CN108932483A - A kind of odor identification method and device based on machine olfaction - Google Patents

Abstract

The odor identification method based on machine olfaction that the invention discloses a kind of, comprising: obtain the hybrid response signal of target mixing smell；Judge in hybrid response signal with the presence or absence of unknown response signal；If so, filtering out unknown response signal from hybrid response signal, target response signal is obtained, and target response signal and the template response signal in presetting database are compared, determine the corresponding odor types of target response signal and content.This method filters out the unknown response signal in hybrid response signal first, has an impact to avoid unknown smell to recognition result, to improve the precision of recognition efficiency, accuracy and smell quantitative analysis before the hybrid response signal of processing mixing smell.Correspondingly, a kind of odor identification device based on machine olfaction, equipment and computer readable storage medium disclosed by the invention, similarly have above-mentioned technique effect.

Description

A method and device for odor recognition based on machine sense of smell

technical field

The present invention relates to the technical field of artificial intelligence, and more specifically, to a method, device, equipment and computer-readable storage medium for smell recognition based on machine sense of smell.

Background technique

At present, the application of machine olfactory technology is becoming more and more extensive. There have been a large number of application cases in many aspects such as industrial gas monitoring, chemical product quality monitoring, indoor and outdoor environmental monitoring, medical diagnosis, and identification of Chinese medicinal materials.

In these application fields, most of the monitored odors are mixed odors, and the odor information in the electronic nose database is limited, so there must be unknown gases. Moreover, the existing odor recognition technology does not take into account the influence of unknown odors on the recognition results when processing the response signal of the sensor, resulting in large errors in the recognition results, reducing the efficiency and accuracy of recognition and the precision of quantitative odor analysis.

Therefore, how to improve the efficiency and accuracy of odor recognition and the precision of odor quantitative analysis is a problem to be solved by those skilled in the art.

Contents of the invention

The object of the present invention is to provide a method, device, equipment and computer-readable storage medium for odor identification based on machine sense of smell, so as to improve the efficiency and accuracy of odor identification and the precision of odor quantitative analysis.

In order to achieve the above object, the embodiment of the present invention provides the following technical solutions:

A method for odor recognition based on machine olfaction, comprising:

Obtain the mixed response signal of the target mixed odor;

judging whether there is an unknown response signal in the mixed response signal;

If so, filter out the unknown response signal from the mixed response signal to obtain a target response signal, and compare the target response signal with a template response signal in a preset database to determine the target response signal corresponding to Odor type and content.

Wherein, the unknown response signal is filtered out from the mixed response signal to obtain a target response signal, and the target response signal is compared with a template response signal in a preset database to determine that the target response signal corresponds to Types and levels of odors, including:

judging whether the mixed response signal is a linear response signal;

If so, use the optimization function in the linear mixed model to filter out the unknown response signal to obtain the target response signal, compare the target response signal with the template response signal in the preset database, and determine by the multiplier method The odor type and content corresponding to the target response signal; H represents the filter mixing matrix, z(t) represents the unknown response signal, t represents the time variable, and x(tP) represents the noise signal.

Wherein, when there is no unknown response signal in the mixed response signal, and when the mixed response signal is a linear response signal, it also includes:

Using the optimization function in the linear mixed model to filter out the unknown response signal to obtain the target response signal;

The target response signal is compared with the template response signal in the preset database, and the odor type and content corresponding to the target response signal are determined by a ridge regression method.

Wherein, when the mixed response signal is a nonlinear response signal, it also includes:

filtering out the unknown response signal by using an optimization function in the nonlinear mixed model to obtain the target response signal;

Comparing the target response signal with the template response signal in the preset database, and determining the odor type and content corresponding to the target response signal through a chaotic genetic algorithm;

Wherein, the optimization function in the nonlinear mixed model is: min||y(t)-g(t)-z(t)||, y(t) represents the mixed response signal, and g(t) represents the nonlinear Response signal, z(t) represents the unknown response signal.

