JP4775088B2 - Odor identification device - Google Patents

Description

  The present invention relates to an odor discriminating apparatus for identifying similarities and differences between a plurality of odors (including all fragrances, odors, etc.). The present invention relates to an odor discriminating apparatus suitable for performing the above.

  In recent years, an odor measurement apparatus using an odor sensor that responds to an odor substance has been developed (see, for example, Patent Document 1). In such an odor measurement apparatus, the odor quality and the odor intensity can be expressed numerically by executing arithmetic processing using a predetermined algorithm based on detection signals acquired by a plurality of odor sensors. It has become. That is, in the odor measuring apparatus, based on the result of measuring a reference odor gas having a known odor, a reference axis is positioned in an m-dimensional odor space formed by detection outputs of m odor sensors, and the reference axis The odor quality is expressed by the positional relationship between the measurement point and the measurement point that is the measurement result of the unknown odor (the odor to be evaluated), specifically the similarity obtained by the degree of proximity / distance between the measurement point and the reference axis. ing. In addition, the odor intensity corresponding to a plurality of standard odor gases is calculated, and the odor index equivalent value is obtained by summing up the contribution of each odor, thereby expressing the odor intensity.

  One promising application of the odor measuring apparatus as described above is to identify whether a product such as a food or drink is a normal product or a defective product (abnormal product) from the viewpoint of odor. For example, in the odor discriminating apparatus described in Patent Document 2, the pass range of the evaluation target product is determined based on the difference in odor quality and the difference in odor intensity with respect to the reference odor, and the measurement result of the odor of the evaluation target product is the pass When it exists in the range, it is judged to be acceptable and when it is outside the range, it is judged to be unacceptable.

  However, in the conventional odor discriminating apparatus, for example, only pass / fail judgment based on the difference in odor quality and the difference in odor intensity with respect to the reference product can be performed, so that practically usable applications are considerably limited. For example, when there is a possibility that a plurality of odors different from the original odor of the product may be mixed during the manufacturing process, depending on how the plurality of odors are mixed, in some cases it can be determined to be acceptable and in another mixing method In the case where it is determined to be unacceptable, it is difficult for the odor discriminating apparatus to accurately determine pass / fail. In addition, since the contents of the difference in odor quality, that is, how the odor quality is different, are not considered in the pass / fail judgment, if the range of acceptable products is narrowed, it is originally judged as a acceptable product. There is a problem in that products that may be acceptable must be judged as rejected products, and conversely, if the range of acceptable products is expanded, products that should be judged as rejected products are judged as acceptable products.

JP 2003-315298 A JP-A-2005-77099

  For this reason, there is a strong demand for a device that can perform pass / fail evaluation and the like based on a more flexible and more precise determination standard than in the past when determining pass / fail of a product from the viewpoint of odor. . The present invention has been made in view of these points, and the main purpose of the present invention is to determine whether a product to be evaluated is acceptable or not from the point of view of odors. An object of the present invention is to provide an odor discriminating apparatus capable of performing evaluation.

An odor identification device according to the present invention , which has been made to solve the above problems,
a) odor measuring means including m (m is an integer of 2 or more) odor sensors having different response characteristics;
b) reference axis acquisition means for positioning a reference axis corresponding to the reference odor in the m-dimensional space formed by the detection output from the odor sensor by measuring the reference odor by the odor measurement means;
c) Measuring the odor of the evaluation object by the odor measuring means to position the measurement point in the m-dimensional space, and based on the positional relationship between the measurement point and the reference axis, the odor intensity of the evaluation object and the reference odor An evaluation processing means for obtaining the difference in odor quality, the difference in odor quality, and the direction of the odor quality difference in the evaluation object with respect to the reference odor, and thereby evaluating the evaluation object odor,
Equipped with a,
The evaluation processing means determines whether or not the evaluation target odor is a pass product based on a determination threshold determined for the difference in odor intensity, the difference in odor quality, and the directionality of the difference in odor quality. The pass / fail judgment means is characterized in that a pass level in the pass / fail judgment means can be set in a plurality of stages by a combination of a plurality of judgment thresholds .

