CN108932483A - A kind of odor identification method and device based on machine olfaction - Google Patents
A kind of odor identification method and device based on machine olfaction Download PDFInfo
- Publication number
- CN108932483A CN108932483A CN201810628801.0A CN201810628801A CN108932483A CN 108932483 A CN108932483 A CN 108932483A CN 201810628801 A CN201810628801 A CN 201810628801A CN 108932483 A CN108932483 A CN 108932483A
- Authority
- CN
- China
- Prior art keywords
- response signal
- target
- hybrid
- unknown
- odor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/08—Computing arrangements based on specific mathematical models using chaos models or non-linear system models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Physics (AREA)
- Evolutionary Biology (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- Genetics & Genomics (AREA)
- Molecular Biology (AREA)
- Physiology (AREA)
- Signal Processing (AREA)
- Computational Linguistics (AREA)
- Nonlinear Science (AREA)
- Algebra (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Investigating Or Analyzing Materials By The Use Of Fluid Adsorption Or Reactions (AREA)
- Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
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
Technical field
The present invention relates to field of artificial intelligence, more specifically to a kind of odor identification based on machine olfaction
Method, apparatus, equipment and computer readable storage medium.
Background technique
Currently, machine olfaction technology using more and more extensive, in industrial gasses monitoring, quality of chemical products monitoring, room
All various aspects such as internal and external environment monitoring, medical diagnosis, Chinese medical herb, have a large amount of application case.
In these application fields, monitored smell mostly mixes smell, and the odiferous information in electronic nose database
It is limited, so certainly existing unknown gas.Also, existing odor identification technology is when handling the response signal of sensor, not
Influence in view of unknown smell to recognition result, it is larger so as to cause recognition result error, reduce the efficiency, accurate of identification
The precision of degree and smell quantitative analysis.
Therefore, the efficiency of odor identification, the precision of accuracy and smell quantitative analysis how are improved, is art technology
Personnel's problem to be solved.
Summary of the invention
The purpose of the present invention is to provide a kind of odor identification method, apparatus, equipment and computer based on machine olfaction
Readable storage medium storing program for executing, to improve the efficiency of odor identification, the precision of accuracy and smell quantitative analysis.
To achieve the above object, the embodiment of the invention provides following technical solutions:
A kind of odor identification method based on machine olfaction, comprising:
Obtain the hybrid response signal of target mixing smell；
Judge in the hybrid response signal with the presence or absence of unknown response signal；
If so, filtering out the unknown response signal from the hybrid response signal, target response signal is obtained, and will
The target response signal is compared with the template response signal in presetting database, determines that the target response signal is corresponding
Odor types and content.
Wherein, described that the unknown response signal is filtered out from the hybrid response signal, target response signal is obtained, and
The target response signal and the template response signal in presetting database are compared, determine the target response signal pair
The odor types and content answered, comprising:
Judge whether the hybrid response signal is linear response signal；
If so, filtering out the unknown response signal using the majorized function in linear mixed model, obtain
The target response signal, by the template response signal in the target response signal and presetting database
It compares, and the corresponding odor types of the target response signal and content is determined by multiplier method；
H indicates that filter hybrid matrix, z (t) indicate unknown response signal, and t indicates time variable, and x (t-P) indicates noise signal.
Wherein, when unknown response signal is not present in the hybrid response signal, and the hybrid response signal is linear
When response signal, further includes:
The unknown response signal is filtered out using the majorized function in the linear mixed model, obtains the target response
Signal；
The target response signal and the template response signal in presetting database are compared, Ridge Regression Modeling Method is passed through
Determine the corresponding odor types of the target response signal and content.
Wherein, when the hybrid response signal is nonlinear response signal, further includes:
The unknown response signal is filtered out using the majorized function in nonlinear mixed model, obtains the target response letter
Number；
The target response signal and the template response signal in presetting database are compared, and pass through Chaos-Genetic
Algorithm determines the corresponding odor types of the target response signal and content；
Wherein, the majorized function in the nonlinear mixed model are as follows: min | | y (t)-g (t)-z (t) | |, y (t) is indicated
Hybrid response signal, g (t) indicate that nonlinear response signal, z (t) indicate unknown response signal.
Wherein, the nonlinear mixed model is generated using Multivariate adaptive regression splines Spline Method, specific generation step are as follows:
The experimental data of mixing smell is obtained, and environmental factor correction is carried out to the experimental data；
Experimental data after correction is divided, obtains multiple sections, and corresponding base letter is generated based on each section
Number；
The highest basic function of accuracy is selected using the respectively theory of error, and the highest basic function of the accuracy is corresponding
Model is determined as the nonlinear mixed model.
