WO2020090018A1

WO2020090018A1 - Odor-sensor inspection device, odor-sensor inspection method, and computer-readable recording medium

Info

Publication number: WO2020090018A1
Application number: PCT/JP2018/040417
Authority: WO
Inventors: 鈴木　亮太; 山田　聡; ひろみ清水; 江藤　力; 純子渡辺
Original assignee: 日本電気株式会社
Priority date: 2018-10-30
Filing date: 2018-10-30
Publication date: 2020-05-07
Also published as: JPWO2020090018A1; JP7205548B2

Abstract

An odor-sensor inspection device 1 for precisely detecting a defect in an odor sensor includes: a calculation unit 2 that calculates a coefficient on the basis of first odor data indicating a first odor and second odor data obtained as a result of the odor sensor measuring the first odor; and an inspection unit 3 that inspects an odor sensor 21b on the basis of the coefficient.

Description

Odor sensor inspection device, odor sensor inspection method, and computer-readable recording medium

The present invention relates to an odor sensor inspection device and an odor sensor inspection method for inspecting an odor sensor, and further to a computer-readable recording medium recording a program for realizing these.

Non-Patent Document 1 discloses an odor sensor provided with a plurality of sensor elements. Specifically, these sensor elements are provided with a sensitive film having different characteristics for each sensor element. In addition, the sensor element is configured to have a unique reaction with respect to the molecule adsorbed on the sensitive film.

However, in the odor sensor having the above-described sensitive film, there are individual differences among the odor sensors depending on the state of the sensitive film after manufacturing, so when measuring odors using different odor sensors, a measurement error occurs between the odor sensors. Occurs.

The state of the sensitive film is, for example, the shape, size, thickness, etc. of the sensitive film. The cause of manufacturing variations is, for example, how the coating is applied to a support member that supports the coating liquid when the coating liquid is used for manufacturing the sensitive film.

If there is a manufacturing defect in the sensitive film, odor analysis cannot be performed accurately, so the accuracy of identifying odor and the accuracy of measuring odor intensity will decrease. Therefore, it is necessary to discard, re-apply, and re-inspect the odor sensor, and therefore it is desired to develop an inspection method that accurately detects that the sensitive film is defective in manufacturing.

One example of an object of the present invention is to provide an odor sensor inspection device, an odor sensor inspection method, and a computer-readable recording medium that detect a defect of an odor sensor with high accuracy.

In order to achieve the above object, the odor sensor inspection device according to one aspect of the present invention,
Based on the first odor data indicating the first odor and the second odor data obtained by measuring the first odor by the odor sensor, a coefficient is calculated, and a calculating unit,
An inspection unit for inspecting the odor sensor based on the coefficient,
It is characterized by having.

Further, in order to achieve the above object, the odor sensor inspection method according to one aspect of the present invention,
(A) calculating a coefficient based on first odor data indicating a first odor and second odor data obtained by measuring the first odor with an odor sensor;
(B) inspecting the odor sensor based on the coefficient;
It is characterized by having.

Further, in order to achieve the above object, a computer-readable recording medium recording the program according to one aspect of the present invention,
On the computer,
(A) calculating a coefficient based on first odor data indicating a first odor and second odor data obtained by measuring the first odor with an odor sensor;
(B) inspecting the odor sensor based on the coefficient;
It is characterized in that a program including an instruction to execute is recorded.

According to the present invention as described above, it is possible to accurately detect a defect in the odor sensor.

FIG. 1 is a diagram showing an example of an odor sensor inspection device. FIG. 2 is a diagram showing an example of a system having an odor sensor inspection device. FIG. 3 is a diagram illustrating an example of the data structure of the reference odor data and the reference target odor data. FIG. 4 is a diagram showing an example of a waveform of odor data. FIG. 5 is a diagram showing an example of the data structure of coefficient data. FIG. 6 is a diagram showing an example of the data structure of the defect degree data. FIG. 7: is a figure which shows an example of operation | movement of an odor sensor inspection apparatus. FIG. 8 is a diagram illustrating an example of a computer that realizes the odor sensor inspection device.

(Embodiment)
Hereinafter, embodiments of the present invention will be described with reference to FIGS. 1 to 8.

[Device configuration]
First, the configuration of the odor sensor inspection device 1 according to the present embodiment will be described with reference to FIG. FIG. 1 is a diagram showing an example of an odor sensor inspection device.

The odor sensor inspection device 1 shown in FIG. 1 is a device that accurately detects a defect in the odor sensor. Further, as shown in FIG. 1, the odor sensor inspection device 1 includes a calculation unit 2 and an inspection unit 3.

Of these, the calculation unit 2 uses the reference odor data (first odor data) indicating the reference odor (first odor) and the measured odor data obtained by the odor sensor measuring the reference odor (second odor data). The odor data) and the coefficient are calculated. The inspection unit 3 inspects the odor sensor based on the coefficient. The coefficient is, for example, a correction coefficient for correcting the measured odor data measured by the target odor sensor, or the reference odor data and the correlation between the measured odor data obtained by measuring the reference odor by the target odor sensor. For example, it is a correlation coefficient.

As described above, in the present embodiment, since the odor sensor is inspected using the calculated coefficient, it is possible to accurately detect the defect of the odor sensor.