Wherein, the nonlinear mixed model is generated using a multivariate adaptive regression spline method, and the specific generation steps are:

Obtaining experimental data of mixed odors, and correcting the experimental data for environmental factors;

Divide the corrected experimental data to obtain multiple intervals, and generate corresponding basis functions based on each interval;

The basis function with the highest accuracy is selected by using the average error method, and the model corresponding to the basis function with the highest accuracy is determined as the nonlinear mixed model.

Wherein, the determination of the odor type and content corresponding to the target response signal through the chaotic genetic algorithm includes:

generating a population of the chaotic genetic algorithm based on the target response signal;

Carry out individual evaluation on the individuals in the population, and perform selection operation, crossover operation and mutation operation on the population that has completed the individual evaluation, so that the population is iteratively updated;

Determine whether the current evolution algebra is greater than the preset maximum evolution algebra;

If yes, output the optimal solution, and determine the type and content of the odor corresponding to the target response signal according to the optimal solution.

An odor recognition device based on machine sense of smell, comprising:

Obtaining module, for obtaining the mixed response signal of target mixed smell;

A judging module, configured to judge whether there is an unknown response signal in the mixed response signal;

An identification module, configured to filter out the unknown response signal from the mixed response signal when there is an unknown response signal in the mixed response signal, obtain a target response signal, and compare the target response signal with the preset database The template response signals are compared to determine the odor type and content corresponding to the target response signal.

Wherein, the identification module includes:

a judging unit, configured to judge whether the mixed response signal is a linear response signal;

An identification unit, configured to use an optimization function in a linear mixed model to filter out the unknown response signal when the mixed response signal is a linear response signal to obtain the target response signal, and compare the target response signal with a preset database Compare the template response signals in , and determine the odor type and content corresponding to the target response signal by the multiplier method;

H represents the filter mixing matrix, z(t) represents the unknown response signal, t represents the time variable, and x(tP) represents the noise signal.

An odor recognition device based on machine sense of smell, comprising:

memory for storing computer programs;

A processor, configured to implement the steps of any one of the above-mentioned odor recognition methods based on machine sense of smell when executing the computer program.

A computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of any one of the above-mentioned odor recognition methods based on machine sense of smell are realized.

From the above schemes, it can be seen that a method for odor recognition based on machine olfaction provided by an embodiment of the present invention includes: acquiring a mixed response signal of a target mixed odor; judging whether there is an unknown response signal in the mixed response signal; The unknown response signal is filtered out from the mixed response signal to obtain the target response signal, and the target response signal is compared with the template response signal in the preset database to determine the odor type and content corresponding to the target response signal.

It can be seen that before processing the mixed response signal of the mixed odor, the method first filters out the unknown response signal in the mixed response signal to avoid the impact of the unknown odor on the recognition result; at the same time, the target response obtained after filtering out the unknown response signal The signal is compared with the template response signal in the preset database to determine the type and content of the odor corresponding to the target response signal, thereby improving the recognition efficiency and accuracy, and also improving the accuracy of odor quantitative analysis.

Correspondingly, the apparatus, equipment and computer-readable storage medium for smell recognition based on machine sense of smell provided by the embodiments of the present invention also have the above-mentioned technical effects.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is a flow chart of a method for smell recognition based on machine sense of smell disclosed in an embodiment of the present invention;

Fig. 2 is a schematic diagram of the response modes of three similar odors disclosed in the embodiment of the present invention:

FIG. 3 is a flow chart of another odor recognition method based on machine sense of smell disclosed in an embodiment of the present invention;

Fig. 4 is a schematic diagram of an odor recognition device based on machine sense of smell disclosed in an embodiment of the present invention;

Fig. 5 is a schematic diagram of an odor recognition device based on machine sense of smell disclosed in an embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

The embodiment of the present invention discloses a method, device, equipment and computer-readable storage medium for smell recognition based on machine sense of smell, so as to improve the efficiency and accuracy of smell recognition and the precision of quantitative analysis of smell.