In the odor discriminating apparatus according to the present invention, not only the magnitude of the odor quality difference between the evaluation object and the reference odor, but also the directionality of the odor quality difference is considered. In the m-dimensional space, the fact that the difference in odor quality with respect to a certain reference odor is directed in a certain direction can be considered as approaching some other odor. Therefore, there may be cases where the other odor is preferable or unfavorable. Therefore, the evaluation processing means judges such a situation according to the direction of the difference in odor quality, and evaluates it. The smell can be evaluated. Thereby, the evaluation of the evaluation target can be performed more flexibly and in detail.

Specifically, in one embodiment of the odor discriminating apparatus according to the present invention, in the m-dimensional space, a projection point of a measurement point into a plane orthogonal to the reference axis is considered, and the projection point and the reference axis penetrate. The directivity of the difference in odor quality can be expressed by the directivity of a straight line connecting the reference point.

  In this configuration, the directionality of the straight line in the plane can be expressed by an angle when a certain direction is used as a reference, and whether or not the evaluation target product is within the acceptable range is determined based on the angle. be able to. Thereby, the directionality of the difference in odor quality can be quantified and a pass / fail judgment can be made appropriately.

[First embodiment]
First, an odor identification apparatus according to an embodiment of the first invention (hereinafter referred to as a first embodiment) will be described with reference to the drawings. FIG. 1 is a schematic block diagram of the odor discriminating apparatus of the first embodiment.

  This odor discriminating apparatus has a response characteristic for measuring a suction port 1 for sucking a sample gas, a pretreatment unit 2 for performing various pretreatments on the sucked sample gas, and a sample gas containing various odor components. A sensor cell 3 having a plurality of different odor sensors 4 incorporated therein, a pump 5 for drawing sample gas into the sensor cell 3, an A / D converter 6 for converting a detection signal from the odor sensor 4 into a digital signal, and digitized detection data A data processing unit 7 that performs analysis processing, a display unit 10 that displays the results of the analysis processing on the screen, a control unit 9 that controls the operation of the entire apparatus, and a pass / fail judgment condition that is connected to the control unit 9 and that is described later by the operator The input unit 11 is used for input.

  In the pre-processing unit 2, removal of moisture contained in the sample gas, concentration / dilution of the target component in the sample gas, removal of interfering components, and the like are performed. The odor sensor 4 is generally a metal oxide semiconductor sensor whose resistance value changes according to the odor component, but in addition to this, a gas adsorption film is formed on the surface of a conductive polymer sensor, a crystal resonator, or a SAW device. A sensor based on another detection mechanism such as a sensor formed with the above-described sensor may be used. The number of odor sensors 4 is not particularly limited, but here, ten (that is, m = 10 in the present invention) is used as an example. Therefore, detection signals from the ten odor sensors 4 are input in parallel to the A / D conversion unit 6, and the A / D conversion unit 6 converts each detection signal into digital data and converts it into parallel or serial. To the data processing unit 7. The data processing unit 7 and the control unit 9 are configured with a personal computer as the center, and each function can be achieved by executing a predetermined processing / control program installed in the computer.

  In the odor discriminating apparatus of the present embodiment, the index value reflects the similarity / difference of the odor quality between the evaluation target and the reference based on the reference axis obtained by measuring the known odor in advance. The similarity and the odor index equivalent value that is an index value of the overall odor intensity are obtained. First, how to obtain the similarity and the odor index equivalent value will be described.

  As described above, since the odor sensors 4 have different response characteristics, a 10-dimensional space formed by using the detection outputs respectively obtained from the ten odor sensors 4 as axes in different directions (directions orthogonal to each other). Considering the above, it is possible to position one measurement point in the 10-dimensional space based on the detection signals obtained from the ten odor sensors 4 as described above.