[System configuration]
Subsequently, the inspection of the odor sensor according to the present embodiment will be described more specifically with reference to FIG. FIG. 2 is a diagram showing an example of a system having an odor sensor inspection device.

As shown in FIG. 2, the system according to the present embodiment includes a reference odor sensor 21a, an odor sensor 21b, an acquisition unit 23, and an output unit 24 in addition to the calculation unit 2 and the inspection unit 3. The reference odor sensor 21a has a sensitive film 22a and a sensitive film 22b. The odor sensor 21b has a sensitive film 22c and a sensitive film 22d. Further, the calculation unit 2 has a preprocessing unit 25 and a coefficient calculation unit 26.

Explain the calculation of the coefficient. When calculating the coefficient, the acquisition unit 23 first acquires the measurement result (reference odor data) obtained by the reference odor sensor 21a, which is the reference odor sensor, measuring the reference odor (reference gas).

Note that the reference odor data may be data obtained from one reference odor sensor, or a plurality of measurement results obtained by measuring the reference odors from the plurality of reference odor sensors 21a may be used. For example, the standard odor data may be obtained by using statistical processing such as average and median.

Subsequently, the calculation unit 2 acquires the measurement result (measured odor data) obtained by measuring the reference odor by the odor sensor 21b for which the correction coefficient is set. After that, the calculation unit 2 uses the reference odor data and the measured odor data obtained by measuring the reference odor, and at the time of measurement, calculates a correction coefficient for correcting the target odor of the measured odor data measured by the odor sensor 21b. calculate. Alternatively, the calculation unit 2 calculates the correlation coefficient using the reference odor data and the measured odor data obtained by measuring the reference odor. Alternatively, the calculation unit 2 calculates the correction coefficient and the correlation coefficient.

Explain the inspection. When performing the inspection, first, the inspection unit 3 acquires a coefficient (correction coefficient, correlation coefficient, or both), and based on the correction coefficient, the correlation coefficient, or both, the odor sensor 21b Determine the degree of failure. The inspector inspects the odor sensor 21b according to the degree of the defect.

The reference odor sensor 21a is a reference odor sensor used when calculating the coefficient, and has one or more sensitive films 22. In the example of FIG. 2, the reference odor sensor 21a has

sensitive films

22a and 22b. Specifically, the reference odor sensor 21a performs measurement using the reference odor, and outputs the reference odor data for each of the

sensitive films

22a and 22b to the acquisition unit 23.

The odor sensor 21b is an odor sensor for which a coefficient is set, and has one or more sensitive films 22. In the example of FIG. 2, the odor sensor 21b has

sensitive films

22c and 22d. Specifically, the odor sensor 21b measures the reference odor and outputs the measured odor data to the calculation unit 2 for each of the

sensitive films

22c and 22d.

Explain the odor sensor 21 (21a, 21b). The odor sensor is a sensor that detects a chemical substance in the air by using an element that detects a chemical substance. Specifically, an odor sensor is a cantilever if it utilizes changes in physical quantities related to device viscoelasticity and dynamic characteristics (mass, moment of inertia, etc.) due to adsorption and desorption of molecules on a sensitive film. Formula, film type, optical type, piezo, vibration response may be used. For example, as the odor sensor, a membrane surface stress sensor (MSS: Membrane-type Surface Stress Sensor) or the like can be considered.

MSS is an odor sensor that has multiple MSS elements. The MSS element has a sensitive film, a support member that supports the sensitive film, a wiring substrate that surrounds the support member, and a plurality of bridges that connect the support member and the wiring substrate. The bridge also has a piezoresistive element.

ㆍ In the sensitive film, when a substance is adsorbed, distortion occurs in the sensitive film. The support member is a member that supports the sensitive film, and has a mechanism of distorting according to the strain of the sensitive film. The bridge is connected to the support member, and when stress is applied to the bridge due to the strain described above, the electrical resistance of the piezoresistive element embedded in the bridge changes. That is, the MSS measures this electrical resistance and outputs odor data representing the measurement result via the wiring board.

Also, the material of the sensitive film of the MSS element differs for each MSS element, but the type of substance detected by the MSS element is not fixed to one. Therefore, the material of the sensitive film differs depending on the set of substances that make up the odor. The sensitive film is formed by depositing a chemical substance on the substrate for constructing the sensitive film. As the attachment method, for example, a method such as liquid application by an inkjet method, a dip method, or vapor deposition is used.

The acquisition unit 23 acquires reference odor data representing the measurement result for each sensitive film 22 from the reference odor sensor 21a. Specifically, the acquisition unit 23 stores the reference odor data acquired from the reference odor sensor 21a in a storage unit (not shown) in advance. For example, the acquisition unit 23 acquires reference odor data in which the sensitized film identification information for identifying each of the sensitized films 22 and the odor data output by each of the responsive films 22 are associated and stored in the storage unit.

The storage unit described above may be provided inside the odor sensor inspection device 1 or may be provided outside the odor sensor inspection device 1. When provided externally, the odor sensor inspection device 1 communicates with a storage unit provided externally to acquire reference odor data.

FIG. 3 is a diagram showing an example of the data structure of the standard odor data and the standard target odor data. The reference odor data 31 in FIG. 3 is “1” and “2” that represent the sensitive film identification information that identifies the

sensitive films

22a and 22b, and “data1” and “data2” that represent the odor data output by each of the

sensitive films

22a and 22b. And are associated and stored.

The calculation unit 2 calculates a coefficient (correction coefficient, correlation coefficient, or both) for each sensitive film 22. The calculation unit 2 also includes a preprocessing unit 25 and a coefficient calculation unit 26. However, the preprocessing unit 25 may not be provided in the calculation unit 2.

Specifically, the calculation unit 2 first acquires the measured odor data obtained by measuring the reference odor from the target odor sensor 21b. The measured odor data 32 in FIG. 3 represents "3" and "4" representing the sensitive film identification information for identifying the

sensitive films

22c and 22d, and the odor data obtained by measuring the reference odors output by the

sensitive films

22c and 22d. "data3" and "data4" are associated with each other.

Subsequently, the calculation unit 2 calculates a coefficient (correction coefficient, correlation coefficient, or both) for each sensitive film 22 using the reference odor data and the measured odor data obtained by measuring the reference odor. Specifically, when the sensitive film “1” (22a) and the sensitive film “3” (22c) are the sensitive films corresponding to each other, the calculation unit 2 responds to the odor data “data1” of the sensitive film “1” and the sensitive film. The coefficient is calculated using the odor data “data3” of the film “3”. Further, when the sensitive film “2” (22b) and the sensitive film “4” (22d) are corresponding sensitive films, the calculation unit 2 calculates the odor data “data2” and the sensitive film “4” of the sensitive film “2”. The coefficient is calculated by using the odor data “data4” of “.

The calculation unit will be described in detail.
The pre-processing unit 25 performs pre-processing on the standard odor data and the measured odor data obtained by measuring the standard odor. Specifically, as shown in Formula 1, the standard odor data and the measured odor data obtained by measuring the standard odor are preprocessed using a linear conversion matrix or the like. However, the linear conversion matrix does not necessarily have to be used for the preprocessing.

The pre-processing includes, for example, (A) acquisition of statistics, (B) downsampling, (C) smoothing, (D) offset removal, (E) feature extraction, and (F) weighting. is there.

Preprocessing will be explained using FIG. FIG. 4 is a diagram showing an example of a waveform of odor data. In FIG. 4, the vertical axis shows the level L of the odor data, and the horizontal axis shows the time t.

(A) Acquisition of statistics will be described.
The acquisition of the statistical amount is, for example, a process of calculating an amplitude and an average. In the process of calculating the amplitude, in the example of FIG. 4, the level L0 at the time t0 and the level Lm at the time tm are acquired, and the difference (Lm−L0) between the level Lm and the level L0 is set as the amplitude. The amplitude is related to the magnitude of the reaction of the sensitive membrane 22. In the example of FIG. 4, the process of calculating the amplitude from the level of the odor data can be expressed using, for example, a linear conversion matrix shown in Formula 2.

In the process of calculating the average, for example, the average of some or all of the levels of odor data is calculated. The process of calculating the average is such that, when calculating the average of the levels from time t0 to time tn in FIG. 4, when the number of odor data sampled during the time from time t0 to time tn is N, the odor data N Calculate the average of the individual levels. By calculating the average in this way, it is expected to reduce noise based on the odor data.

Note that the process of calculating the average from the level of the odor data can be expressed by using a linear conversion matrix as shown in Formula 3, for example.

(B) Down sampling will be described.
The downsampling is, for example, a process (thinning process) of acquiring odor data at predetermined intervals. By using downsampling, the amount of data can be suppressed, so that the processing speed can be improved.

The downsampling process can be expressed using, for example, a linear transformation matrix as shown in Equation 4. The linear transformation matrix of the example of Expression 4 thins out the odor data to 1/3.

(C) Smoothing will be described.
The smoothing is a process of applying a moving average filter, a Gaussian filter, a median filter, or the like to the odor data. By using the smoothing, it is possible to reduce the noise on the level of the odor data. In the case of linear conversion, the smoothing process uses, for example, a moving average filter or a Gaussian filter. For example, the kernels of the equations 5 and 6 are used, respectively.

Explain how to use a specific kernel by using an example of moving average filter. The kernel of the moving average filter is expressed by Equation 5. Here, if the size of the kernel is (1 × 3), the kernel is expressed as (1/3/1/3 1/3). The odor data can be smoothed by the moving average filter by multiplying the kernel and the odor data while shifting them and further adding them.

That is, the processing of the moving average filter can be expressed by using, for example, a linear conversion matrix shown in Expression 7. Also in the Gaussian filter, the odor data can be smoothed by Gaussian by performing the same calculation after determining the kernel by setting the size of the kernel to (1 × 3) and the magnitude of the variance to 1, for example.

In the case of non-linear conversion, for example, in the case of the median filter, if the kernel size is (1 × 3), the data is selected while shifting the odor data by three, and if the median of each is taken, it is smoothed by the median filter. Can be converted. Note that the moving average filter, Gaussian filter, and median filter are examples, and smoothing may be performed using another filter such as a Gabor filter. The presented numerical value is an example, and another numerical value may be used to determine the kernel.

(D) The removal of offset will be described.
The removal of the offset is, for example, a process of subtracting an average value of a part or all of the levels from each level of the odor data, or a process of subtracting the level of the odor data at a predetermined time from each level of the odor data. By removing such an offset, the bias can be removed. Note that the process of subtracting the average value of the levels from the level of the odor data is expressed using, for example, a linear conversion matrix shown in Expression 8.

Further, the process of subtracting the level L0 of the odor data at time t0 in FIG. 4 from each level L of the odor data is expressed by, for example, a linear conversion matrix shown in Expression 9.

(E) The extraction of the feature amount of the rate constant contribution will be described.
The odor data may be converted such that the magnitude of the contribution of the speed of adsorption / desorption of molecules to the sensitive film (rate constant) is used as the characteristic amount. For example, a linear transformation matrix as shown in Expression 10 may be used.

(F) Weighting will be described.
Weighting is a process of assigning weight by designating the odor data of a place to be emphasized. By weighting, any important part of the odor data can be emphasized. For weighting, a linear transformation matrix as shown in Expression 11 may be used. The linear conversion matrix shown in Expression 11 is a linear conversion matrix for making the levels L0 and Lm notice in the odor data acquired from the time immediately before the rising time t0 of the waveform in FIG. 4 to the time immediately after the falling time tm. Is. That is, the values 0.1, 0.3, and 0.1 of Expression 11 make the levels L1 and Lm and the level L in the vicinity thereof large.

Further, when performing regression analysis or discriminant analysis using odor data, for example, a linear model such as Expression 12 is used as it is or as a part of an equation. When this linear model is used, its output is determined based on the total value obtained by adding the weights w _i and x _i together. That is, it is considered that the larger the absolute values of w _i and x _i are, the more they contribute to the output. Therefore, by weighting the coefficient of the created linear model, the odor data of the part that contributes to the output of regression analysis or discriminant analysis can be emphasized. When the coefficient w _{i of the} linear model is used as the weight, for example, a linear conversion matrix shown in Expression 12 is used. At this time, the odor data may or may not be standardized.

Note that two or more of the above-mentioned pretreatments may be combined and used. For example, (C) smoothing may be performed after (D) offset is removed.

The calculation of the correction coefficient will be described.
The coefficient calculation unit 26 calculates the correction coefficient α by, for example, minimizing the equation shown in Expression 13. For the minimization, for example, the least squares method or the stochastic gradient descent method may be used.

Alternatively, the coefficient calculation unit 26 may calculate the correction coefficient α by applying the above-described linear conversion matrix (pre-processing) and then minimizing the coefficient, as shown in Expression 14, for example.

Furthermore, the coefficient calculation unit 26 may calculate the correction coefficient α _k by minimizing the condition for each condition, as shown in Expression 15. Here, the condition is a measurement condition in which the temperature, the humidity, the type of the standard odor, and the like are combined.

Alternatively, the coefficient calculation unit 26 may calculate the correction coefficient α _k by applying and minimizing the above-described linear conversion matrix (preprocessing) for each condition, as shown in Expression 16. ..

The calculation of the correlation coefficient will be described.
When calculating the correlation coefficient, the coefficient calculation unit 26 calculates the correlation coefficient r using the reference odor data and the measured odor data obtained by measuring the reference odor. Specifically, the coefficient calculation unit 26 divides the covariance of the reference odor data and the measured odor data obtained by measuring the reference odor by the respective standard deviations to obtain a correlation coefficient indicating the strength of the linear relationship. Calculate r.

Alternatively, when calculating the correlation coefficient for each of the above-described conditions, the coefficient calculating unit 26 uses the reference odor data and the measured odor data obtained by measuring the reference odor for each condition, and the correlation coefficient r for each condition. You may calculate _k .

The inspection unit 3 determines the degree of failure of the odor sensor 21b based on the coefficient. Specifically, the inspection unit 3 first acquires the coefficient (correction coefficient, correlation coefficient, or both) calculated by the coefficient calculation unit 26. Subsequently, the inspection unit 3 generates coefficient data and stores it in a storage unit (not shown). Inspecting unit 3 is, for example, the correction coefficient alpha, alpha _k, and generates coefficient data by using a correlation coefficient r, r _k, the storage unit provided within the odor sensor inspection apparatus 1 as described above, or , Is stored in an external storage unit.

Note that the calculation unit 2 may generate the coefficient data and store the coefficient data in the storage unit provided inside the odor sensor inspection device 1 or the storage unit provided outside as described above.

FIG. 5 is a diagram showing an example of the data structure of coefficient data. The coefficient data 51 shown in FIG. 5 includes "S1" that represents sensor identification information that identifies the odor sensor 21b, "3" and "4" that represents sensitive film identification information that identifies the

sensitive films

22c and 22d, and the sensitive film 22c. , 22d are stored in association with “cr3” and “cr4”, which represent the correction coefficient α, and “rr3” and “rr4”, which represent the correlation coefficient r.

The coefficient data 52 shown in FIG. 5 is “S1” indicating sensor identification information for identifying the odor sensor 21b, “con1” and “con2” indicating conditions, and sensitive film identification information for identifying the sensitive films 22c and 22d. “3” and “4”, “cr3_1”, “cr4_1”, “cr3_2”, and “cr4_2” that represent the correction coefficients α _k for the

sensitive films

22c and 22d for each condition, and “rr3_1” that represents the correlation coefficient r _k. "Rr4_1", "rr3_2", and "rr4_2" are stored in association with each other.

Subsequently, the inspection unit 3 selects any one of the correction coefficients α, α _k and the correlation coefficients r, r _k corresponding to each of the sensitive films 22 of the odor sensor 21b, or the correction coefficient α and the correlation coefficient r, or The inspection is performed using the correction coefficient α _k and the correlation coefficient r _k , the correction coefficient α and the correlation coefficient r _k , or the correction coefficient α _k and the correlation coefficient r. Specifically, the inspection unit 3 selects one of the correction coefficients α, α _k and the correlation coefficients r, r _k corresponding to the obtained

sensitive films

22c, 22d, or the correction coefficient α and the correlation coefficient r, Alternatively, using the correction coefficient α _k and the correlation coefficient r _k , or the correction coefficient α and the correlation coefficient r _k , or the correction coefficient α _k and the correlation coefficient r, the preset defect degree data described later is referred to. The degree of failure of the odor sensor 21b is determined.

For example, the test unit 3, when using the correction coefficient alpha or alpha _k, based on the magnitude of the correction coefficient alpha or alpha _k, determines defective degree. For example, when the correction coefficient α or α _k is less than or equal to the predetermined threshold value th1, the inspection unit 3 passes the odor sensor 21b. In addition, the inspection unit 3 rejects the odor sensor 21b when it is larger than the predetermined threshold value th1. The predetermined threshold th1 is set, for example, by experiment or simulation.

Here, since the correction coefficient α or α _k is a coefficient used for correcting the measured odor data to match the reference odor data, the measured odor data is multiplied. Therefore, when the correction coefficient α or α _k is large, the noise itself included in the measured odor data also becomes large, and the odor may not be accurately measured. Therefore, by rejecting a sensor having a large correction coefficient α or α _k, the odor sensor 21 that outputs a large amount of noise can be excluded.

Or, the inspection unit 3 determines failure of the original case, its size using the correlation coefficient r or r _k. Inspecting unit 3, if the correlation coefficient r or _{r k} is the predetermined threshold th2 or more, the odor sensor 21b as acceptable. In addition, the inspection unit 3 fails the odor sensor 21b when it is smaller than the predetermined threshold value th2. The predetermined threshold th2 is set by, for example, an experiment or a simulation.

Here, the correlation coefficient r or r _k are the indicator of the strength of the linear relationship of two data, if there is some problem, the reference odor data and measurement odor data measured with the reference odor The odor sensor 21b is excluded because it is considered to have a weak relationship with.

Furthermore, the inspection unit 3 may make the determination using only one of the above-described correction coefficient and correlation coefficient, or may use the correction coefficient α and the correlation coefficient r or the correction coefficient α _k and the correlation coefficient r _k. Alternatively, the inspection may be performed using the correction coefficient α and the correlation coefficient r _k , or the correction coefficient α _k and the correlation coefficient r.

Specifically, the inspection unit 3 may pass the odor sensor 21b when passing either one of the inspection using the correction coefficient and the inspection using the correlation coefficient. Alternatively, the inspection unit 3 may pass the odor sensor 21b when both the inspection using the correction coefficient and the inspection using the correlation coefficient have passed.

Further, in the case of failure, the inspection unit 3 sets a degree of failure such as “reinspection”, “recoating”, or “disposal” according to the values of the correction coefficient, the correlation coefficient, etc. The treatment after the inspection may be determined. FIG. 6 is a diagram showing an example of the data structure of the defect degree data. In the example of FIG. 6, the defect levels “level1”, “level2”, “level3”, and “level4” are associated with “pass”, “re-inspection”, “re-application”, and “disposal” representing the inspection result of the odor sensor 21 and stored. It is stored in the department. For example, the degree of failure is determined by experiments and simulations.

The output unit 24 acquires the output information indicating the inspection result (defective degree) converted into the format that can be output from the inspection unit 3, and outputs the image and the sound generated based on the output information. The output unit 24 is, for example, an image display device using a liquid crystal, an organic EL (Electro Luminescence), or a CRT (Cathode Ray Tube). Further, the image display device may have an audio output device such as a speaker. The output unit 24 may be a printing device such as a printer.

[Device operation]
Next, the operation of the odor sensor inspection device 1 according to the embodiment of the present invention will be described with reference to FIG. FIG. 7 is a flowchart showing an example of the operation of the odor sensor inspection device. In the following description, FIGS. 1 to 6 will be referred to as appropriate. Moreover, in this Embodiment, the odor sensor inspection method is implemented by operating the odor sensor inspection apparatus 1. Therefore, the description of the odor sensor inspection method according to the present embodiment will replace the operation description of the odor sensor inspection device 1 below.

As shown in FIG. 7, first, the acquisition unit 23 acquires reference odor data for each sensitive film 22 from the reference odor sensor 21a (step A1). Specifically, in step A1, the acquisition unit 23 acquires reference odor data in which the sensitive film identification information for identifying each of the sensitive films 22 and the odor data output by each of the sensitive films 22 are associated with each other, and is stored in the storage unit. Remember. See the reference odor data 31 in FIG. If the standard odor data is already stored in the storage unit, the process of step A1 is unnecessary.

Subsequently, the calculation unit 2 acquires the measured odor data measured using the reference odor for each of the sensitive films 22 from the odor sensor 21b for which the correction coefficient is set (step A2). Specifically, in step A2, the calculation unit 2 measures the odor measured using the reference odor in which the sensitive film identification information for identifying each sensitive film 22 and the odor data output by each sensitive film 22 are associated with each other. Get the data.

Subsequently, the preprocessing unit 25 of the calculation unit 2 performs preprocessing on the standard odor data and the measured odor data (step A3). Specifically, in step A3, the preprocessing unit 25 preprocesses the reference odor data and the measured odor data by using a linear conversion matrix or the like, as shown in Formula 1. However, the pre-processing of step A3 may be omitted.

The preprocessing includes, for example, (A) acquisition of statistics, (B) downsampling, (C) smoothing, (D) offset removal, (E) feature extraction, and (F) weighting. Processing.

Subsequently, the coefficient calculation unit 26 of the calculation unit 2 calculates a coefficient (correction coefficient, correlation coefficient, or both) for each sensitive film 22 by using the reference odor data and the measured odor data (step). A4).

Specifically, in step A4, when the sensitive film “1” (22a) and the sensitive film “3” (22c) are corresponding sensitive films, the coefficient calculation unit 26 determines that the odor data of the sensitive film “1”. The coefficient is calculated using "data1" and the odor data "data3" of the sensitive film "3". Further, when the sensitive film “2” (22b) and the sensitive film “4” (22d) are corresponding sensitive films, the calculation unit 2 calculates the odor data “data2” and the sensitive film “4” of the sensitive film “2”. The coefficient is calculated by using the odor data “data4” of “.

The coefficient calculating unit 26 calculates the correction coefficient α by minimizing it as shown in Expression 13, for example. Alternatively, the coefficient calculation unit 26 may calculate the correction coefficient α by applying the above-described linear conversion matrix (pre-processing) and then minimizing it, as shown in Expression 14, for example.

Alternatively, the coefficient calculation unit 26 may calculate the correction coefficient α _k by minimizing each condition, as shown in Expression 15. Here, the condition is a measurement condition in which the temperature, the humidity, the type of the standard odor, and the like are combined. Alternatively, the coefficient calculation unit 26 may calculate the correction coefficient α _k by applying the above-described linear conversion matrix (preprocessing) for each condition and then minimizing the coefficient, as shown in Expression 16, for example.

Alternatively, when calculating the correlation coefficient, the coefficient calculation unit 26 may calculate the correlation coefficient r using the reference odor data and the measured odor data obtained by measuring the reference odor. Specifically, the coefficient calculation unit 26 divides the covariance of the reference odor data and the measured odor data obtained by measuring the reference odor by the respective standard deviations to obtain a correlation coefficient indicating the strength of the linear relationship. Calculate r.

Or, the coefficient calculating unit 26 calculates when calculating a correlation coefficient for each condition mentioned above, the reference odor data, by using the measurement odor data reference odor was measured, the correlation coefficient r _k for each condition You may.

Subsequently, the inspection unit 3 inspects the odor sensor 21b using the coefficient (step A5). Specifically, in step A5, the inspection unit 3 first acquires the coefficient calculated by the coefficient calculation unit 26. Subsequently, the inspection unit 3 generates coefficient data and stores it in a storage unit (not shown).

Inspecting unit 3 is, for example, the correction coefficient alpha, alpha _k, and generates coefficient data by using a correlation coefficient r, r _k, the storage unit provided within the odor sensor inspection apparatus 1 as described above, or , Is stored in an external storage unit. See

coefficient data

51 and 52 in FIG.

Subsequently, the inspection unit 3 selects any one of the correction coefficients α, α _k and the correlation coefficients r, r _k corresponding to each of the sensitive films 22 of the odor sensor 21b, or the correction coefficient α and the correlation coefficient r, or The inspection is performed using the correction coefficient α _k and the correlation coefficient r _k , the correction coefficient α and the correlation coefficient r _k , or the correction coefficient α _k and the correlation coefficient r.

Specifically, the inspection unit 3 selects one of the correction coefficients α, α _k and the correlation coefficients r, r _k corresponding to the obtained

sensitive films

22c, 22d, or the correction coefficient α and the correlation coefficient r, Alternatively, the correction coefficient α _k and the correlation coefficient r _k , the correction coefficient α and the correlation coefficient r _k , or the correction coefficient α _k and the correlation coefficient r are used to refer to pre-stored defect degree data described later. The degree of failure of the odor sensor 21b is determined. See the defect degree data 61 in FIG.

For example, the test unit 3, when using the correction coefficient alpha or alpha _k, based on the magnitude of the correction coefficient alpha or alpha _k, determines defective degree. For example, when the correction coefficient α or α _k is less than or equal to the predetermined threshold value th1, the inspection unit 3 passes the odor sensor 21b. In addition, the inspection unit 3 rejects the odor sensor 21b when it is larger than the predetermined threshold value th1.

As described above, when the correction coefficient α or α _k is large, the noise itself included in the measured odor data also becomes large, and the odor may not be accurately measured. Therefore, by rejecting the odor sensor 21b having a large correction coefficient α or α _k , the odor sensor 21b having a large noise can be excluded.

Thus, the correlation coefficient r or r _k are the indicator of the strength of the linear relationship of two data, if there is some problem, and the reference odor data, measured odor data reference odor was measured The odor sensor 21b is excluded because it is considered to have a weak relationship with.

Specifically, the inspection unit 3 may pass the odor sensor 21b when either the inspection using the correction coefficient or the inspection using the correlation coefficient has passed. Alternatively, the inspection unit 3 may pass the odor sensor 21b when both the inspection using the correction coefficient and the inspection using the correlation coefficient have passed.

Subsequently, the output unit 62 acquires the output information indicating the determination result (defective degree) converted into the format that can be output from the inspection unit 3, and outputs the image and the sound generated based on the output information ( Step A6).

[Effects of this Embodiment]
As described above, according to the present embodiment, since the odor sensor is inspected using the calculated coefficient (correction coefficient, correlation coefficient, or both), it is possible to accurately detect a defect in the odor sensor.

Also, if there are manufacturing variations due to the shape, size, thickness, etc. of each sensitive film, it is not possible to accurately inspect the odor sensor for defects. However, according to the present embodiment, since the coefficient calculated for each sensitive film is used to perform the inspection for each sensitive film, it is possible to accurately detect the defect of the odor sensor.

[program]
The program according to the embodiment of the present invention may be a program that causes a computer to execute steps A1 to A6 shown in FIG. The odor sensor inspection device and the odor sensor inspection method according to the present embodiment can be realized by installing and executing this program on a computer. In this case, the processor of the computer functions as the acquisition unit 23, the calculation unit 2 (the preprocessing unit 25, the coefficient calculation unit 26), and the inspection unit 3, and performs the processing.

Also, the program in the present embodiment may be executed by a computer system constructed by a plurality of computers. In this case, for example, each computer may function as either the calculation unit 2 (preprocessing unit 25, coefficient calculation unit 26) or the inspection unit 3.

[Physical configuration]
Here, a computer that realizes the odor sensor inspection device by executing the program according to the embodiment will be described with reference to FIG. 8. FIG. 8 is a block diagram showing an example of a computer that realizes the odor sensor inspection device.

As shown in FIG. 8, the computer 110 includes a CPU 111, a main memory 112, a storage device 113, an input interface 114, a display controller 115, a data reader / writer 116, and a communication interface 117. These units are connected to each other via a bus 121 so as to be able to perform data communication with each other. The computer 110 may include a GPU (Graphics Processing Unit) or an FPGA (Field-Programmable Gate Array) in addition to or instead of the CPU 111.

The CPU 111 expands the program (code) according to the present embodiment stored in the storage device 113 into the main memory 112 and executes these in a predetermined order to perform various calculations. The main memory 112 is typically a volatile storage device such as a DRAM (Dynamic Random Access Memory). Further, the program in the present embodiment is provided in a state of being stored in computer-readable recording medium 120. The program in the present embodiment may be distributed on the Internet connected via communication interface 117.

Further, as a specific example of the storage device 113, a semiconductor storage device such as a flash memory can be cited in addition to a hard disk drive. The input interface 114 mediates data transmission between the CPU 111 and an input device 118 such as a keyboard and a mouse. The display controller 115 is connected to the display device 119 and controls the display on the display device 119.

The data reader / writer 116 mediates data transmission between the CPU 111 and the recording medium 120, reads a program from the recording medium 120, and writes the processing result in the computer 110 to the recording medium 120. The communication interface 117 mediates data transmission between the CPU 111 and another computer.

Specific examples of the recording medium 120 include general-purpose semiconductor storage devices such as CF (Compact Flash (registered trademark)) and SD (Secure Digital), magnetic recording media such as a flexible disk, or CD- An optical recording medium such as a ROM (Compact Disk Read Only Memory) can be given.

It should be noted that the odor sensor inspection device 1 according to the present embodiment can be realized by using hardware corresponding to each unit instead of using a computer in which a program is installed. Further, the odor sensor inspection device 1 may be partially implemented by a program and the rest may be implemented by hardware.

[Appendix]
Regarding the above-described embodiment, the following supplementary notes will be disclosed. The whole or part of the exemplary embodiments described above can be represented by (Supplementary Note 1) to (Supplementary Note 12) described below, but the present invention is not limited to the following description.

(Appendix 1)
Based on the first odor data indicating the first odor and the second odor data obtained by measuring the first odor by the odor sensor, a coefficient is calculated, and a calculating unit,
An inspection unit for inspecting the odor sensor based on the coefficient,
An odor sensor inspection device comprising:

(Appendix 2)
The odor sensor inspection device according to attachment 1,
The odor sensor has a plurality of sensitive films,
The said calculation part calculates the said coefficient for every said sensitive film. The odor sensor inspection apparatus characterized by the above-mentioned.

(Appendix 3)
The odor sensor inspection device according to attachment 2,
The odor sensor inspection device, wherein the first odor data is generated for each of the sensitive films based on the reference first odor.

(Appendix 4)
The odor sensor inspection device according to

attachment

2 or 3,
The odor sensor inspection device, wherein the inspection unit inspects each of the sensitive films based on the coefficient calculated for each of the sensitive films.

(Appendix 5)
(A) calculating a coefficient based on first odor data indicating a first odor and second odor data obtained by measuring the first odor with an odor sensor;
(B) inspecting the odor sensor based on the coefficient;
An odor sensor inspection method comprising:

(Appendix 6)
The odor sensor inspection method according to attachment 5,
The odor sensor has a plurality of sensitive films,
In the step (a), the odor sensor inspection method is characterized in that the coefficient is calculated for each of the sensitive films.

(Appendix 7)
The odor sensor inspection method according to attachment 6,
The odor sensor inspection method, wherein the first odor data is generated for each of the sensitive films based on the reference first odor.

(Appendix 8)
The odor sensor inspection method according to attachment 6 or 7,
In the step (b), the odor sensor inspection method is characterized by inspecting each of the sensitive films based on the coefficient calculated for each of the sensitive films.

(Appendix 9)
On the computer,
(A) calculating a coefficient based on first odor data indicating a first odor and second odor data obtained by measuring the first odor with an odor sensor;
(B) inspecting the odor sensor based on the coefficient;
A computer-readable recording medium recording a program including an instruction to execute.

(Appendix 10)
The computer-readable recording medium according to attachment 9,
The odor sensor has a plurality of sensitive films,
In the step (a), the computer-readable recording medium, wherein the coefficient is calculated for each of the sensitive films.

(Appendix 11)
The computer-readable recording medium according to attachment 10,
The computer-readable recording medium, wherein the first odor data is generated for each of the sensitive films based on the reference first odor.

(Appendix 12)
The computer-readable recording medium according to supplementary note 10 or 11,
A computer-readable recording medium, characterized in that, in the step (b), an inspection is performed for each of the sensitive films based on the coefficient calculated for each of the sensitive films.

Although the present invention has been described with reference to the exemplary embodiments, the present invention is not limited to the above exemplary embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

According to the present invention as described above, it is possible to accurately detect a defect in the odor sensor. INDUSTRIAL APPLICABILITY The present invention is useful in the field where it is necessary to inspect an odor sensor for defects.

1 Odor sensor inspection device 2 Calculation unit 3

Inspection unit

21, 21a,

21b Odor sensor

22, 22a, 22b, 22c, 22d Sensitive film 23 Acquisition unit 24 Output unit 25 Preprocessing unit 26 Coefficient calculation unit 31 Standard odor data 32 Measured odor Data 51 Coefficient data 61 Defect degree data 110 Computer 111 CPU
112 Main Memory 113 Storage Device 114 Input Interface 115 Display Controller 116 Data Reader / Writer 117 Communication Interface 118 Input Equipment 119 Display Device 120 Recording Medium 121 Bus

Claims

Based on the first odor data indicating the first odor, and the second odor data obtained by measuring the first odor by the odor sensor, a coefficient calculating unit,
Inspecting the odor sensor based on the coefficient, an inspection unit,
An odor sensor inspection device comprising:
The odor sensor inspection device according to claim 1,
The odor sensor has a plurality of sensitive films,
The odor sensor inspection device, wherein the calculating means calculates the coefficient for each of the sensitive films.
The odor sensor inspection device according to claim 2, wherein
The odor sensor inspection device, wherein the first odor data is generated for each of the sensitive films based on the reference first odor.
The odor sensor inspection device according to claim 2 or 3, wherein
The odor sensor inspection device, wherein the inspection means inspects each of the sensitive films based on the coefficient calculated for each of the sensitive films.
(A) calculating a coefficient based on first odor data indicating a first odor and second odor data obtained by measuring the first odor with an odor sensor;
(B) inspecting the odor sensor based on the coefficient;
An odor sensor inspection method comprising:
The odor sensor inspection method according to claim 5,
The odor sensor has a plurality of sensitive films,
In the step (a), the odor sensor inspection method is characterized in that the coefficient is calculated for each of the sensitive films.
The odor sensor inspection method according to claim 6,
The odor sensor inspection method, wherein the first odor data is generated for each of the sensitive films based on the reference first odor.
The odor sensor inspection method according to claim 6 or 7, wherein
In the step (b), the odor sensor inspection method is characterized by inspecting each of the sensitive films based on the coefficient calculated for each of the sensitive films.
On the computer,
(A) calculating a coefficient based on first odor data indicating a first odor and second odor data obtained by measuring the first odor with an odor sensor;
(B) inspecting the odor sensor based on the coefficient;
A computer-readable recording medium recording a program including an instruction to execute.
The computer-readable recording medium according to claim 9,
The odor sensor has a plurality of sensitive films,
In the step (a), the computer-readable recording medium, wherein the coefficient is calculated for each of the sensitive films.
The computer-readable recording medium according to claim 10,
The computer-readable recording medium, wherein the first odor data is generated for each of the sensitive films based on the reference first odor.
The computer-readable recording medium according to claim 10 or 11, wherein
A computer-readable recording medium, characterized in that, in the step (b), an inspection is performed for each of the sensitive films based on the coefficient calculated for each of the sensitive films.