Referring to Fig. 1, a kind of odor recognition method based on machine sense of smell provided by the embodiment of the present invention, comprises:

S101. Obtain a mixed response signal of the target mixed odor;

S102. Determine whether there is an unknown response signal in the mixed response signal; if yes, execute S103; if not, execute S104;

S103. Filter out the unknown response signal from the mixed response signal to obtain the target response signal, compare the target response signal with the template response signal in the preset database, and determine the odor type and content corresponding to the target response signal;

Specifically, since there are unknown response signals in the mixed response signals, in order to eliminate the influence of the unknown response signals on the odor analysis process, it is necessary to filter out the unknown response signals first. Among them, since different response signals similar to odors have roughly the same response mode collected by the same sensor, the unknown response signal can be represented by a specific function.

For example: the response patterns of these three similar odors of emperor tangerine pulp, navel orange pulp and shatangju pulp are shown in Figure 2. As can be seen from Figure 2, the same sensor has the same odor response pattern, so there is an unknown odor (generated by a single gas When a single odor produced by a single gas component or a compound odor produced by multiple gas components), different unknown odors have different numerical values on each sensor, but the odor response pattern on each sensor must be the same. Based on this, a specific objective function can be optimized to generate a recognition model that takes into account unknown odors.

S104. Compare the mixed response signal with the template response signal in the preset database, and determine the odor type and content corresponding to the mixed response signal.

Specifically, when there is no unknown response signal in the mixed response signal, the mixed response signal can be directly compared with the template response signal in the preset database, so as to determine the odor type and content corresponding to the mixed response signal.

It can be seen that this embodiment provides a method for odor recognition based on machine olfaction. Before processing the mixed response signal of the mixed odor, the method first filters out the unknown response signal in the mixed response signal, so as to avoid the impact of the unknown odor on the recognition result. At the same time, the target response signal obtained after filtering out the unknown response signal is compared with the template response signal in the preset database to determine the type and content of the odor corresponding to the target response signal, thereby improving the recognition efficiency and accuracy , also improves the precision of odor quantitative analysis.

The embodiment of the present invention discloses an odor recognition method based on machine sense of smell. Compared with the previous embodiment, this embodiment further explains and optimizes the technical solution.

Referring to Fig. 3, another odor recognition method based on machine olfaction provided by an embodiment of the present invention includes:

S301. Obtain a mixed response signal of the target mixed odor;

S303. Determine whether there is an unknown response signal in the mixed response signal; if yes, execute S303; if not, execute S307;

S303. Determine whether the mixed response signal is a linear response signal; if yes, execute S304; if not, execute S305;

S304. Use the optimization function in the linear mixed model to filter out the unknown response signal, obtain the target response signal, compare the target response signal with the template response signal in the preset database, and determine the odor type corresponding to the target response signal by the multiplier method and content;

Filter mixing matrix, z(t) represents the unknown response signal, t represents the time variable, and x(tP) represents the noise signal.

Moreover, the multiplier method is used in constrained nonlinear programming, which belongs to the penalty function method. It adds corresponding penalties to the Lagrange function of the constrained problem, which can ensure that the iteration point will converge to the minimum point of the constrained problem, and more importantly It can guarantee the stability of numerical calculation.

It should be noted that the optimization function in the linear mixed model can be obtained based on the following process. The response of a single sensor to a single odor can be described by a finite impulse response filter under linear mixture conditions while considering both the steady-state and transient responses of a machine olfactory sensor array. For example: if there are m sensors and n source signals (source odors), the following convolutional mixed mathematical model can be established:

Among them, y(t), x(t), and n(t) are respectively an m×1 sensor response column vector composed of m sensor responses, an n×1 signal source column vector composed of n signal sources, and An m×1 noise vector composed of the observation noise of m sensors, s represents the length of the impulse response vector, and H represents the filter mixing matrix and has nothing to do with time. It should be noted that in the research and application of machine olfaction, it is assumed that x(t) is basically unchanged during a measurement process (about 30 seconds to 1 minute).

