JP2013024641A - Crack detector of noodle, crack detection method and discrimination system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a crack detector of noodles capable of easily detecting degrees of generation of cracks in noodles.SOLUTION: Illumination light is vertically irradiated to the axis direction of a sample S consisting of noodles from a light source part 1, an image of the sample S is acquired by an imaging part 2 by transmitted light of the sample S, a density histogram is created by an image analysis part 4 on the basis of the image, at least one of an average value MEN, a contrast value CNT, a dispersion value VAR, an energy value EGY, and an entropy value EPY of density levels is calculated as an evaluation value from the density histogram by a determination part 5, and a generation amount of cracks in the sample S is detected by comparing the calculated evaluation value with a reference evaluation value preliminarily calculated for reference noodles whose generation amount of the cracks is known.

Description

The present invention relates to an apparatus and method for detecting cracks in noodles, and more particularly to an apparatus and method for detecting the degree of occurrence of cracks in noodles such as pasta.
The present invention also relates to a noodle sorting system for sorting noodles in which occurrence of cracks of a predetermined amount or more is detected.

In general, noodles such as pasta are manufactured by extruding a noodle material using a die having a nozzle hole having a shape corresponding to its cross-sectional shape and then drying, but the noodle material shrinks during drying. Due to various factors such as internal stress of the noodle material generated during extrusion, cracks along the axial direction may occur on the surface of the noodles.
If such cracks are very small, they are blocked by the expansion of the noodle material when the noodles are boiled, so this does not affect the texture. The appearance will be lost and the quality of the noodles will be degraded when boiled.
For this reason, it is desired to evaluate the degree of cracks generated in noodles.

Conventionally, as a technique for detecting a crack, for example, Patent Document 1 discloses an apparatus for investigating a crack generated on the surface of an outer wall of a concrete structure. An infrared image of the surface of the structure is acquired and image analysis is performed to detect a region with a temperature difference as a crack candidate. The telescope can visually recognize the crack candidate by adjusting the collimating direction and focus of the telescope to the detected crack candidate. To do.
Patent Document 2 discloses an apparatus for evaluating the appearance quality of rice. Paragraph 0065 of Patent Document 2 describes a method for identifying cracked rice from an image of rice using an edge discrimination method. Yes.

Japanese Patent No. 4598881 Special table 2010-540957

However, the apparatus of Patent Document 1 is for investigating cracks in the outer wall of a concrete structure. Even if a crack generated in noodles can be detected by an infrared image, the crack is visually recognized by a telescope. It is not practical to do.
Moreover, in patent document 2, although the crack is detected by extracting an edge component perpendicular | vertical with respect to the major axis of rice, in noodles, most cracks generate | occur | produce along the axial direction of a noodle string. Therefore, if axial streaks are formed on the surface during extrusion molding of the noodle material, it becomes difficult to accurately detect only cracks by the edge discrimination method. Furthermore, if each crack is detected by the edge discrimination method, there is a problem that it takes an excessive burden and time to detect the degree of occurrence of cracks in noodles.

This invention was made in order to solve such a conventional problem, and it aims at providing the crack detection apparatus and method of noodles which can detect easily the generation | occurrence | production degree of the crack in noodles. To do.
Another object of the present invention is to provide a noodle sorting system that sorts noodles with a crack generation amount detected by such a crack detection device being a predetermined value or more.

  The noodle crack detection device according to the present invention is acquired by a light source unit that irradiates illumination light to the noodles, an imaging unit that acquires an image of the noodles by transmitted light of the noodles with respect to the illumination light irradiated from the light source unit, and an imaging unit. And a crack detection unit that detects a generation amount of cracks in the noodles based on the generated density histogram.

The light source unit can irradiate the noodles with multicolor light as illumination light, and the crack detection unit can perform RGB decomposition on the noodle image acquired by the imaging unit and create a density histogram for at least one color of RGB. .
Further, the crack detection unit calculates at least one of an average value of density levels, a contrast value, a dispersion value, an energy value, and an entropy value as an evaluation value from the generated density histogram, and based on the calculated evaluation value It is preferable to detect the amount of cracks generated in the noodles. In this case, the crack detection unit detects the amount of crack generation in the noodles by comparing the calculated evaluation value of the noodles with a reference evaluation value calculated in advance for a reference noodle with a known amount of crack generation. Can do.