  The measurement results of a plurality of gases having the same odor quality but different concentrations are positioned as different measurement points in the 10-dimensional space, and are positioned so as to move away from the origin of the 10-dimensional space as the concentration increases. Therefore, it is possible to draw a single line connecting each measurement point with the increase in the concentration of the odor component starting from the origin. Since it is difficult and easy to understand the 10-dimensional space, considering the 2-dimensional space formed by the detection outputs (CH1, CH2) of the two odor sensors, the same odor component as shown in FIG. The measurement points which are the results of measuring a plurality of gases having different concentrations are positioned as U1, U2, U3, for example. Therefore, it is possible to draw a line Ta connecting the origin and these measurement points U1, U2, U3. Further, if the same measurement is performed for other gases having different odor qualities, the measurement points are positioned as W1, W2, and W3, for example, and a line Tb that does not coincide with the line Ta can be drawn. The same applies to a 10-dimensional odor space.

  Therefore, by measuring each reference gas having n kinds of predetermined reference odors while changing the concentration, n lines are drawn in a 10-dimensional space to define each line as a reference axis. The standard odor type may be determined as appropriate depending on the odor type to be evaluated. In FIG. 2, the lines Ta and Tb are shown as straight lines. However, when the output of the odor sensor 4 has non-linearity, it may not necessarily be a straight line but may be a curved line. But that curve can be used as the reference axis.

  FIG. 3 is a conceptual diagram showing a state in which the reference axes Ta, Tb, Tc, and Td, which are the results of measuring the reference smells A, B, C, and D, are positioned in the 10-dimensional space. A measurement point P, which is the result of measuring an odor for a certain purpose, is similarly positioned in the 10-dimensional space. If the object odor is exactly the same as the reference odor A, the measurement point P should be on the reference axis Ta, but if the object odor is qualitatively different from the reference, it is shown in FIG. As shown, the measurement point P exists at a position away from the reference axis Ta. At this time, the closer the odor quality is, the closer the measurement point P is to a position closer to the reference axis Ta. Therefore, as shown in FIG. 3, a straight line S connecting the origin O and the measurement point P is drawn. An angle θa formed by S and the reference axis Ta is obtained. Based on this angle θa, an index value indicating the degree of similarity of the target odor to the reference odor A is calculated.

  Specifically, when the angle θa is 0, the similarity is set to 100%, and when the angle θa is equal to or greater than a predetermined angle, the similarity is set to 0%, and the similarity is set to 0 according to the angle θa therebetween. It shall be specified in a range of ˜100%. However, since there is a difference between the response sensitivity of the odor sensor 4 and the sensitivity of the human nose (olfaction), it is preferable to perform a process of correcting the difference when determining the similarity from the angle θa. As a correction method, even if the degree of similarity by the above calculation is the same, the similarity should be relatively higher as the threshold sensitivity of the person's nose with respect to each reference is lower. Similarities can be similarly obtained for the other reference axes Tb, Tc, and Td.

  On the other hand, the odor index equivalent value is obtained as follows. Since the position of each point on the reference axes Ta, Tb, Tc, Td should correspond to the concentration of the reference odor component, as shown in FIG. A perpendicular line is drawn from P, and a density corresponding to the reference is obtained based on the positions of the intersections Qa to Qd of the perpendicular lines on the reference axes Ta to Td. Then, the odor concentration is calculated by dividing the substance concentration by the olfactory threshold of the odor for each reference odor (the limit concentration that can be felt by the nose), and logarithmically converted to obtain an odor index equivalent value. In this way, the odor index equivalent value is obtained for each odor, and the odor index equivalent value is added according to the degree of contribution to calculate a total odor index equivalent value.

  As described above, the similarity as the odor quality index value and the odor index equivalent value as the odor intensity index value can be obtained. The odor quality index value and the odor intensity index value may be obtained by a method other than the above. Such a method can use various methods already proposed in Patent Documents 1 and 2, etc., and in any case, an index value that reflects similarity / difference between an evaluation target and a standard, The index value reflecting the odor intensity of the odor to be evaluated can be determined clearly.

  Next, a characteristic odor evaluation processing operation performed under the control of the control unit 9 in the odor identification apparatus of the present embodiment will be described.