Wherein, described that the corresponding odor types of the target response signal and content are determined by Chaos Genetic Algorithm, packet
It includes:
The population of the Chaos Genetic Algorithm is generated based on the target response signal；
Individual evaluation is carried out to the individual in the population, and Selecting operation is carried out to the population for completing individual evaluation, is handed over
Operation and mutation operator are pitched, the population is made to be iterated update；
Judge whether current evolutionary generation is greater than preset maximum evolutionary generation；
If so, output optimal solution, and the corresponding odor types of the target response signal are determined according to the optimal solution
And content.
A kind of odor identification device based on machine olfaction, comprising:
Module is obtained, for obtaining the hybrid response signal of target mixing smell；
Judgment module, for judging in the hybrid response signal with the presence or absence of unknown response signal；
Identification module, for when there are when unknown response signal, believing from the hybrid response in the hybrid response signal
The unknown response signal is filtered out in number, obtains target response signal, and will be in the target response signal and presetting database
Template response signal compare, determine the corresponding odor types of the target response signal and content.
Wherein, the identification module includes:
Judging unit, for judging whether the hybrid response signal is linear response signal；
Recognition unit is used for when the hybrid response signal is linear response signal, using in linear mixed model
Majorized function filters out the unknown response signal, obtains the target response signal, by the target response signal and present count
Compared according to the template response signal in library, and by multiplier method determine the corresponding odor types of the target response signal and
Content；
H indicates that filter hybrid matrix, z (t) indicate unknown response signal, and t indicates time variable, and x (t-P) indicates noise signal.
A kind of odor identification equipment based on machine olfaction, comprising:
Memory, for storing computer program；
Processor realizes the gas described in above-mentioned any one based on machine olfaction when for executing the computer program
The step of taste recognition methods.
A kind of computer readable storage medium is stored with computer program on the computer readable storage medium, described
The step of the odor identification method described in above-mentioned any one based on machine olfaction is realized when computer program is executed by processor
Suddenly.
By above scheme it is found that a kind of odor identification method based on machine olfaction provided in an embodiment of the present invention, packet
It includes: obtaining the hybrid response signal of target mixing smell；Judge in the hybrid response signal with the presence or absence of unknown response signal；
If so, filter out the unknown response signal from the hybrid response signal, target response signal is obtained, and by the target
Response signal is compared with the template response signal in presetting database, determines the corresponding smell class of the target response signal
Type and content.
As it can be seen that the method filters out in hybrid response signal first before the hybrid response signal of processing mixing smell
Unknown response signal, recognition result is had an impact to avoid unknown smell；Meanwhile it being obtained after unknown response signal will be filtered out
Target response signal compared with the template response signal in presetting database, so that it is determined that target response signal is corresponding out
Odor types and content, which thereby enhance recognition efficiency and accuracy, also improve the precision of smell quantitative analysis.
Correspondingly, a kind of odor identification device based on machine olfaction, equipment and computer provided in an embodiment of the present invention
Readable storage medium storing program for executing similarly has above-mentioned technique effect.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of odor identification method flow diagram based on machine olfaction disclosed by the embodiments of the present invention；
Fig. 2 is the response modes schematic diagram of three kinds disclosed by the embodiments of the present invention similar smells:
Fig. 3 is another odor identification method flow diagram based on machine olfaction disclosed by the embodiments of the present invention；
Fig. 4 is a kind of odor identification device schematic diagram based on machine olfaction disclosed by the embodiments of the present invention；
Fig. 5 is a kind of odor identification equipment schematic diagram based on machine olfaction disclosed by the embodiments of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The embodiment of the invention discloses a kind of, and odor identification method, apparatus, equipment and computer based on machine olfaction can
Storage medium is read, to improve the efficiency of odor identification, the precision of accuracy and smell quantitative analysis.
Referring to Fig. 1, a kind of odor identification method based on machine olfaction provided in an embodiment of the present invention, comprising:
S101, the hybrid response signal for obtaining target mixing smell；
S102, judge in hybrid response signal with the presence or absence of unknown response signal；If so, executing S103；If it is not, then holding
Row S104；
S103, unknown response signal is filtered out from hybrid response signal, obtain target response signal, and target response is believed
It number is compared with the template response signal in presetting database, determines the corresponding odor types of target response signal and content；
Specifically, since there are unknown response signals in hybrid response signal, in order to eliminate unknown response signal to smell
The influence of analytic process needs to filter out unknown response signal first.Wherein, due to the different response signals of similar smell, pass through
The same collected response modes of sensor are roughly the same, therefore can be indicated unknown response signal using specific function.