Since the content (or strength) of the signal source cannot be negative, the following objective function can be constructed:

It can be seen from the above formula that it is the least squares method with constraints, and the signal source vector x(t) can be calculated by using the Lagrange multiplier method. However, because there may be collinearity (complete correlation or high correlation) between sensors, the Lagrange multiplier method will face ill-conditioned solution problems. In order to overcome the limitations of the Lagrange multiplier method, the ridge regression algorithm can be used to solve it. The ridge regression formula is as follows:

Among them, X is the source odor vector, H is the transfer matrix, H ^* is the pseudo-inverse matrix of H, Y is the sensor response matrix, and γ (0≤γ≤1) is the regularization parameter, which can be obtained according to cross-validation. However, if there is an unknown odor, since the sensor array response matrix A or the filter response matrix H _p is described for n kinds of source odors in a specific application scenario, and does not contain the response information of the unknown gas, the above optimization objective Solving, there must be a relatively large error.

In order to solve the above problems, according to the different response signals of similar odors, the optimization can be carried out through the principle that the response modes collected by the same sensor are roughly the same, and then the optimal value in the linear mixed model can be obtained. If there is no discrimination, it will be discarded; the z(t) components of the other 9 sensors are solved by one of the following formulas.

Among them, formula (1) is: x _i (t) ≥ 0, x _i (t) represents the noise signal; formula (2) is: z (t) = 0;

Formula (3) includes four solution modes, wherein, formula (3-1) is: z(t)=Ae ^at , formula (3-2) is: z(t)=1-Be ^bt , formula (3- 3) is: z(t)=Cec ^ct , formula (3-4) is: z(t)=kt ² +lt+m.

When the z(t) components of all sensors are equal to 0, it means that there is no unknown odor. Therefore, this functional model can take into account the presence of unknown odors.

Among them, when the response signal increases linearly and gradually reaches a steady state, Ae ^at is used to solve z(t); when the response signal decreases linearly and gradually reaches a steady state, 1-Be ^bt is used to solve z(t); When the response signal is a nonlinear response signal and changes to a pre-stabilized state, use Ce ^ct to solve z(t); when the response signal is a nonlinear response signal and changes to a stable state, use kt ² +lt +m Solve for z(t).

S305. Use the optimization function in the nonlinear mixed model to filter out the unknown response signal, obtain the target response signal, and execute S306;

S306. Compare the target response signal with the template response signal in the preset database, and determine the odor type and content corresponding to the target response signal through a chaotic genetic algorithm.

When the mixed response signal is a nonlinear response signal, the optimization function in the nonlinear mixed model can be used to filter out the unknown response signal to obtain the target response signal; where, the optimized function in the nonlinear mixed model is: min||y(t )-g(t)-z(t)||, y(t) represents a mixed response signal, g(t) represents a nonlinear response signal, and z(t) represents an unknown response signal.

S307. Compare the mixed response signal with the template response signal in the preset database, and determine the odor type and content corresponding to the mixed response signal.

In this embodiment, the nonlinear mixed model is generated using the multivariate adaptive regression spline method, and the specific generation steps are:

Obtain the experimental data of mixed odors, and correct the experimental data for environmental factors;

Divide the corrected experimental data to obtain multiple intervals, and generate corresponding basis functions based on each interval;

The basis function with the highest accuracy is selected by the average error method, and the model corresponding to the basis function with the highest accuracy is determined as a nonlinear mixed model. The smallest result of the equipartition error method corresponds to the basis function with the highest accuracy.

Wherein, the experimental data of the mixed odor can also be used to generate a preset odor database for comparison with the mixed response signal to be identified.

The correction of environmental factors to the experimental data includes: using a neural network method to compensate the temperature, humidity and atmospheric pressure of the experimental data.

