.

In the embodiment of the invention, the corresponding characteristic information is determined based on the data processing of the structured data and the unstructured data, a food standard data portrait is formed, and the numerical value information and the characteristic information are input into a standard database for subsequent data comparison.

It should be noted that the unstructured data include, but are not limited to, spectrograms, chromatograms, rice appearance analysis pictures and documents.

As a preferred embodiment, the information processing includes a structured processing of unstructured data and an extraction of feature information. In the embodiment of the invention, the unstructured data is effectively converted and processed, and the structural features of the unstructured data are extracted.

As a preferred embodiment, the structuring of the unstructured data comprises a batch process of continuous spectra. In the embodiment of the invention, intermittent processing is adopted to carry out efficient feature extraction on the image.

As a preferred embodiment, the feature information extraction includes: typical wave number of spectrogram, initial temperature of rheological map, and gene sequence of liquid quality test, and establishing food characteristic standard database by using the above structural characteristic information. In the embodiment of the invention, different maps are effectively characterized to generate corresponding characteristic information so as to build a library later.

In a specific embodiment of the invention, taking a spectrogram as an example, the characteristic information corresponding to the spectrogram comprises wavelength and absorption intensity, and the near infrared spectrums of different rice are very similar in wave crest and wave form as can be seen from the near infrared spectrums of the rice sample. The rice near infrared spectrum has obvious absorption peaks near the wavelengths of 1200, 1440, 1570 and 1720nm, the difference of the light absorption values is large in the wave bands of 1000-1100 nm, 1200-1300 nm and 1420-1799 nm, and the effective wavelength for identifying rice varieties can be distributed in the wave bands. However, the sample measured by the near infrared spectroscopy is rice which is not purified and the like, contains components such as starch, protein, moisture and the like, the background of the spectrum is complex, and the spectrum peaks of different molecules and various groups are overlapped in the same near infrared spectral region, so that the spectrum of the rice sample cannot be directly analyzed by a conventional method. In addition, by adopting the near-infrared diffuse reflection spectrum, the absorbance is influenced by the reflection of the surface of the sample, so that the spectrum distortion is caused, and background interferences such as spectrum translation, scattering errors and the like are generated. Therefore, the noise and error are reduced and the weak signal of the near infrared spectrum data set is recovered by information processing technology such as derivative spectrum, multivariate scattering correction, wavelet transformation and the like for correcting system error. Because the number of wavelength points of the near-infrared spectrogram is large, the characteristic wavelength of model analysis can be screened by adopting methods such as a competitive self-adaptive re-weighting sampling method, a principal component analysis method and the like, and the characteristic data of the spectrum can be extracted. After principal component analysis is adopted, principal components with the accumulated contribution rate larger than 80% are selected and reserved, the original 800 spectrum points are compressed into 3-8 principal components, a large amount of overlapped information is eliminated, then, the obtained principal component scores can be used as characteristic data of a rice variety identification model, and finally, a rice near infrared spectrum information principal component analysis characteristic data table corresponding to a near infrared spectrogram is obtained.

In a specific embodiment of the present invention, taking a chromatogram as an example, the feature information corresponding to the chromatogram includes retention time and peak area of each feature peak. And taking a GC-MS spectrum of a certain brand of wine as an example for explanation, comparing characteristic peaks shared by certain brands of wine produced in batches in different years to serve as characteristic peaks of the certain brand of wine, and recording peak areas and retention time as characteristic information to obtain a data table corresponding to the GC-MS spectrum.

In a specific embodiment of the invention, taking a rice appearance analysis picture as an example, the characteristic information corresponding to the rice appearance analysis picture comprises thousand grain weight, area, perimeter, aspect ratio, length, width, whole polished rice area, whole polished rice perimeter, whole polished rice aspect ratio, whole rice length, whole polished rice width, whole polished rice rate, whole polished rice thousand grain weight, small broken rice rate, transparency, precision, chalkiness grain rate, chalkiness degree, roundness and whole polished rice roundness of rice, in a model for identifying rice varieties, the conventional indexes can be subjected to dimensionality reduction through principal component analysis, multiple variables with certain relevance originally are combined into a group of new independent variables to replace original variables, principal components with an accumulated contribution rate larger than 80% are selected and reserved, and the principal component scores can be used as characteristic data of a rice variety identification model, and finally obtaining a rice appearance information principal component analysis characteristic data table corresponding to the near-rice appearance analysis diagram.

It should be noted that the national standard of rice includes the quality requirement and the sanitary requirement of rice. The characteristic information of the food characteristic portrait for multi-source information fusion comprises quality requirements, including information such as broken rice content, processing precision, imperfect grain content, yellow rice grain content, chalkiness degree, moisture content, amylose content and the like, and health standards including information such as color, smell, toxic and harmful bacteria, plant seeds, fungicidin limit indexes, pollutant limit indexes, pesticide maximum residue limit indexes and the like. The representation in the characters can be replaced by numbers for distinguishing, for example, the finishing and the adaptive finishing of the processing precision in the national standard can be respectively represented by 1 and 2. The character expression is expressed by numerical value to be processed by information structuring, thereby facilitating the characteristic image. And sorting the characteristic information to obtain a data table corresponding to the rice national standard document.