Such as: response modes such as Fig. 2 institute of emperor mandarin orange pulp, navel orange pulp with Citrus shatangju pulp these three similar smells
Show, as seen from Figure 2, same sensor smell response modes having the same, thus there are unknown smell (by pure gas at
Divide the single smell generated or the compound smell generated by multiple gases ingredient) when, different unknown smells are on each sensor
The numerical values recited of generation is different, but smell response modes on each sensor must be identical.Based on this, can incite somebody to action
Specific objective function optimizes, to generate the identification model for taking into account unknown smell.
S104, hybrid response signal and the template response signal in presetting database are compared, determines hybrid response
The corresponding odor types of signal and content.
Specifically, when unknown response signal is not present in hybrid response signal, then it can be directly by hybrid response signal
It is compared with the template response signal in presetting database, so that it is determined that the corresponding odor types of hybrid response signal and containing
Amount.
As it can be seen that present embodiments providing a kind of odor identification method based on machine olfaction, the method is mixed in processing
Before the hybrid response signal of smell, the unknown response signal in hybrid response signal is filtered out, first to avoid unknown smell pair
Recognition result has an impact；Meanwhile it will filter out in the target response signal obtained after unknown response signal and presetting database
Template response signal compares, so that it is determined that the corresponding odor types of target response signal and content out, which thereby enhances knowledge
Other efficiency and accuracy, also improve the precision of smell quantitative analysis.
The odor identification method based on machine olfaction that the embodiment of the invention discloses a kind of, relative to a upper embodiment, originally
Embodiment has made further instruction and optimization to technical solution.
Referring to Fig. 3, odor identification method of the another kind based on machine olfaction provided in an embodiment of the present invention, comprising:
S301, the hybrid response signal for obtaining target mixing smell；
S303, judge in hybrid response signal with the presence or absence of unknown response signal；If so, executing S303；If it is not, then holding
Row S307；
S303, judge whether hybrid response signal is linear response signal；If so, executing S304；If it is not, then executing
S305；
S304, unknown response signal is filtered out using the majorized function in linear mixed model, obtain target response signal, it will
Target response signal is compared with the template response signal in presetting database, and determines target response signal by multiplier method
Corresponding odor types and content；
Filter hybrid matrix, z (t) indicate unknown response signal, and t indicates time variable, and x (t-P) indicates noise signal.
Also, multiplier method is used for constrained Non-Linear Programming, belongs to penalty function method, is the Lagrange in restricted problem
Corresponding punishment is added in function, both can guarantee that iteration point can converge to the minimal point of restricted problem, it is often more important that can guarantee
The stability that numerical value calculates.
It should be noted that the majorized function in the linear mixed model can be obtained based on following processes.Linearly
When considering the steady-state response and transient response of machine olfaction sensor array under mixing condition and simultaneously, single sensor is to single
The response of smell can be described with finite impulse response filter.Such as: m sensor, n source signal (source gas if it exists
Taste), then it can establish following convolved mixtures mathematical model:
Wherein, y (t), x (t), n (t) are made of the response of m sensor respectively the sensor of m × 1 response column vector, by
The signal source column vector of n × 1 of n signal source composition, m × 1 being made of the observation noise of m sensor noise vector, s
Indicate the length of impulse response vector, H indicates filter hybrid matrix and is unrelated with the time.It should be noted that being smelt in machine
Feel in research and application, all hypothesis x (t) are basically unchanged in one-shot measurement process (- 1 minute about 30 seconds).
Since the content (or intensity) of signal source cannot be negative, therefore following objective function can be constructed:
It can be seen that by above formula, be the least square method of with constraint conditions, letter can be calculated with Lagrange multiplier method
Number source vector x (t).But since there may be synteny (perfectly correlated or highly relevant) between sensor, so Lagrange multiplies
Sub- method can face ill Solve problems.For the limitation for overcoming Lagrange multiplier method, the solution of ridge regression algorithm can be used.Ridge is returned
Formula of making a public possession is as follows:
Wherein, X is source gas taste vector, and H is transfer matrix, H*For the pseudo inverse matrix of H, Y is sensor response matrix, γ (0
≤ γ≤1) it is regularization parameter, it can be obtained according to cross validation.But if wherein there is unknown smell, due to sensor array
Column response matrix A or filter response matrix HpIt is to be described for the n provenance smell of specific application scene, without not
Know the response message of gas, so solving using above-mentioned optimum target, certainly exists bigger error.