  Here, it is assumed that there is a possibility that the above-mentioned standards A, B, C, and D may be mixed as impurities in the evaluation object, and the case where the pass / fail of the evaluation object product is determined according to the degree of the odor due to the impurity is considered. Therefore, it is ideal that the odor to be evaluated does not include any of the standards A, B, C, and D. This ideal state may be a state with a certain smell or a completely odorless state. The reference axes Ta, Tb, Tc, and Td as described above are created in advance by measuring these reference smells A, B, C, and D, and information representing these reference axes is stored in the reference axis storage unit 8. Let me.

  The person in charge sets pass condition items from the input unit 11 prior to measurement of the evaluation target product. FIG. 4 is an example of this pass condition item. That is, as the condition items, there are similarities (in the above% display) αa, αb, αc, αd and overall odor intensity with respect to odors A, B, C, and D, and pass conditions in which respective threshold values are arbitrarily set. Can be set at multiple levels. In this example, three levels of pass conditions are set. That is, the acceptance condition 1 is that the degree of similarity αa, αb, αc, αd is not limited as long as the condition that the odor intensity is 5 or less is satisfied. In addition, the pass condition 2 is that if the odor intensity is 10 or less and the similarity αc to the odor quality C is 20% or less, the other similarities αa, αb, and αd do not matter. Furthermore, the pass condition 3 is that the odor intensity is 20 or less, the similarity αa is 50% or less, the similarity αb is 10% or less, the similarity αc is 15% or less, and the similarity αd is 5% or less. In addition, all the conditions that the similarity αa is equal to or lower than the similarity αb are satisfied. Fails if none of these three levels of pass conditions are met.

  The passing condition input from the input unit 11 as described above is set in the data processing unit 7. When the odor to be evaluated is set in the suction port 1 and the measurement by the odor sensor 4 is executed as described above, the data processing unit 7 uses the reference axis based on the information stored in the reference axis storage unit 8. The above arithmetic processing with Ta, Tb, Tc, and Td set is executed, thereby calculating the similarities αa, αb, αc, and αd for the odors A, B, C, and D, respectively. An odor index equivalent value that is an odor intensity is also calculated. After that, it is checked one by one whether each numerical value meets the set pass condition.

  Now, for example, when the product No. 1 shown in FIG. 5 is given as an odor, since the odor strength is 10, it is determined that the pass condition 1 is unacceptable, and then the odor strength is 10 under the pass condition 2. Since the degree of similarity αc: 17 is satisfied, it is determined to be acceptable. Therefore, the pass / fail judgment result is passed under pass condition 2. When a plurality of products to be evaluated are measured in order, by determining the measurement result, a pass or fail result is obtained as shown in FIG. In this way, the products to be evaluated are judged at different levels from the viewpoint of the similarity of odor quality to multiple odors and from the viewpoint of odor intensity, and pass / fail judgments are made. Can be classified.

[Second Embodiment]
Next, an odor discriminating apparatus according to an embodiment of the second invention (hereinafter referred to as the second embodiment) will be described with reference to the drawings. The basic configuration of the odor discriminating apparatus of the second embodiment is the same as that of the first embodiment, but the arithmetic processing in the data processing unit 7 is different. Therefore, this point will be described.

  The point of calculating the odor quality similarity to the reference odor of the evaluation target odor and the odor index equivalent value is the same as in the second embodiment. Furthermore, in this second embodiment, the direction of the difference in odor quality with respect to the reference odor is also added as an index for evaluation.

  Consider a situation where one reference axis Ta is positioned by measurement of a certain reference odor A in a 10-dimensional odor space as shown in FIG. As described above, when the odor quality of the evaluation target odor is qualitatively different from the reference odor A, the measurement point P due to the evaluation target odor exists at a position away from the reference axis Ta. As described above, the similarity can be calculated based on the angle θ formed by the line S connecting the measurement point P and the origin O and the reference axis Ta. The degree of similarity based on the angle θ represents how much the odor quality is different, but no consideration is given to how it differs. That is, as shown in FIG. 6, the angle θ is the same regardless of the position of the measurement point P on the concentric circle G centered on the reference axis Ta, and there is no difference in the similarity.