As a preferred embodiment, referring to fig. 4, fig. 4 is a schematic flow chart of an embodiment of establishing a contribution type model in step S2 in fig. 2 provided by the present invention, and specifically includes steps S401 to S402, where:

in step S401, determining a contribution index of each characteristic parameter to the food index by using a principal component analysis method and a principal component regression method according to a plurality of characteristic parameters and the food index corresponding to the at least one food category;

in step S402, the contribution type model is determined according to the plurality of characteristic parameters and the corresponding contribution indexes.

In the embodiment of the invention, the corresponding contribution type model is effectively established according to the various characteristic parameters and the corresponding contribution indexes.

As a preferred example, the food index includes at least one of a nutrition index, a flavor index, a food safety index, a place of origin index, a breed index, and a brand index, and the contribution model is represented by the following formula:

In the embodiment of the invention, the contributing model is effectively characterized through the establishment of the formula, so that the direct application of the subsequent model is facilitated.

In the above formula, f (x) is a characteristic function of the food, including the nutritional index f_yyFlavor index f_fwFood safety index f_aqAnd the place of birth f_cdVariety f_pzBrand f_pp；x₁,x₂… (n-1, 2, …, natural number) which represents parameters characteristic of food products, such as the content of nutritional components, spectroscopy, bops, chromatography, electrochemistry, mass spectrometry, mechanical, thermal, optical, biological parameters, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (a) reflects the contribution of the nutritional ingredient to the total nutrition of the food.

In a particular embodiment of the invention, the nutritional characteristics of the food product are described by a nutritional index: nutritional index:

x₁,x₂… (n is 1,2, …, and is a natural number) each indicating the content of a nutrient component, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (a) reflects the contribution of the nutritional ingredient to the total nutrition of the food.

In a particular embodiment of the invention, the flavor profile of the food product is described by a flavor index: flavor index:

x₁,x₂… (n is 1,2, …, and is a natural number), each of which indicates the content of a flavor substance, the ratio of the absorption values of the probe 2 and the probe 5 of the electronic nose probe 1, the absorption ratio of the probe 5 of the electronic tongue probe 2, the mechanical parameter elastic modulus, …, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (a) reflects the contribution of the flavour object to the overall flavour of the food product.

In the inventionIn a particular example, the safety profile of a food product is described by a safety index: safety index:

x₁,x₂… (n is 1,2, …, natural number) each indicating the content of a hazardous substance (e.g., heavy metal cadmium, pesticide residue …) in food, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (d) reflects the contribution of the safety detection index information to the overall food safety of the food.

As a preferred embodiment, referring to fig. 5, fig. 5 is a schematic flowchart of an embodiment of establishing the discriminant model in step S2 in fig. 2 according to the present invention, which specifically includes step S501 to step S503, where:

in step S501, determining a contribution index of each characteristic parameter to the food index according to a plurality of characteristic parameters and the food index corresponding to the at least one food category;

in step S502, determining a corresponding discrimination function value according to the plurality of feature parameters and the corresponding contribution index;

in step S503, the discrimination function value is calculated according to the food index of the at least one food type in the preset type, and a preset range of values corresponding to the food index is determined.

In the embodiment of the invention, the corresponding discrimination function value is obtained by establishing the discrimination model, and the standard numerical value preset range is effectively obtained according to the discrimination function value, so that the comparison and the identification of the subsequent food to be detected are ensured.

In one specific embodiment of the present invention, the discriminant model is represented by the following equation:

in the above formula, f (x) is the discriminant function of food classification, including the nutritional index f_yyWind and windTaste index f_fwFood safety index f_aqAnd the place of birth f_cdVariety f_pzBrand f_pp；x₁,x₂… (n-1, 2, …, natural number) which represents parameters characteristic of food products, such as the content of nutritional components, spectroscopy, bops, chromatography, electrochemistry, mass spectrometry, mechanical, thermal, optical, biological parameters, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (a) reflects the contribution of the nutritional ingredient to the total nutrition of the food. Wherein, taking the region as an example, if f (x) belongs to (p1, p 2)]The food production place is judged as area 1, otherwise, other areas.

In a particular embodiment of the invention, the characteristics of the origin of the food product are described by the origin index: the producing area index:

x₁,x₂… denotes the parameters of regional characteristics in the food product (e.g. content of mineral iron, content of copper, ratio of iron to copper, …, a respectively₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (a) reflects the contribution of the regional characteristic parameter to the total food region; if f (x) belongs to (p1, p 2)]The food production place is judged as area 1, otherwise, other areas.