To solve the above-mentioned problems, it can be acquired according to the different response signals of similar smell by same sensor
To the roughly the same principle of response modes optimize, and then obtain excellent in linear mixed model
Do not have discrimination, will give up；Z (t) component of other 9 sensors takes one of following formula to be solved.
Wherein, formula (1) are as follows: xi(t) >=0, xi(t) noise signal is indicated；Formula (2) are as follows: z (t)=0；
Formula (3) includes four kinds of Solution models, wherein formula (3-1) are as follows: z (t)=Aeat, formula (3-2) are as follows: z (t)=1-
Bebt, formula (3-3) are as follows: z (t)=Cect, formula (3-4) are as follows: z (t)=kt2+lt+m。
When all the sensors z (t) component be equal to 0, then imply the absence of unknown smell.Therefore, which can
Take into account unknown smell there are the case where.
Wherein, when response signal linearly increases, and progressively reaches stable state, using AeatIt solves z (t)；When response is believed
It is number linear to reduce, and when progressively reaching stable state, using 1-BebtIt solves z (t)；When response signal be nonlinear response signal,
And when mutation to state before stabilization, using CectIt solves z (t)；When response signal is nonlinear response signal, and it is mutated to steady
When state after fixed, using kt2+ lt+m solves z (t).
S305, unknown response signal is filtered out using the majorized function in nonlinear mixed model, obtain target response signal,
And execute S306；
S306, target response signal and the template response signal in presetting database are compared, and is lost by chaos
Propagation algorithm determines the corresponding odor types of target response signal and content.
When hybrid response signal is nonlinear response signal, filtered out using the majorized function in nonlinear mixed model
Unknown response signal obtains target response signal；Wherein, the majorized function in nonlinear mixed model are as follows: min | | y (t)-g
(t)-z (t) | |, y (t) indicates that hybrid response signal, g (t) indicate that nonlinear response signal, z (t) indicate unknown response signal.
S307, hybrid response signal and the template response signal in presetting database are compared, determines hybrid response
The corresponding odor types of signal and content.
In the present embodiment, nonlinear mixed model is generated using Multivariate adaptive regression splines Spline Method, specific to generate step
Suddenly are as follows:
The experimental data of mixing smell is obtained, and environmental factor correction is carried out to experimental data；
Experimental data after correction is divided, obtains multiple sections, and corresponding base letter is generated based on each section
Number；
The highest basic function of accuracy is selected using the respectively theory of error, and by the corresponding model of the highest basic function of accuracy
It is determined as nonlinear mixed model.Respectively the minimum result of the theory of error corresponds to the highest basic function of accuracy.
Wherein, the experimental data of the mixing smell can be also used for generating preset odor data library, for it is to be identified
Hybrid response signal compare.
It is described that carry out environmental factor correction to experimental data include: the method using neural network to the temperature of experimental data
Degree, humidity and atmospheric pressure compensate.
Such as: under specific application, temperature and humidity compensation, the input of neural network are carried out using three layers of BP neural network
For sensor response, environment temperature, ambient humidity and atmospheric pressure, export as the sensor response after correction.It so can be perfect
Preset odor data library (data that database is deposited can not cover entire temperature, humidity and atmospheric pressure range), passes through
Correction can make the data coverage in database more extensive.
Wherein, the basic function is generally linear function or Quadratic Function Optimization, can be to the ginseng in function by neural network
Number is modified, to improve the accuracy of basic function.
In the present embodiment, the corresponding odor types of target response signal and content are determined by Chaos Genetic Algorithm, wrapped
It includes:
The population of Chaos Genetic Algorithm is generated based on target response signal；
Individual evaluation is carried out to the individual in population, and Selecting operation is carried out to the population for completing individual evaluation, intersects and transports
Calculation and mutation operator, make population be iterated update；
Judge whether current evolutionary generation is greater than preset maximum evolutionary generation；
If so, output optimal solution, and the corresponding odor types of target response signal and content are determined according to optimal solution.
Specifically, can be used to the concentration and type estimating to mix smell using the result that Chaos Genetic Algorithm obtains, wherein
The accuracy of the estimation result of concentration is up to 0.01mg/m3, to improve the precision of smell quantitative analysis.