  Therefore, as shown in FIG. 7, a plane H orthogonal to the reference axis Ta in the 10-dimensional space is considered. Here, the plane H includes the measurement point P, but the measurement point P is not necessarily included. When the plane H that does not include the measurement point P is selected, the projection point of the measurement point P is drawn on the plane H. Then, a line Sp connecting the point Ta ′ through which the reference axis Ta passes on the plane H and the measurement point P (or its projection point) is drawn. The direction of this line around the point Ta 'is defined as the direction of the difference in odor quality. In order to quantify this directionality, the line S0 connecting the point Ta ′ and the point Tb ′ through which another reference axis (here, Tb) passes is used as a reference, and the angle of the line Sp with reference to this S0. Obtain θ1. In the example of FIG. 7, the directionality of the difference in odor quality at the measurement point P is θ1, but the directionality of the difference in odor quality at other measurement points P ′ on the concentric circle G showing the same similarity is θ2. Become. By using the index value based on the angle expression defined in this way in addition to the similarity and the odor index equivalent value, it becomes possible to evaluate the difference in odor quality in more detail.

  For example, in the past, when the pass / fail of the evaluation target product has been determined only by the similarity of the odor quality and the odor intensity, the pass / fail is the same whether it is located at the measurement point P or at the measurement point P ′. . On the other hand, if the reference odor B that is the basis of the reference axis Tb is a scent that is comfortable for humans, as shown in FIG. 8, a fan-shaped range extending at the same angle on both sides of the line S0 is defined as an acceptable range, It can be distinguished that the measurement point P is acceptable because it falls within this range, while the measurement point P ′ is outside this range and fails.

  Thus, by introducing a new indicator of the direction of the difference in odor quality, it becomes possible to perform pass / fail judgment and evaluation more flexibly than in the past.

  Note that the above embodiment is merely an example of the present invention, and it is obvious that modifications, corrections, additions, and the like can be appropriately made within the scope of the present invention.

The schematic block block diagram of the smell identification apparatus of one Example (1st Example) of 1st invention. Explanatory drawing of the evaluation process operation | movement by the smell identification device of 1st Example. Explanatory drawing of the evaluation process operation | movement of the smell identification apparatus of 1st Example. The figure which shows an example of the pass condition item input-set by the smell identification apparatus of 1st Example. The figure which shows an example of the pass / fail determination result by the smell identification apparatus of 1st Example. Explanatory drawing of evaluation process operation | movement by the smell identification apparatus of one Example (2nd Example) of 2nd invention. Explanatory drawing of the evaluation process operation | movement by the smell identification device of 2nd Example. Explanatory drawing of the evaluation process operation | movement by the smell identification device of 2nd Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Suction port 2 ... Pre-processing part 3 ... Sensor cell 4 ... Odor sensor 5 ... Pump 6 ... A / D conversion part 7 ... Data processing part 8 ... Reference axis memory | storage part 9 ... Control part 10 ... Display part 11 ... Input part

Claims (2)

a) odor measuring means including m (m is an integer of 2 or more) odor sensors having different response characteristics;
b) reference axis acquisition means for positioning a reference axis corresponding to the reference odor in the m-dimensional space formed by the detection output from the odor sensor by measuring the reference odor by the odor measurement means;
c) Measuring the odor of the evaluation object by the odor measuring means to position the measurement point in the m-dimensional space, and based on the positional relationship between the measurement point and the reference axis, the odor intensity of the evaluation object and the reference odor An evaluation processing means for obtaining the difference in odor quality, the difference in odor quality, and the direction of the difference in odor quality in the evaluation object with respect to the reference odor, and evaluating the evaluation object odor by these, and
Equipped with a,
The evaluation processing means determines whether or not the evaluation target odor is a pass product based on a determination threshold determined for the difference in odor intensity, the difference in odor quality, and the directionality of the difference in odor quality. An odor discriminating apparatus, wherein the acceptance level in the acceptance / rejection determination means can be set in a plurality of stages by a combination of a plurality of determination threshold values .   In the m-dimensional space, the projection point of the measurement point into the plane orthogonal to the reference axis is considered, and the difference in the odor quality is determined by the directivity of the straight line connecting the projection point and the reference point through the reference axis. The odor discriminating apparatus according to claim 1, wherein directionality is expressed.

Images