For example, in a specific application, a three-layer BP neural network is used for temperature and humidity compensation. The input of the neural network is the sensor response, ambient temperature, ambient humidity, and atmospheric pressure, and the output is the corrected sensor response. In this way, the preset odor database can be perfected (the data stored in the database cannot cover the entire range of temperature, humidity and atmospheric pressure), and the data coverage in the database can be made wider through correction.

Wherein, the base function is generally a linear function or a quadratic function, and the parameters in the function can be corrected through the neural network to improve the accuracy of the base function.

In this embodiment, the type and content of the odor corresponding to the target response signal are determined by a chaotic genetic algorithm, including:

Generate a population of chaotic genetic algorithm based on the target response signal;

Carry out individual evaluation on the individuals in the population, and perform selection operation, crossover operation and mutation operation on the population that has completed the individual evaluation, so that the population can be iteratively updated;

Determine whether the current evolution algebra is greater than the preset maximum evolution algebra;

If so, output the optimal solution, and determine the type and content of the odor corresponding to the target response signal according to the optimal solution.

Specifically, the results obtained by using the chaotic genetic algorithm can be used to estimate the concentration and type of the mixed odor, wherein the accuracy of the estimation result of the concentration can reach 0.01 mg/m ³ , thereby improving the accuracy of the quantitative analysis of the odor.

It can be seen that this embodiment provides another odor recognition method based on machine olfaction. When the method processes the mixed response signal of mixed odor, it first judges whether the mixed response signal is a linear response signal or a nonlinear response signal. When the mixed response signal When it is a linear response signal, use the optimization function in the linear mixed model to filter out the unknown response signal to obtain the target response signal, compare the target response signal with the template response signal in the preset database, and determine the target response signal through the multiplier method Corresponding odor type and content; when the mixed response signal is a nonlinear response signal, use the optimization function in the nonlinear mixed model to filter out the unknown response signal to obtain the target response signal, and compare the target response signal with the template in the preset database The response signals are compared, and the odor type and content corresponding to the target response signal are determined by the chaotic genetic algorithm, thereby improving the recognition efficiency and accuracy, and also improving the accuracy of odor quantitative analysis.

It should be noted that when there is no unknown response signal in the mixed response signal, and the mixed response signal is a linear response signal, it also includes:

Using the optimization function in the linear mixed model to filter out the unknown response signal to obtain the target response signal;

The target response signal is compared with the template response signal in the preset database, and the odor type and content corresponding to the target response signal are determined by a ridge regression method.

Specifically, when there is no unknown response signal in the mixed response signal, and the mixed response signal is a linear response signal, the above-mentioned optimization function and ridge regression method in the linear mixed model can be used to determine the odor type corresponding to the target response signal and content to make the solution process easier and the identification results more accurate. Wherein, since the existing odor identification method can already identify the type and content of the odor, how to identify the type and content of the odor is not described in detail in the present invention. The odor identification method disclosed by the invention is more operable in processing steps, and its processing model is also more optimized, so the obtained identification results are more accurate.

Based on any of the above-mentioned embodiments, it should be noted that the judging whether there is an unknown response signal in the mixed response signal, that is, judging whether there is an odor that does not exist in the preset database in the obtained mixed response signal, and if it exists, it means that the mixed response signal is unknown. There is an unknown response signal in the response signal; if it does not exist, it means that there is no unknown response signal in the mixed response signal, that is, all the signals in the mixed response can be found in the preset database.

The following is an introduction to an odor recognition device based on machine sense of smell provided by an embodiment of the present invention. The device for smell recognition based on machine sense of smell described below and the method for smell recognition based on machine sense of smell described above can be cross-referenced.

Referring to Fig. 4, an odor recognition device based on machine sense of smell provided by an embodiment of the present invention includes:

An acquisition module 401, configured to acquire a mixed response signal of a target mixed odor;

A judging module 402, configured to judge whether there is an unknown response signal in the mixed response signal;

An identification module 403, configured to filter out the unknown response signal from the mixed response signal when there is an unknown response signal in the mixed response signal, obtain a target response signal, and compare the target response signal with a preset database Compare the template response signals in , and determine the odor type and content corresponding to the target response signal.