As it can be seen that present embodiments providing another odor identification method based on machine olfaction, the method is mixed in processing
Close smell hybrid response signal when, first determine whether hybrid response signal be linear response signal or nonlinear response signal,
When hybrid response signal is linear response signal, unknown response signal is filtered out using the majorized function in linear mixed model,
Target response signal is obtained, target response signal and the template response signal in presetting database are compared, and by multiplying
Sub- method determines the corresponding odor types of target response signal and content；When hybrid response signal is nonlinear response signal, benefit
Unknown response signal is filtered out with the majorized function in nonlinear mixed model, obtains target response signal, and target response is believed
It number is compared with the template response signal in presetting database, and determines that target response signal is corresponding by Chaos Genetic Algorithm
Odor types and content, which thereby enhance recognition efficiency and accuracy, also improve the precision of smell quantitative analysis.
It should be noted that when unknown response signal is not present in the hybrid response signal, and the hybrid response is believed
Number be linear response signal when, further includes:
The unknown response signal is filtered out using the majorized function in the linear mixed model, obtains the target response
Signal；
The target response signal and the template response signal in presetting database are compared, Ridge Regression Modeling Method is passed through
Determine the corresponding odor types of the target response signal and content.
Specifically, when unknown response signal is not present in hybrid response signal, and the hybrid response signal is linear rings
When induction signal, majorized function and Ridge Regression Modeling Method in above-mentioned linear mixed model can be used and determine the target response signal pair
The odor types and content answered, so that solution procedure is more easy, recognition result is more accurate.Wherein, due to existing gas
Taste recognition methods can recognize that the type and content of smell, so type and content for how to identify smell, this hair
It is bright not elaborate.Odor identification method disclosed by the invention has more operability in processing step, handles model
It is more optimized, therefore the recognition result obtained is more acurrate.
Based on above-mentioned any embodiment, it should be noted that whether there is unknown sound in the judgement hybrid response signal
Induction signal judges in the hybrid response signal got with the presence or absence of the smell being not present in presetting database, and if it exists, then
Illustrate that there are unknown response signals in hybrid response signal；If it does not exist, then illustrate that there is no unknown sound in hybrid response signal
Induction signal, i.e., all signals in hybrid response can be found in the preset database.
A kind of odor identification device based on machine olfaction provided in an embodiment of the present invention is introduced below, is hereafter retouched
A kind of odor identification device based on machine olfaction stated and a kind of above-described odor identification method based on machine olfaction
It can be cross-referenced.
Referring to fig. 4, a kind of odor identification device based on machine olfaction provided in an embodiment of the present invention, comprising:
Module 401 is obtained, for obtaining the hybrid response signal of target mixing smell；
Judgment module 402, for judging in the hybrid response signal with the presence or absence of unknown response signal；
Identification module 403, for when in the hybrid response signal there are when unknown response signal, from the hybrid response
It filters out the unknown response signal in signal, obtains target response signal, and by the target response signal and presetting database
In template response signal compare, determine the corresponding odor types of the target response signal and content.
Wherein, the identification module includes:
Judging unit, for judging whether the hybrid response signal is linear response signal；
Recognition unit is used for when the hybrid response signal is linear response signal, using in linear mixed model
Majorized function filters out the unknown response signal, obtains the target response signal, by the target response signal and present count
Compared according to the template response signal in library, and by multiplier method determine the corresponding odor types of the target response signal and
Content；
H indicates that filter hybrid matrix, z (t) indicate unknown response signal, and t indicates time variable, and x (t-P) indicates noise signal.
Wherein, further includes:
Unknown response signal is not present for working as in Linear processing module in the hybrid response signal, and the mixing is rung
When induction signal is linear response signal, the unknown response signal is filtered out using the majorized function in the linear mixed model,
Obtain the target response signal；The target response signal and the template response signal in presetting database are compared,
The corresponding odor types of the target response signal and content are determined by Ridge Regression Modeling Method.
Wherein, further includes:
Nonlinear processing module, for being mixed using non-linear when the hybrid response signal is nonlinear response signal
Majorized function in molding type filters out the unknown response signal, obtains the target response signal；The target response is believed
It number is compared with the template response signal in presetting database, and the target response signal is determined by Chaos Genetic Algorithm
Corresponding odor types and content；
Wherein, the majorized function in the nonlinear mixed model are as follows: min | | y (t)-g (t)-z (t) | |, y (t) is indicated
Hybrid response signal, g (t) indicate that nonlinear response signal, z (t) indicate unknown response signal.