In a particular embodiment of the invention, the breed characteristics of the food product are described by a breed index: variety index:

x₁,x₂… shows the peak time and the infrared characteristic absorption wave number of a chromatogram (e.g. gas chromatography, liquid chromatography) in a food, …, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (d) reflects the contribution of the variety information to the total food variety; if f (x) belongs to (p1, p 2)]If not, the food variety is judged as variety 1.

In a particular embodiment of the present invention,the brand identity of a food product is described by a brand index: brand index:

x₁,x₂… shows the peak time and the infrared characteristic absorption wave number of a chromatogram (e.g. gas chromatography, liquid chromatography) in a food, …, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nReflects the contribution of the brand information to the total food brand; if f (x) belongs to (p1, p 2)]If so, the food brand is judged as brand A, otherwise, the food brand is judged as other brands.

As a preferred embodiment, referring to fig. 6, fig. 6 is a schematic flow chart of an embodiment of determining the quality index of the food to be tested in step S3 in fig. 2, which is provided by the present invention, and specifically includes steps S601 to S602, where:

in step S601, determining the characteristic parameter and the corresponding contribution index according to the food index;

in step S602, the food index is multiplied by the corresponding contribution index, then accumulated, and added to the preset coefficient, so as to determine the quality index corresponding to the food to be measured.

In the embodiment of the invention, aiming at the food to be measured, the established contribution model is utilized to calculate the corresponding quality index and feed back the food quality information.

As a preferred embodiment, referring to fig. 7, fig. 7 is a flowchart illustrating an embodiment of identifying whether the food to be tested belongs to the preset category in step S3 in fig. 2 according to the present invention, which specifically includes steps S701 to S703, where:

in step S701, determining the characteristic parameter and the corresponding contribution index according to the food index;

in step S702, the food index is multiplied by the corresponding contribution index, then accumulated, and then added to the preset coefficient to determine the corresponding quality index;

in step S703, it is determined whether the quality index belongs to the preset range of values, and if so, the food to be tested belongs to the preset category.

In the embodiment of the invention, aiming at the food to be detected, the established discriminant model is utilized to calculate the corresponding quality index, and the quality index is compared with the numerical value preset range to identify the preset type.

As a preferred embodiment, referring to fig. 8, fig. 8 is a schematic flowchart of an embodiment of step S703 in fig. 7 provided by the present invention, and specifically includes steps S801 to S802, where:

in step S801, determining whether the food to be tested belongs to the preset range of values of different preset categories according to the nutritional index, the flavor index, the food safety index, the origin index, the variety index and the brand index of the food to be tested;

in step S802, if yes, the food to be tested belongs to the preset category.

In the embodiment of the invention, the food is identified based on the indexes of various foods, whether the foods belong to the corresponding preset types is judged, and the multi-aspect identification is effectively completed.

In a specific embodiment of the present invention, for example, the identification of the shredded green pepper meat as pork or beef is taken as an example, there is a certain difference in physical, chemical and biological information between pork and beef. Therefore, a multi-source information fusion mode can be established to judge whether the shredded meat in the shredded meat with green pepper is pork or beef. At present, the near infrared technology, the electronic nose technology and the PCR gene technology are commonly used for distinguishing the pig and the beef. And then the pigs are put into the pig feed. The near infrared spectrum, the electronic nose data, the DNA information and the like of the beef are directly obtained and input into an original standard database, a multi-source information distinguishing model is established, and then the information of the measured sample is compared with the pig and beef information in the original standard database, so that whether the meat in the green pepper shredded meat is the pork or the beef can be more effectively and accurately identified;

wherein, purchasing raw materials known as pork and beef, respectively placing the raw materials into a meat grinder to be pulped into muddy flesh, respectively adding the pork into the beef according to the mass ratio of 0 percent (pure beef), 10 percent, 30 percent, 50 percent, 70 percent, 90 percent and 100 percent (pure pork), uniformly stirring to prepare test samples, and sampling for 20 times in each ratio to obtain 140 samples in total;

wherein, the acquisition of near infrared spectrum data: and collecting the meat sample diffuse reflection spectrum by adopting a Fourier transform near infrared instrument. Each sample was made to weigh 50g, and each sample was scanned 3 times in duplicate, and the average was taken as the raw data for that sample. Scanning wave band range 10000-4000 cm^-1Resolution of 4cm^-1The number of scans was 32. In order to improve the signal-to-noise ratio, the original spectrum is preprocessed, principal component analysis is adopted to carry out dimensionality reduction and decorrelation processing on the spectrum data, and feature data are extracted;

wherein, the collection of electron nose data: weighing 3-10 g of sample in a gas collection bottle, collecting gas in a headspace manner for 30min at room temperature, detecting according to the operation and working process of the electronic nose, and extracting characteristic data of the signal response value of the electronic nose sensor by adopting methods such as main component analysis, discriminant factor analysis and the like.

Wherein, the acquisition of gene data: extracting and purifying the genomic DNA of the sample by using a DNA extraction kit, dissolving the extracted DNA sample in a buffer solution, and storing at-20 ℃. Designing a forward universal primer according to the gene sequence of the mitochondrial cytochrome b of the cattle and the pig; designing a special reverse primer according to the species-specific DNA sequence of the cattle and the pig. The PCR reaction was carried out in a 50. mu.L PCR reaction system. Then, detecting the result of the PCR amplification product by agarose gel electrophoresis to obtain characteristic gene data of the pork and the beef;

the method comprises the following steps of inputting spectral information, electronic nose information and gene information of pork and beef into an original standard database, and establishing a multisource information fusion discrimination model of the spectrum, the electronic nose and the gene:

x₁,x₂… denotes in sub-table the characteristic data of the near infrared characteristic absorption wave number and absorption intensity, the electric nose sensor signal response value, DNA data, a₀,a₁,a₂… is a coefficient; will be provided withx_nAfter normalization, a_nThe size of (d) reflects the contribution of the variety information to the total food variety; if f (x) belongs to (p1, p 2)]Judging the shredded meat in the shredded meat with green pepper to be pork, otherwise, judging the shredded meat to be beef;

the method comprises the following steps of carrying out corresponding identification, collecting the green pepper shredded meat, processing the green pepper shredded meat raw materials, selecting the shredded meat, placing the shredded meat in a meat grinder, and pulping for later use. And (3) placing the measured 50g sample at the sample cell light hole of a spectrometer for scanning, repeatedly scanning for 3 times, taking an average value as original spectrum data of the sample, preprocessing the spectrum, performing dimensionality reduction on the near infrared spectrum data after meat sample preprocessing by adopting principal component analysis and the like, and extracting characteristic data of the spectrum. And simultaneously, taking a measured 3-10 g sample to collect a sensor signal response value, performing dimensionality reduction on the meat electronic nose data, and extracting feature data of the signal response value. And extracting the DNA of the measured meat paste to obtain gene data. And substituting the measured data into a multi-source information fusion variety distinguishing model to be compared with the standard portrait to obtain whether the meat in the green pepper shredded meat is pork or beef.

In a specific embodiment of the invention, taking the identification of judging whether the rice is Thai jasmine rice as an example, the Thai jasmine rice has excellent quality, higher taste quality and higher nutritional value, and is popular with consumers in various countries, but adverse phenomena such as inferior quality, adulteration and the like often occur in the Thai jasmine rice market. Therefore, the establishment of the rapid identification method of the Thailand jasmine rice has important significance for promoting the sustainable and healthy development of the high-quality rice industry. The Thai jasmine scented rice has certain difference with other varieties of scented rice in physical, chemical and biological information, so that the Thai jasmine scented rice can be identified based on near infrared spectrum, electronic nose and multi-source information fusion of conventional indexes, and the problem of low accuracy in identifying the Thai jasmine scented rice by common methods such as near infrared spectrum information and the like can be solved;

wherein, firstly, preparing a rice raw material, and selecting one Thai jasmine rice produced in Thailand Siamese Jufu and four Thai jasmine rice produced in Thailand Wuwenfu, wherein 5 Thailand jasmine rice are selected. Then, different varieties of non-Thai jasmine rice produced in various countries and regions, namely, 134 non-Thai jasmine rice, were prepared. Representative non-thailand jasmine rice among them includes: jinjiangding Jiatai scented rice, Cambodia jasmine scented rice and Thailand rice. Arranging and combining raw rice materials:

wherein, the sample with the Thailand jasmine rice content of 100 percent: any two of the five Thai jasmine rice are arranged and combined, the content ratio of the two Thai jasmine rice in each combination is 9:1, 7:3, 5:5, 3:7 and 1:9, and 50 parts of samples are used in total. Taking any three of the five Thai jasmine rice in a permutation and combination way, wherein the content ratio of the three Thai jasmine rice in each combination is 1:1:1, 1:2:2, 2:1:2 and 2:2:1, and the total weight is 40 parts of samples;

wherein, the Thailand jasmine scented rice is a sample with the content of 92-98 percent: the Jinjiangding Jiatai scented rice, the Cambodia jasmine scented rice and the Thailand rice are respectively proportioned with five Thailand jasmine scented rice, so that the content of the Thailand jasmine scented rice in the mixed rice is respectively 92%, 94%, 96% and 98%, and 60 parts of samples are used in total. Five Thai jasmine scented rice are added, so that the total Thai jasmine scented rice sample comprises 155 parts;

wherein, the Thai jasmine rice content is 20-80% of the sample (non-Thai jasmine rice): the Jinjiangding Jiatai scented rice, the Cambodia jasmine scented rice and the Thailand rice are respectively proportioned with five Thailand jasmine scented rice, so that the content of the Thailand jasmine scented rice in the mixed rice is respectively 20%, 40%, 60% and 80%, and 60 parts of samples are used in total. The original 134 non-Thai jasmine rice samples are added, so the total non-Thai jasmine rice samples comprise 194 parts;

wherein, near infrared spectrum data acquisition: preheating the near infrared spectrometer for 30min, performing instrument self-inspection, performance test and white board reference, pouring the rice sample into a sample box, filling, flattening with a sample box cover, placing the sample in a specified position, and starting spectral measurement. Respectively collecting the spectrums of Thai jasmine rice samples and non-Thai jasmine rice samples, repeatedly loading and scanning each sample for 6 times, and taking the average value as spectrum original data. To improve the signal-to-noise ratio of the spectrum, and to eliminate baseline and other background interference, pre-processing of the near infrared spectrum is required. The spectrum pretreatment can be carried out on the original spectrum of the sample, and four methods of first derivative correction, baseline correction, multivariate scattering correction and trend removing correction are selected. Modeling the original spectral data and the spectral data after preprocessing the data, and selecting an optimal preprocessing method by comparing modeling results;

wherein, electronic nose data acquisition: 5g of the sample is weighed, placed in a 20mL sample bottle and sealed. Test parameters are as follows: the carrier gas is synthetic dry air, the flow rate is 150mL/min, the headspace generation time is 120s, the headspace generation temperature is 60 ℃, the stirring speed is 500r/min, the headspace injection volume is 2.5mL, the headspace injection speed is 2.3mL/s, the total volume of the injection needle is 5.0mL, the injection needle temperature is 60 ℃, the acquisition time is 120s, and the delay time is 300 s. 4 replicates per sample;

wherein, the rice routine index is collected, 10g of rice sample is weighed, the rice sample is flatly paved on a scanner of a rice appearance detector to scan rice appearance pictures, and a seed rice appearance quality detection analyzer system is utilized to calculate the thousand grain weight, the area, the perimeter, the length-width ratio, the length, the width, the whole polished rice area, the whole polished rice perimeter, the whole polished rice length-width ratio, the whole polished rice length, the whole polished rice width, the whole polished rice rate, the whole polished rice thousand grain weight, the small broken rice rate, the transparency, the precision, the chalky grain rate, the chalky whiteness, the roundness and the whole polished rice roundness; measuring the whiteness of the sample by using a color difference meter;

extracting characteristic information, before the model is established, performing dimensionality reduction treatment on conventional indexes of rice, electronic nose data and preprocessed rice near infrared spectrum data by adopting principal component analysis, extracting the characteristic data, and establishing a multi-source information fusion discrimination model of the Thailand jasmine scented rice:

x₁,x₂… represents the characteristic data of the near infrared characteristic absorption wave number and absorption intensity in food, the response value of the electronic nose sensor signal, the conventional index data, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (d) reflects the contribution of the variety information to the total food variety; if f (x) belongs to (p1, p 2)]Then, it is determined as ThailandJasmine scented rice, otherwise non-Thai jasmine scented rice;

wherein, the identification process is as follows: and processing the conventional index of the rice to be identified, the electronic nose and the near infrared spectrum data, extracting characteristic data, substituting the characteristic data into a multi-source information variety discrimination model, and comparing the characteristic data with the established standard food portrait to obtain whether the rice is Thailand jasmine rice.

In a specific embodiment of the present invention, the pork is identified as lean meat, African swine fever, etc. At present, the problem of safety of pork always appears endlessly, the problem of clenbuterol is frequent up to now, and the African swine fever appears all over the country in recent years, which has great harm to the health of consumers. Therefore, it is very interesting to establish an effective pork safety identification method. Normal pork and pork containing clenbuterol have good appearance. The pork quality and the physicochemical information are different, the physical, chemical and biological information of normal pork is input into an original standard database, and the information of clenbuterol and African swine fever virus is also input into the standard database, so that a safety characteristic contribution model of the pork is established. In the process of detecting the pork safety problem, only the physical, chemical and biological information of the pork to be detected needs to be measured, the physical, chemical and biological information is compared with the information of normal pork, and whether the measured sample has the same information of clenbuterol and African swine fever virus in a standard database or not is judged, so that whether the pork has the safety problem or not can be judged;

wherein, the normal fresh pork is neutral or weakly alkaline, the pH value of the pork is approximately 5.6-6.0 after the pork is slaughtered and placed for 6 hours under natural conditions, and the pH value of the pork with residual lean meat is obviously lower than the range. Firstly, purchasing fresh pork samples which do not contain known hazardous substances such as clenbuterol, African swine fever virus and the like and are placed for different time periods, and measuring the pH value of the pork samples; research shows that the pork using the clenbuterol has obvious difference from various nutrient components such as creatine, amino acid, carnitine, vitamin B6, saccharides, carnosine, phospholipid and the like in normal pork tissues through main component analysis, and the nutrient components of the pork sample without hazardous substances are measured; determining retention time of clenbuterol (including clenbuterol, salbutamol, ractopamine, terbutaline, iprobuterol, chlorpropaline, cimaterol, fenoterol and the like, which is generally referred to as clenbuterol hydrochloride in China) on a chromatogram by using a high performance liquid chromatography; performing direct mass spectrometry on clenbuterol components by using internal electrospray extraction ionization mass spectrometry in a positive ion detection mode and using methanol water as an extracting agent under the condition of no need of any sample pretreatment to obtain a chemical fingerprint spectrogram in a certain mass-to-charge ratio range and obtain characteristic fragment ion information (mass-to-charge ratio) of clenbuterol;

among them, African Swine Fever (ASF) is a virulent infectious disease caused by African Swine Fever Virus (ASFV), and the current detection method for African swine fever virus is mainly based on fluorescent quantitative PCR, and has high specificity and good sensitivity. Firstly, designing and synthesizing primers and probes according to A104R, B602L, CP204L, B646L and K205R gene sequences of ASFV gene II type HLJ/18 strain published in GenBank and P72 gene nucleotide sequence of ASFV SY18 strain (MH713612), using the primers and probes as real-time fluorescent quantitative PCR primers and probes, constructing recombinant plasmid, and carrying out PCR reaction in a PCR reaction system to obtain DNA information of African swine fever virus;

the pH value and the nutritional ingredient information of normal pork which does not contain hazardous substances such as clenbuterol, African swine fever virus and the like, the retention time of clenbuterol on a chromatogram, the characteristic fragment ion information on a mass spectrogram and the gene information of the African swine fever virus are input into an original standard database to be used as the standard for detecting whether a pork sample contains clenbuterol and the African swine fever virus. Establishing a safety contribution model of the pork,

x₁,x₂… respectively indicates pH value, nutrient content, retention time of pork on chromatogram, ion information (mass-to-charge ratio) of characteristic fragment on mass spectrum, DNA information extracted from meat sample, and a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of (a) reflects the contribution of the safety detection index information to the total food safety of the food;

wherein, the identification process is as follows: collecting a pork sample to be detected, firstly, measuring the pH value of the pork sample by using a pH value method, then measuring the nutritional ingredients of the pork sample, carrying out pretreatment, then obtaining chromatogram and mass spectrum information by using a high performance liquid chromatogram and an internal electrospray extraction ionization mass spectrum, extracting DNA information of the pork sample, substituting the DNA information into a safety characteristic contribution model of the pork, comparing the DNA information with the multi-source information of normal pork, clenbuterol and African swine fever virus in an original standard library, and finally judging whether the pork contains harmful substances with safety problems, such as clenbuterol, African swine fever and the like.

In a specific embodiment of the present invention, the brand identification of white spirit is taken as an example. The outstanding taste quality and unique brewing process of certain brand of wine as the national wine in China enjoy the reputation at home and abroad, but the phenomenon of counterfeit and shoddy certain brand of wine often occurs in the wine product market. It is therefore of great importance to establish a method for identifying a certain brand of wine quickly and accurately. Certain brand wine has certain difference with other wine in physical and chemical information. Therefore, a standard database can be established by obtaining multi-source information of a real certain brand of wine, and a brand discrimination model is obtained to discriminate whether the brand of wine is the real certain brand of wine;

firstly, purchasing a certain Guizhou brand wine produced in batches in different years, purchasing a certain brand wine with unknown authenticity in the market, then identifying the certain brand wine by a professional wine taster expert, and additionally purchasing white spirit produced by other manufacturers;

wherein, infrared spectrum data acquisition: and selecting a primary infrared spectrogram and a second derivative spectrogram of the white spirit sample obtained by using an FTIR (infrared Fourier transform infrared spectrometer). The characteristic peaks of esters, organic acids, phenols, ketones, carboxylic acids, acetates and the like are different between true brands and false brands, so that the positions of the characteristic peaks and the absorption peak intensities are different. The infrared spectrogram is subjected to second derivative processing to distinguish overlapping peaks, and the quantity and the peak positions of automatic peaks and cross peaks of the true and false certain brand of wine are different, so that the true and false certain brand of wine can be directly judged;

wherein, near infrared spectrum data acquisition: all samples were equilibrated overnight at a certain temperature and humidity, and the near infrared spectra of the samples were collected using the transmission sampling system of the near infrared spectrometer. Selecting a proper collection wave number range according to requirements, taking an instrument built-in background as a reference, using 32 times of scanning for each sample and the reference, and preprocessing a spectrum in order to improve the signal-to-noise ratio of the spectrum so as to obtain preprocessed near-infrared spectrogram information of real and fake certain brands of wine;

wherein, gas chromatography-mass spectrometry data acquisition: detecting certain brands of wine produced in batches in different years, common white spirits of other varieties and map information of certain brands of wine by using a gas chromatography-mass spectrometry method, selecting a common peak of certain brands of wine which appears in the true certain brands of wine and does not appear in other white spirits as a characteristic peak of the true certain brands of wine, and recording peak areas and retention time as characteristic information;

the collected spectrogram and chromatogram information is collected, characteristic information of true and false certain brand wine is extracted, and then the characteristic information is input into a standard database to establish a standard portrait, namely a brand discrimination model of certain brand wine:

x₁,x₂… respectively represents retention time, peak area, infrared characteristic absorption wavenumber, absorption intensity, near infrared characteristic absorption wavenumber and absorption intensity of chromatogram in a certain brand of wine, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nReflects the contribution of the brand information to the total food brand; if f (x) belongs to (p1, p 2)]Judging the liquor to be certain brand liquor, otherwise, judging the liquor to be other liquor;

wherein, the identification process is as follows: collecting a white spirit sample to be identified, detecting the white spirit sample to obtain an infrared spectrogram, a near infrared spectrogram and a gas chromatography-mass spectrometry fingerprint of the white spirit, extracting characteristic information, substituting the characteristic information into a brand discrimination model of certain brand of wine, and comparing the characteristic information with a standard characteristic portrait of certain brand of wine in an original standard database to obtain whether the white spirit is the certain brand of wine or not. The model is also suitable for judging white spirits of other brands, and the model is established by collecting feature data of white spirits of the brands to be identified to perform feature portrait.

In a specific embodiment of the invention, the identification of saccharides in the formula is taken as an example. Sugar is an important substance in milk powder, and lactose is the main substance. Lactose forms organic acid after the action of lactase in human body, can promote the absorption of calcium ions and the growth of probiotics, and can promote the absorption of calcium ions and the growth of beneficial bacteria in intestinal tract. However, lactose intolerance phenomena such as abdominal pain and diarrhea occur in elderly people, infants with intestinal dysplasia and lactose intolerant people when a large amount of lactose is eaten. Sucrose, a commonly used food additive, is also often added to milk powder to increase the sugar content, but too much sucrose consumption can have an impact on health. At present, galactooligosaccharides are also often added to milk powder as important prebiotics. Therefore, the establishment of a method for effectively and rapidly detecting the contents of glucose, sucrose, galactose, lactose and other sugars in the milk powder is very important;

wherein, a plurality of brands and types of formula milk powder are purchased, wherein the formula milk powder is suitable for different ages and special groups, such as milk powder suitable for lactose intolerant people, and the nutritional formula is detected according to a standard method, so that high-quality milk powder meeting the sugar content standard is obtained. Then selecting near infrared spectrum, high performance liquid chromatography and ion chromatography-mass spectrometry combined technology to obtain the spectrum and chromatographic information of the milk powder;

wherein, near infrared spectrum data acquisition: the milk powder samples are subjected to near-infrared spectrograms by a near-infrared spectrometer. Selecting a proper collection wave number range according to requirements, taking an instrument built-in background as a reference, using 32 times of scanning for each sample and reference, and preprocessing a spectrogram to extract spectral characteristic information in order to improve the signal-to-noise ratio;

wherein, the data acquisition of the high performance liquid chromatography: determining saccharides in the milk powder samples by using a high performance liquid chromatograph, selecting a common peak of all samples which meet the requirement that milk powder appears (does not appear) but does not appear (appears) in other milk powder as a characteristic peak of the special formula milk powder, and recording peak area and retention time as characteristic information;

wherein, the ion chromatography-mass spectrometry data acquisition: detecting saccharides in the milk powder samples by adopting an ion chromatography-mass spectrometry, selecting a common peak of all samples which meet the requirement that milk powder appears (does not appear) but does not appear (appears) in other milk powder as a characteristic peak of the milk powder with a special formula, and recording peak area and retention time as characteristic information;

the collected spectrogram and chromatogram information are collected, characteristic information of the milk powder meeting the formula requirements and not meeting the requirements is extracted, and then the characteristic information is input into a standard database to establish a standard portrait, namely a saccharide contribution model of the milk powder:

x₁,x₂… (n is 1,2, …, and is a natural number) each indicating the content of a different saccharide in the milk powder, a₀,a₁,a₂… is a coefficient; x is to be_nAfter normalization, a_nThe size of the milk powder reflects the contribution of different saccharides in the milk powder to the total nutrition of the milk powder;

wherein, the identification process is as follows: verifying and identifying the milk powder suitable for the lactose intolerant person marked on the product label, collecting the chromatographic information and the spectral information of the milk powder and extracting characteristic data, substituting the characteristic data into a sugar contribution model of the milk powder, comparing the characteristic image with a characteristic image of standard milk powder which is built in an original standard database and does not contain or contains a small amount of lactose and is suitable for the lactose intolerant person, and determining the milk powder suitable for the lactose intolerant person if the coincidence rate is high; verifying and identifying a milk powder product with sucrose or other saccharides in a certain range marked on a product label, collecting chromatographic information and spectral information of the milk powder and extracting characteristic data, substituting the characteristic data into a saccharide contribution model of the milk powder, comparing the characteristic data with a characteristic image of standard milk powder with saccharides meeting requirements in a certain range established in an original standard database, and if the coincidence rate is high, determining the standard milk powder with the saccharides in the label marking range of the milk powder, wherein the standard milk powder is qualified if the saccharides in the milk powder content in the label marking range is met; if the coincidence rate of the characteristic images of the milk powder and the standard milk powder is low, the content of the sugar to be detected of the milk powder is not in the range, the milk powder does not meet the standard requirement, and the milk powder is unqualified.

An embodiment of the present invention further provides a food multi-source information fusion device, and with reference to fig. 9, fig. 9 is a schematic structural diagram of an embodiment of the food multi-source information fusion device provided by the present invention, including:

an obtaining unit 901, configured to obtain multiple food information of at least one food type, and establish a standard database of corresponding food types according to the multiple food information;

a processing unit 902, configured to establish a contribution type model and/or a discriminant type model according to data in the standard database;

the application unit 903 is configured to collect food indexes of food to be tested, determine quality indexes of the food to be tested based on the contribution model and the food indexes, and/or identify whether the food to be tested belongs to a preset category based on the discriminant model and the food indexes.

The more specific implementation manner of each unit of the food multi-source information fusion device can be referred to the description of the accessing food multi-source information fusion method, and has similar beneficial effects, and the detailed description is omitted here.

The embodiment of the invention also provides a computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and when the program is executed by a processor, the food multi-source information fusion method is realized.

Generally, computer instructions for carrying out the methods of the present invention may be carried using any combination of one or more computer-readable storage media. Non-transitory computer readable storage media may include any computer readable medium except for the signal itself, which is temporarily propagating.

The embodiment of the invention also provides computing equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein when the processor executes the program, the food multi-source information fusion method is realized.

According to the computer-readable storage medium and the computing device provided by the above embodiments of the present invention, the content specifically described for implementing the food multi-source information fusion method according to the present invention can be referred to, and the beneficial effects similar to those of the food multi-source information fusion method described above are achieved, and are not repeated herein.

The invention discloses a food multi-source information fusion method, a device and a storage medium, wherein, firstly, food information of different food types is effectively obtained, and a corresponding standard database is established; then, performing corresponding processing based on the input data of the standard database, and establishing a contribution type model and/or a discrimination type model; and finally, inputting the food indexes of the food to be detected into the contribution type model to obtain the quality index which feeds back the quality of each aspect of the food to be detected, and inputting the quality index into the discrimination type model to determine whether the food to be detected belongs to the preset type.

According to the technical scheme, the credibility of the data is enhanced by adopting multi-source information fusion, the contribution type model is established so as to accurately and quickly determine the quality index of the food to be detected, the discrimination type model is established so as to efficiently discriminate whether the food to be detected belongs to the preset variety or not, and efficient, accurate and comprehensive monitoring on the food is ensured.

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.

Claims

1. A food multi-source information fusion method is characterized by comprising the following steps:

2. The food multi-source information fusion method according to claim 1, wherein the establishing a standard database of corresponding food types according to the plurality of food information comprises:

3. The food multi-source information fusion method according to claim 1, wherein the database entry data comprises a plurality of characteristic parameters, and the establishment process of the contribution type model comprises the following steps:

4. The food multi-source information fusion method of claim 3, wherein the food index comprises at least one of a nutrition index, a flavor index, a food safety index, a place of origin index, a variety index and a brand index, and the contribution model is represented by the following formula:

5. The food multi-source information fusion method according to claim 4, wherein the database entry data comprises a plurality of characteristic parameters, and the establishment process of the discriminant model comprises the following steps:

6. The food multi-source information fusion method according to claim 4, wherein the determining the quality index of the food to be tested based on the contribution model and the food index comprises:

7. The food multi-source information fusion method according to claim 5, wherein the identifying whether the food to be tested belongs to a preset category based on the discriminant model and the food index comprises:

8. The food multi-source information fusion method according to claim 7, wherein the quality index includes at least one of a nutrition index, a flavor index, a food safety index, a place of production index, a variety index and a brand index, and the determining whether the quality index belongs to the preset range of values, if so, the food to be tested belongs to the preset category includes:

and if so, the food to be detected belongs to the preset category.

9. A food multi-source information fusion device, comprising:

10. A computer-readable storage medium, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the food multi-source information fusion method according to any one of claims 1-8.