The method for detecting cracks in noodles according to the present invention irradiates noodles with illumination light, obtains an image of the noodles by transmitted light of the noodles with respect to the illuminated illumination light, creates a density histogram from the obtained image of the noodles, This is a method for detecting the generation amount of cracks in noodles based on the created density histogram.
Further, the noodle sorting system according to the present invention includes a transport means for transporting noodles, the crack detection apparatus for detecting the amount of cracks generated in the noodles transported by the transport means, and a crack detected by the crack detection apparatus. And a separation means for separating noodles whose generation amount is a predetermined value or more from the transport means.

  According to this invention, the imaging unit obtains an image of the noodles by the transmitted light of the noodles with respect to the illumination light irradiated from the light source unit, and the crack detection unit creates the density histogram from the image of the noodles and the created density histogram Since the amount of cracks generated in noodles is detected based on the above, the degree of cracks generated in noodles can be easily detected.

It is a block diagram which shows the structure of the crack detection apparatus of the noodles which concerns on Embodiment 1 of this invention. 6 is a diagram showing density histograms for RGB colors of an image of noodles subjected to RGB decomposition in Embodiment 1. FIG. It is a block diagram which shows the structure of the noodle sorting system which concerns on Embodiment 2. FIG.

Hereinafter, the present invention will be described in detail based on a preferred embodiment shown in the drawings.
Embodiment 1
FIG. 1 shows the configuration of a noodle crack detection apparatus according to Embodiment 1 of the present invention. The crack detection apparatus has a light source unit 1 arranged to face a sample S made of noodles, and an imaging unit 2 is arranged to face the light source unit 1 so as to sandwich the sample S. Further, the crack detection unit 3 is connected to the imaging unit 2.

The light source unit 1 irradiates the sample S with illumination light from a direction perpendicular to the axial direction of the sample S. For example, an LED light board including a white LED (light emitting diode) as a light source can be used.
The imaging unit 2 acquires a color image of the sample S based on the transmitted light of the sample S with respect to the illumination light irradiated from the light source unit 1 and can be configured by a general-purpose digital camera, a CCD camera, or the like.

The crack detection unit 3 includes an image analysis unit 4 connected to the imaging unit 2 and a determination unit 5 connected to the image analysis unit 4.
The image analysis unit 4 performs RGB decomposition on the color image of the sample S acquired by the imaging unit 2 and creates a density histogram for at least one color of RGB.
The determination unit 5 calculates an evaluation value related to the amount of occurrence of cracks from the density histogram created by the image analysis unit 4, and compares the evaluation value with a reference evaluation value calculated in advance, thereby determining the cracks in the sample S. Determine the amount generated. Here, as the evaluation value, at least one of density level average value MEN, contrast value CNT, dispersion value VAR, energy value EGY, and entropy value EPY can be employed.

  Here, assuming that the frequency of the density level i appearing in the image when the density of each pixel of the image is normalized to, for example, the L gradation of level 0 to level (L-1) is P (i), the average value MEN , Contrast value CNT, dispersion value VAR, energy value EGY, and entropy value EPY can be expressed by the following equations, respectively.

  The average value MEN indicates the average density level in the image of the sample S. The contrast value CNT becomes a large value when the histogram is biased toward a high density level. The variance value VAR takes a large value when the frequency of the density level far from the average value is high. The energy value EGY becomes a large value when the histogram is concentrated at a specific density level. The entropy value EPY becomes a large value when the histogram is distributed over a wide range of density levels.

Next, the operation of the noodle crack detection apparatus shown in FIG. 1 will be described.
First, the average value MEN, the contrast value CNT, the dispersion value VAR, the energy value EGY, and the entropy are compared with reference samples S0, S1, and S2 in which the generation amount of cracks is known in advance using the crack detection device of FIG. The respective evaluation values of the value EPY were calculated.
Samples S0 to S2 are all made of the same material but have different amounts of cracks. S0 is a sample in which no cracks are observed, S1 is a sample in which a small amount of cracks is observed, and S2 is It is a sample in which a large amount of cracks is observed.
A 1000 cd / m ² LED light board was used as the light source unit 1, and a digital camera was used as the imaging unit 2.

The sample S0 was placed at a predetermined position between the light source unit 1 and the imaging unit 2, and illumination light composed of multicolor light was irradiated from the light source unit 1 perpendicular to the axial direction of the sample S0. At this time, the sample S0 was placed on a transparent mounting table disposed between the light source unit 1 and the imaging unit 2, but the sample S0 was directly placed on the light exit surface of the LED light board used as the light source unit 1. It may be placed.
Then, a 24-bit color image of the sample S0 was acquired by the imaging unit 2 by capturing the transmitted light of the sample S0 with respect to the illumination light from the light source unit 1.
The acquired color image was subjected to RGB decomposition by the image analysis unit 4 of the crack detection unit 3, and a density histogram for each of the three RGB colors was created with L = 256 gradations. Further, the determination unit 5 calculates the average value MEN, the contrast value CNT, the dispersion value VAR, the energy value EGY, and the entropy value EPY from each density histogram.
Similarly, also for the samples S1 and S2, color images are acquired, density histograms for RGB colors are created, and the average value MEN, contrast value CNT, dispersion value VAR, energy value EGY, and entropy value EPY are obtained from each density histogram. Calculated.