Wherein, the Nonlinear processing module further include:
Generation unit, for generating nonlinear mixed model using Multivariate adaptive regression splines Spline Method；
The generation unit includes:
Subelement is obtained, carries out environmental factor school for obtaining the experimental data of mixing smell, and to the experimental data
Just；
Subelement is divided, for dividing to the experimental data after correction, obtains multiple sections, and be based on each section
Generate corresponding basic function；
Subelement is determined, for selecting the highest basic function of accuracy using the respectively theory of error, and most by the accuracy
The corresponding model of high basic function is determined as the nonlinear mixed model.
Wherein, Nonlinear processing module is specifically used for:
The population of the Chaos Genetic Algorithm is generated based on the target response signal；
Individual evaluation is carried out to the individual in the population, and Selecting operation is carried out to the population for completing individual evaluation, is handed over
Operation and mutation operator are pitched, the population is made to be iterated update；
Judge whether current evolutionary generation is greater than preset maximum evolutionary generation；
If so, output optimal solution, and the corresponding odor types of the target response signal are determined according to the optimal solution
And content.
A kind of odor identification equipment based on machine olfaction provided in an embodiment of the present invention is introduced below, is hereafter retouched
A kind of odor identification equipment based on machine olfaction and a kind of above-described odor identification method based on machine olfaction stated
And device can be cross-referenced.
Referring to Fig. 5, a kind of odor identification equipment based on machine olfaction provided in an embodiment of the present invention, comprising:
Memory 501, for storing computer program；
Processor 502 is realized when for executing the computer program and is smelt described in above-mentioned any embodiment based on machine
The step of odor identification method of feel.
A kind of computer readable storage medium provided in an embodiment of the present invention is introduced below, one kind described below
Computer readable storage medium can phase with a kind of above-described odor identification method, device and equipment based on machine olfaction
Mutual reference.
A kind of computer readable storage medium is stored with computer program on the computer readable storage medium, described
The odor identification method based on machine olfaction as described in above-mentioned any embodiment is realized when computer program is executed by processor
The step of.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.
The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention.
Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, of the invention
It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one
The widest scope of cause.
Claims (10)
1. a kind of odor identification method based on machine olfaction characterized by comprising
Obtain the hybrid response signal of target mixing smell；
Judge in the hybrid response signal with the presence or absence of unknown response signal；
If so, filtering out the unknown response signal from the hybrid response signal, target response signal is obtained, and will be described
Target response signal is compared with the template response signal in presetting database, determines the corresponding gas of the target response signal
Taste type and content.
2. the odor identification method according to claim 1 based on machine olfaction, which is characterized in that described from the mixing
It filters out the unknown response signal in response signal, obtains target response signal, and by the target response signal and present count
It is compared according to the template response signal in library, determines the corresponding odor types of the target response signal and content, comprising:
Judge whether the hybrid response signal is linear response signal；
If so, filtering out the unknown response signal using the majorized function in linear mixed model, the target response is obtained
Signal compares the target response signal and the template response signal in presetting database, and is determined by multiplier method
The corresponding odor types of the target response signal and content；
H indicates that filter hybrid matrix, z (t) indicate unknown response signal, and t indicates time variable, and x (t-P) indicates noise signal.
3. the odor identification method according to claim 2 based on machine olfaction, which is characterized in that when the hybrid response
In signal be not present unknown response signal, and the hybrid response signal be linear response signal when, further includes:
The unknown response signal is filtered out using the majorized function in the linear mixed model, obtains the target response letter
Number；
The target response signal and the template response signal in presetting database are compared, determined by Ridge Regression Modeling Method
The corresponding odor types of the target response signal and content.
4. the odor identification method according to claim 2 based on machine olfaction, which is characterized in that when the hybrid response
When signal is nonlinear response signal, further includes:
The unknown response signal is filtered out using the majorized function in nonlinear mixed model, obtains the target response signal；
The target response signal and the template response signal in presetting database are compared, and pass through Chaos Genetic Algorithm
Determine the corresponding odor types of the target response signal and content；
Wherein, the majorized function in the nonlinear mixed model are as follows: min | | y (t)-g (t)-z (t) | |, y (t) indicates mixing
Response signal, g (t) indicate that nonlinear response signal, z (t) indicate unknown response signal.