Wherein, the identification module includes:

a judging unit, configured to judge whether the mixed response signal is a linear response signal;

An identification unit, configured to use an optimization function in a linear mixed model to filter out the unknown response signal when the mixed response signal is a linear response signal to obtain the target response signal, and compare the target response signal with a preset database Compare the template response signals in , and determine the odor type and content corresponding to the target response signal by the multiplier method;

H represents the filter mixing matrix, z(t) represents the unknown response signal, t represents the time variable, and x(tP) represents the noise signal.

Among them, also include:

A linear processing module, configured to use an optimization function in the linear mixed model to filter out the unknown response signal when there is no unknown response signal in the mixed response signal, and the mixed response signal is a linear response signal, to obtain The target response signal; comparing the target response signal with the template response signal in the preset database, and determining the type and content of the odor corresponding to the target response signal through ridge regression method.

Among them, also include:

The nonlinear processing module is used to filter out the unknown response signal by using the optimization function in the nonlinear mixed model to obtain the target response signal when the mixed response signal is a nonlinear response signal; Comparing it with the template response signal in the preset database, and determining the odor type and content corresponding to the target response signal through a chaotic genetic algorithm;

Wherein, the optimization function in the nonlinear mixed model is: min||y(t)-g(t)-z(t)||, y(t) represents the mixed response signal, and g(t) represents the nonlinear Response signal, z(t) represents the unknown response signal.

Wherein, the nonlinear processing module also includes:

generating unit for generating nonlinear mixed models using multivariate adaptive regression splines;

The generating unit includes:

Acquiring sub-units for obtaining experimental data of mixed odors, and correcting the experimental data for environmental factors;

Divide the subunits, which are used to divide the corrected experimental data to obtain multiple intervals, and generate corresponding basis functions based on each interval;

A determining subunit is configured to select the basis function with the highest accuracy by using the average error method, and determine the model corresponding to the basis function with the highest accuracy as the nonlinear mixed model.

Among them, the nonlinear processing module is specifically used for:

generating a population of the chaotic genetic algorithm based on the target response signal;

Carry out individual evaluation on the individuals in the population, and perform selection operation, crossover operation and mutation operation on the population that has completed the individual evaluation, so that the population is iteratively updated;

Determine whether the current evolutionary algebra is greater than the preset maximum evolutionary algebra;

If yes, output the optimal solution, and determine the type and content of the odor corresponding to the target response signal according to the optimal solution.

The following is an introduction to an odor recognition device based on machine sense of smell provided by an embodiment of the present invention. The smell recognition device based on machine sense of smell described below and the method and device for smell recognition based on machine sense of smell described above can be cross-referenced .

Referring to Fig. 5, an embodiment of the present invention provides an odor recognition device based on machine sense of smell, including:

Memory 501, for storing computer programs;

The processor 502 is configured to implement the steps of the odor recognition method based on machine sense of smell described in any of the above embodiments when executing the computer program.

A computer-readable storage medium provided by an embodiment of the present invention is introduced below. The computer-readable storage medium described below and the method, device, and equipment for odor recognition based on machine sense of smell described above can be referred to each other.

A computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the odor recognition method based on machine sense of smell as described in any of the above embodiments are implemented.

Each embodiment in this specification is described in a progressive manner, each embodiment focuses on the difference from other embodiments, and the same and similar parts of each embodiment can be referred to each other.

The above description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention will not be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. A method for odor identification based on machine sense of smell, characterized in that, comprising: Obtain the mixed response signal of the target mixed odor; judging whether there is an unknown response signal in the mixed response signal; If so, filter out the unknown response signal from the mixed response signal to obtain a target response signal, and compare the target response signal with a template response signal in a preset database to determine the target response signal corresponding to Odor type and content. 2. The odor recognition method based on machine sense of smell according to claim 1, wherein the unknown response signal is filtered out from the mixed response signal to obtain a target response signal, and the target response signal is Compared with the template response signal in the preset database, the odor type and content corresponding to the target response signal are determined, including: judging whether the mixed response signal is a linear response signal; If so, use the optimization function in the linear mixed model to filter out the unknown response signal to obtain the target response signal, compare the target response signal with the template response signal in the preset database, and determine by the multiplier method The odor type and content corresponding to the target response signal; H represents the filter mixing matrix, z(t) represents the unknown response signal, t represents the time variable, and x(tP) represents the noise signal. 3. The odor recognition method based on machine sense of smell according to claim 2, wherein, when there is no unknown response signal in the mixed response signal, and when the mixed response signal is a linear response signal, it also includes: Using the optimization function in the linear mixed model to filter out the unknown response signal to obtain the target response signal; The target response signal is compared with the template response signal in the preset database, and the odor type and content corresponding to the target response signal are determined by a ridge regression method. 4. the method for odor recognition based on machine sense of smell according to claim 2, is characterized in that, when described mixed response signal is nonlinear response signal, also comprises: filtering out the unknown response signal by using an optimization function in the nonlinear mixed model to obtain the target response signal; Comparing the target response signal with the template response signal in the preset database, and determining the odor type and content corresponding to the target response signal through a chaotic genetic algorithm; Wherein, the optimization function in the nonlinear mixed model is: min||y(t)-g(t)-z(t)||, y(t) represents the mixed response signal, and g(t) represents the nonlinear Response signal, z(t) represents the unknown response signal. 5. the odor recognition method based on machine sense of smell according to claim 4, is characterized in that, described nonlinear mixed model utilizes multivariate self-adaptive regression spline method to generate, and concrete generation step is: Obtaining experimental data of mixed odors, and correcting the experimental data for environmental factors; Divide the corrected experimental data to obtain multiple intervals, and generate corresponding basis functions based on each interval; The basis function with the highest accuracy is selected by using the average error method, and the model corresponding to the basis function with the highest accuracy is determined as the nonlinear mixed model. 6. The method for identifying odors based on machine sense of smell according to claim 4, wherein said determination of the odor type and content corresponding to said target response signal by a chaotic genetic algorithm comprises: generating a population of the chaotic genetic algorithm based on the target response signal; Carry out individual evaluation on the individuals in the population, and perform selection operation, crossover operation and mutation operation on the population that has completed the individual evaluation, so that the population is iteratively updated; Determine whether the current evolution algebra is greater than the preset maximum evolution algebra; If yes, output the optimal solution, and determine the type and content of the odor corresponding to the target response signal according to the optimal solution. 7. A smell recognition device based on machine sense of smell, characterized in that it comprises: Obtaining module, for obtaining the mixed response signal of target mixed smell; A judging module, configured to judge whether there is an unknown response signal in the mixed response signal; An identification module, configured to filter out the unknown response signal from the mixed response signal when there is an unknown response signal in the mixed response signal, obtain a target response signal, and compare the target response signal with the preset database The template response signals are compared to determine the odor type and content corresponding to the target response signal. 8. The odor identification device based on machine sense of smell according to claim 7, wherein the identification module comprises: a judging unit, configured to judge whether the mixed response signal is a linear response signal; An identification unit, configured to use an optimization function in a linear mixed model to filter out the unknown response signal when the mixed response signal is a linear response signal to obtain the target response signal, and compare the target response signal with a preset database The template response signal in is compared, and the odor type and content corresponding to the target response signal are determined by the multiplier method; H represents the filter mixing matrix, z(t) represents the unknown response signal, t represents the time variable, and x(tP) represents the noise signal. 9. A smell recognition device based on machine sense of smell, characterized in that, comprising: memory for storing computer programs; A processor, configured to implement the steps of the method for smell recognition based on machine olfaction according to any one of claims 1-6 when executing the computer program. 10. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the method based on any one of claims 1-6 is implemented. Steps in an odor recognition method for machine olfaction.