FIG. 2 shows density histograms for the RGB colors of the samples S0 to S2 created in this way. The horizontal axis represents density, and the vertical axis represents frequency.
Table 1 below shows the average value MEN, contrast value CNT, dispersion value VAR, energy value EGY, and entropy value EPY for the RGB colors of the samples S0 to S2.

As shown in FIG. 2, the peak value of the density level decreases and the broader histogram is obtained as the direction from the sample S0 to the sample S2, that is, as the generation amount of cracks increases, for any color of RGB. Recognize.
For this reason, as shown in Table 1, the average value MEN, the contrast value CNT, and the energy value EGY tend to decrease, while the variance value VAR and the entropy value EPY tend to increase as the amount of occurrence of cracks increases.
The evaluation values of the samples S0 to S2 calculated in advance as described above and shown in Table 1 are used as reference evaluation values for detecting the amount of occurrence of cracks.

  Next, an evaluation value is measured for the sample S for which the amount of cracks to be detected is detected. That is, the sample S is placed at a predetermined position between the light source unit 1 and the imaging unit 2, illumination light is irradiated from the light source unit 1 perpendicular to the axial direction of the sample S, and imaging is performed with the transmitted light of the sample S. A color image of the sample S is acquired by the unit 2. The acquired color image is RGB-separated by the image analysis unit 4 and a density histogram for each of the RGB colors is created. An energy value EGY and an entropy value EPY are respectively calculated.

  Then, in the determination unit 5, for example, the dispersion value VAR of the sample S for the color B of RGB is compared with the dispersion value VAR of the samples S0 to S2 for the color B shown in Table 1 above, and based on the comparison result. The amount of cracks in the sample S is determined. If the size of the dispersion value VAR of the sample S is approximately equal to or smaller than the dispersion value VAR = 49 of the sample S0, it is determined that no crack is generated in the sample S. From the dispersion value VAR = 49 of the sample S0 If the dispersion value VAR of the sample S1 is between 80, it is determined that a crack equal to or smaller than the crack generation amount of the sample S1 has occurred, and the dispersion value VAR of the sample S2 is VAR = 784 from the dispersion value VAR = 80 of the sample S1. If it is between, the crack generation amount of the sample S1 or more and the crack generation amount of the sample S2 or less is determined to be generated, and if it is larger than the dispersion value VAR = 784 of the sample S2, the crack generation amount of the sample S2 or more. It can be determined that the crack is occurring.

  Similarly, by comparing the dispersion value VAR of the sample S with respect to the color R or the color G of RGB with the dispersion values VAR of the samples S0, S1, and S2 with respect to the color R or the color G shown in Table 1, respectively. The amount of cracks generated in S can be determined.

  A sample can also be obtained by comparing any evaluation value of the average value MEN, contrast value CNT, energy value EGY, and entropy value EPY other than the dispersion value VAR of the sample S with the reference evaluation values of the corresponding samples S0, S1, and S2. It is possible to determine the amount of cracks generated in S. However, regarding the average value MEN, the contrast value CNT, and the energy value EGY, it is determined that the smaller the evaluation value of the sample S is, the more cracks are generated than the reference sample, and the entropy value EPY is As the evaluation value of the sample S becomes larger than the reference evaluation value, it is determined that the amount of cracks generated is larger than that of the reference sample.

  For example, the average value MEN of the sample S for the color B is compared with the average value MEN = 160.0 of the sample S0 for the color B shown in Table 1 and the average value MEN = 137.1 of the sample S1. If the value MEN is approximately equal to or larger than the average value MEN = 160.0 of the sample S0, it is determined that there is no crack in the sample S. From the average value MEN = 160.0 of the sample S0 If the average value MEN of the sample S1 is between 137.1, it is determined that a crack equal to or less than the crack generation amount of the sample S1 has occurred, and if the average value MEN of the sample S1 is smaller than 137.1, the sample S1 It can be determined that more cracks than the amount of cracks generated have occurred.

Further, not only one evaluation value but also a plurality of evaluation values among average value MEN, contrast value CNT, dispersion value VAR, energy value EGY and entropy value EPY are obtained, and cracks are generated based on the result of comparison with the reference evaluation value. You may determine the generation amount.
In this case, only when a plurality of evaluation values show the same tendency at the same time, the generation amount of cracks can be determined correspondingly. For example, both the average value MEN and the variance value VAR are calculated for the sample S, and only when the average value MEN of the sample S is larger than the reference evaluation value and the variance value VAR of the sample S is smaller than the reference evaluation value, When the crack generation amount is determined to be smaller than the reference sample crack generation amount, on the other hand, when the average value MEN of the sample S is smaller than the reference evaluation value and the variance value VAR of the sample S is larger than the reference evaluation value Only, it is determined that the sample S has more cracks than the reference sample.
By using a plurality of evaluation values at the same time, it becomes possible to detect the amount of crack generation with higher reliability.

  Alternatively, when a plurality of evaluation values are compared with a reference evaluation value, when at least one evaluation value among the plurality of evaluation values indicates the occurrence of a larger amount of cracks than the reference sample, based on the evaluation value It can also be determined that the amount of cracks generated is greater than that of the reference sample. In this way, for example, a sample that cannot be shipped as a product due to the possibility of occurrence of a crack of a predetermined amount or more can be more reliably determined, and the quality of the shipped product can be ensured. Become.

In the first embodiment, the sample is irradiated with illumination light composed of multicolor light from the light source unit 1 using a white LED, and the color image acquired by the imaging unit 2 is RGB-decomposed by the image analysis unit 4 to obtain RGB. The density histogram for each color has been created. However, the present invention is not limited to this. After the image analysis unit 4 performs RGB decomposition on the color image, only the density histogram for one or two colors of RGB is created. Also good. Moreover, as a light source, not only LED but various light sources can be used.
Further, the light source unit 1 incorporates a monochromatic light source. For example, the sample is irradiated with monochromatic light as illumination light from the light source unit 1 in a dark room, and the image analysis unit 4 directly creates a density histogram from the image acquired by the imaging unit 2. It can also be configured as follows. In this way, RGB decomposition in the image analysis unit 4 becomes unnecessary, and the processing is simplified.

  As described above, according to the crack detection apparatus of the first embodiment, by comparing the evaluation value calculated for the sample S with the reference evaluation values of the reference samples S0 to S2, the cracks in the sample S are compared. The amount of generation can be detected, and the degree of occurrence of cracks in the sample S can be easily detected.

Embodiment 2
FIG. 3 shows the configuration of the noodle sorting system according to the second embodiment. The separation system includes a transport conveyor 11 that transports the sample S and a transport conveyor drive unit 12 that drives the transport conveyor 11. The sample S is placed on the transport conveyor 11 one by one above the transport conveyor 11. The feeder 13 to supply is arrange | positioned. The noodle crack detection apparatus shown in the first embodiment is arranged on the downstream side of the feeder 13 with respect to the conveyance direction by the conveyance conveyor 11. The light source unit 1 of the crack detection device is located below the transfer conveyor 11, the image pickup unit 2 is disposed above the transfer conveyor 11 so as to face the light source unit 1, and the crack detection unit 3 is connected to the image pickup unit 2. Has been.
In addition, the conveyor belt of the conveyance conveyor 11 is formed from the material which has translucency so that the illumination light from the light source part 1 may permeate | transmit.

On the downstream side of the transport conveyor 11, a chute 14 having an inclination such that the placed sample S slides down is disposed, and further on the downstream side of the chute 14, a discharge conveyor 15 that discharges the sample S as a product is provided. The discharge conveyor drive unit 16 is connected to the discharge conveyor 15.
Further, the chute 14 is rotatably attached so that the inclination angle is further increased, and a chute drive unit 17 for rotating the chute 14 is connected to the chute 14. A collection unit 18 for collecting the sample S determined to be abnormal is disposed.

And the control part 19 is connected to the crack detection part 3, the conveyance conveyor drive part 12, the feeder 13, the discharge conveyor drive part 16, and the chute drive part 17 of a crack detection apparatus.
In addition, a conveyance means is comprised by the conveyance conveyor 11, the conveyance conveyor drive part 12, the discharge conveyor 15, and the discharge conveyor drive part 16, and the sorting means is comprised by the chute | shoot 14, the chute drive part 17, and the control part 19. FIG.

  First, the chute drive unit 17 is controlled by the control unit 19, and the rotation position where the chute 14 substantially connects between the downstream end of the transport conveyor 11 and the upstream end of the discharge conveyor 15 as shown by a solid line in FIG. 3. While being maintained at A, under the control of the control unit 19, the transport conveyor 11 and the discharge conveyor 15 are driven by the transport conveyor drive unit 12 and the discharge conveyor drive unit 16, respectively.

  In this state, the samples S are supplied one by one from the feeder 13 onto the conveyor 11. When the sample S is transported by the transport conveyor 11 and reaches just above the light source unit 1 of the crack detection device, the illumination light emitted from the light source unit 1 is irradiated onto the sample S via a translucent conveyor belt, An image of the sample S is acquired by the imaging unit 2 by the transmitted light of the sample S, a density histogram is created by the crack detection unit 3, and the average value MEN, the contrast value CNT, the dispersion value VAR, the energy value EGY, and the entropy value EPY Of these, at least one evaluation value is calculated, and the amount of cracks generated in the sample S is detected based on the calculated evaluation value.

  The amount of crack generation detected by the crack detection unit 3 is output to the control unit 19, and the control unit 19 compares the amount of crack generation in the sample S with a predetermined value set in advance. And the control part 19 judges with the quality which can be shipped as a product with respect to the sample S whose generation | occurrence | production amount of a crack is less than a predetermined value, and from the conveyance conveyor 11 without rotating the chute | shoot 14. It is conveyed to the discharge conveyor 15. That is, the sample S is transported from the downstream end of the transport conveyor 11 onto the chute 14 that is maintained at the rotational position A, slides down the upper surface of the chute 14 and reaches the upstream end of the discharge conveyor 15. Discharged.

On the other hand, for the sample S in which the amount of cracks detected by the crack detection unit 3 is greater than or equal to a predetermined value, the control unit 19 determines that the product does not have a quality that can be shipped as a product, and the chute driving unit 17 And the chute 14 is rotated to the rotation position B indicated by the broken line in FIG. As a result, the sample S cannot reach the upstream end of the discharge conveyor 15 from the downstream end of the transport conveyor 11 and is recovered as an abnormal product by the recovery unit 18 disposed below the chute 14.
After the sample S falls on the collecting unit 18, the chute 14 is returned to the rotation position A by the chute driving unit 17.

In this way, only the sample S having a crack generation amount equal to or greater than a predetermined value is automatically sorted and collected in the collecting unit 18, and only the sample S having a crack generation amount less than the predetermined value is discharged from the discharge conveyor 15. can do.
The noodle crack detection apparatus and noodle sorting system according to the present invention can be widely applied to various types of noodles such as udon, hiyagigi, raw noodles, cold noodles, Chinese noodles, buckwheat noodles, etc. in addition to pasta such as spaghetti .

  DESCRIPTION OF SYMBOLS 1 Light source part, 2 Imaging part, 3 Crack detection part, 4 Image analysis part, 5 Judgment part, 11 Conveyor conveyor, 12 Conveyor drive part, 13 Feeder, 14 chute, 15 Discharge conveyor, 16 Discharge conveyor drive part, 17 chute Drive unit, 18 recovery unit, 19 control unit, S, S0, S1, S2 sample.

Claims (6)

A light source unit for irradiating noodles with illumination light;
An imaging unit that acquires an image of the noodles by transmitted light of the noodles with respect to illumination light emitted from the light source unit;
A crack detection unit that creates a density histogram from the image of the noodles acquired by the imaging unit and detects a generation amount of cracks in the noodles based on the created density histogram. Crack detection device. The light source unit irradiates the noodles with multicolor light as illumination light,
2. The noodle crack detection apparatus according to claim 1, wherein the crack detection unit performs RGB decomposition on the noodle image acquired by the imaging unit and creates a density histogram for at least one color of RGB.   The crack detection unit calculates at least one of an average value of density levels, a contrast value, a dispersion value, an energy value, and an entropy value as an evaluation value from the generated density histogram, and based on the calculated evaluation value, The apparatus for detecting cracks in noodles according to claim 1 or 2, wherein the amount of cracks generated in the noodles is detected.   The crack detection unit detects a crack generation amount in the noodles by comparing the calculated evaluation value of the noodles with a reference evaluation value calculated in advance for a reference noodle with a known crack generation amount. Item 4. A crack detection apparatus for noodles according to Item 3. Irradiate noodles with illumination light,
An image of the noodles is obtained by the transmitted light of the noodles with respect to the irradiated illumination light,
Create a density histogram from the acquired images of the noodles,
A method for detecting cracks in noodles, wherein the amount of cracks in the noodles is detected based on the created density histogram. Conveying means for conveying noodles;
The crack detection device according to any one of claims 1 to 4, which detects the amount of cracks generated in the noodles conveyed by the conveying means.
A noodle sorting system, comprising: a sorting unit that sorts noodles having a crack generation amount detected by the crack detection device equal to or greater than a predetermined value from the transport unit.