5. the odor identification method according to claim 4 based on machine olfaction, which is characterized in that the non-linear mixing
Model is generated using Multivariate adaptive regression splines Spline Method, specific generation step are as follows:
The experimental data of mixing smell is obtained, and environmental factor correction is carried out to the experimental data；
Experimental data after correction is divided, obtains multiple sections, and corresponding basic function is generated based on each section；
The highest basic function of accuracy is selected using the respectively theory of error, and by the corresponding model of the highest basic function of the accuracy
It is determined as the nonlinear mixed model.
6. the odor identification method according to claim 4 based on machine olfaction, which is characterized in that described to be lost by chaos
Propagation algorithm determines the corresponding odor types of the target response signal and content, comprising:
The population of the Chaos Genetic Algorithm is generated based on the target response signal；
Individual evaluation is carried out to the individual in the population, and Selecting operation is carried out to the population for completing individual evaluation, intersects and transports
Calculation and mutation operator, make the population be iterated update；
Judge whether current evolutionary generation is greater than preset maximum evolutionary generation；
If so, output optimal solution, and determine the corresponding odor types of the target response signal according to the optimal solution and contain
Amount.
7. a kind of odor identification device based on machine olfaction characterized by comprising
Module is obtained, for obtaining the hybrid response signal of target mixing smell；
Judgment module, for judging in the hybrid response signal with the presence or absence of unknown response signal；
Identification module, for when in the hybrid response signal there are when unknown response signal, from the hybrid response signal
It filters out the unknown response signal, obtains target response signal, and by the mould in the target response signal and presetting database
Plate response signal compares, and determines the corresponding odor types of the target response signal and content.
8. the odor identification device according to claim 7 based on machine olfaction, which is characterized in that the identification module packet
It includes:
Judging unit, for judging whether the hybrid response signal is linear response signal；
Recognition unit, for utilizing the optimization in linear mixed model when the hybrid response signal is linear response signal
Function filters out the unknown response signal, obtains the target response signal, by the target response signal and presetting database
In template response signal compare, and the corresponding odor types of the target response signal are determined by multiplier method and are contained
Amount；
H indicates that filter hybrid matrix, z (t) indicate unknown response signal, and t indicates time variable, and x (t-P) indicates noise signal.
9. a kind of odor identification equipment based on machine olfaction characterized by comprising
Memory, for storing computer program；
Processor, realization is as claimed in any one of claims 1 to 6 when for executing the computer program is smelt based on machine
The step of odor identification method of feel.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program is realized as claimed in any one of claims 1 to 6 based on machine olfaction when the computer program is executed by processor
Odor identification method the step of.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810628801.0A CN108932483B (en) | 2018-06-19 | 2018-06-19 | Smell identification method and device based on machine olfaction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810628801.0A CN108932483B (en) | 2018-06-19 | 2018-06-19 | Smell identification method and device based on machine olfaction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108932483A true CN108932483A (en) | 2018-12-04 |
| CN108932483B CN108932483B (en) | 2022-03-25 |
Family
ID=64446013
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810628801.0A Active CN108932483B (en) | 2018-06-19 | 2018-06-19 | Smell identification method and device based on machine olfaction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108932483B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110057975A (en) * | 2019-04-22 | 2019-07-26 | 广东工业大学 | A kind of olfactory descriptor acquiring method, system and relevant apparatus |
| CN111024776A (en) * | 2019-12-19 | 2020-04-17 | 安徽康佳同创电器有限公司 | Electronic nose, smell identification method and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1482453A (en) * | 2003-07-11 | 2004-03-17 | 华东理工大学 | Machine olfaction odor distinguishing method based on modularized composite neural net |
| US20090048689A1 (en) * | 2007-08-14 | 2009-02-19 | David Andrew Pelton | Actuation response oscillation detection monitor |
| CN101470121A (en) * | 2007-12-29 | 2009-07-01 | 广东工业大学 | Built-in bionic smell recognition method and device |
| CN101788517A (en) * | 2010-02-23 | 2010-07-28 | 广东工业大学 | Pungent traditional Chinese medicine odor fingerprint map construction system and method based on bionic olfaction |
| CN104460948A (en) * | 2013-09-15 | 2015-03-25 | 南京大五教育科技有限公司 | Olfactory sensing technology |
| CN108123724A (en) * | 2017-12-22 | 2018-06-05 | 西安烽火电子科技有限责任公司 | A kind of communication system based on shortwave narrowband waveform |
-
2018
- 2018-06-19 CN CN201810628801.0A patent/CN108932483B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1482453A (en) * | 2003-07-11 | 2004-03-17 | 华东理工大学 | Machine olfaction odor distinguishing method based on modularized composite neural net |
| US20090048689A1 (en) * | 2007-08-14 | 2009-02-19 | David Andrew Pelton | Actuation response oscillation detection monitor |
| CN101470121A (en) * | 2007-12-29 | 2009-07-01 | 广东工业大学 | Built-in bionic smell recognition method and device |
| CN101788517A (en) * | 2010-02-23 | 2010-07-28 | 广东工业大学 | Pungent traditional Chinese medicine odor fingerprint map construction system and method based on bionic olfaction |
| CN104460948A (en) * | 2013-09-15 | 2015-03-25 | 南京大五教育科技有限公司 | Olfactory sensing technology |
| CN108123724A (en) * | 2017-12-22 | 2018-06-05 | 西安烽火电子科技有限责任公司 | A kind of communication system based on shortwave narrowband waveform |
Non-Patent Citations (4)
| Title |
|---|
| BILIANG XIA ET AL: "Classification of Fruit Species Based on Effective Sensor Array Selection", 《2016 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS-CHINA (ICCE-CHINA)》 * |
| ELIZABETH A. BALDWIN ET AL: "Electronic Noses and Tongues: Applications for the Food and Pharmaceutical Industries", 《SENSORS》 * |
| 彭柯 等: "基于机器嗅觉的柑橘品种无损检测与识别", 《江西农业大学学报》 * |
| 高大启: "机器嗅觉在气味类别与浓度同时估计中的应用", 《2004年中国香料香精学术研讨会》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110057975A (en) * | 2019-04-22 | 2019-07-26 | 广东工业大学 | A kind of olfactory descriptor acquiring method, system and relevant apparatus |
| CN111024776A (en) * | 2019-12-19 | 2020-04-17 | 安徽康佳同创电器有限公司 | Electronic nose, smell identification method and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108932483B (en) | 2022-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Baresel et al. | Fitness function design to improve evolutionary structural testing | |
| Li et al. | Auxiliary model based least squares iterative algorithms for parameter estimation of bilinear systems using interval-varying measurements | |
| CN105637439B (en) | The system and method for data acquisition | |
| CN108932483A (en) | A kind of odor identification method and device based on machine olfaction | |
| CN112817299A (en) | Industrial wastewater treatment data management cloud platform and control method thereof | |
| Fithian et al. | Semiparametric exponential families for heavy-tailed data | |
| CN113695366B (en) | Intelligent screening method and system for fly ash chelating agent based on environmental monitoring | |
| Go et al. | Machine learning applied to sensor data analysis | |
| Xue et al. | An optimal data-driven approach to distribution independent fault detection | |
| Parag et al. | Exact Bayesian inference for phylogenetic birth-death models | |
| CN117934247A (en) | Carbon emission factor prediction method and device based on time sequence decomposition, electronic equipment and storage medium | |
| Kalemkerian | Modelling and parameter estimation for discretely observed fractional iterated ornstein–uhlenbeck processes | |
| CN105277343B (en) | Method and device for detecting service life of filter element | |
| Veyna et al. | Parameter uncertainty modeling for multiobjective robust control design. Application to a temperature control system in a proton exchange membrane fuel cell | |
| JP7228599B2 (en) | X-ray analysis device, X-ray analysis system, analysis method, and program | |
| CN112598534A (en) | Method and device for evaluating correlation between power consumption and total domestic production value | |
| CN117648983A (en) | Time sequence cause and effect discovery method, device, equipment and medium based on intervention data | |
| Uematsu et al. | Closed-loop identification of plant and disturbance models based on data-driven generalized minimum variance regulatory control | |
| CN115965149A (en) | Water quality index prediction method based on LSTM algorithm model | |
| CN115102790A (en) | Big data based network flow abnormity sensing system and method | |
| Zajic | A Hammerstein-bilinear approach with application to heating ventilation and air conditioning systems | |
| Sandhu et al. | A comparative analysis of fuzzy, neuro-fuzzy and fuzzy-GA based approaches for software reusability evaluation | |
| CN114168654A (en) | Model establishing method, abnormality detecting apparatus, abnormality detecting medium, and abnormality detecting device | |
| Rodriguez-Brenes et al. | Beyond the pair approximation: Modeling colonization population dynamics | |
| CN111190045A (en) | Voltage abnormity prediction method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |