JP4529700B2 - Grain quality discrimination device - Google Patents

Description

The present invention relates to a grain quality discriminating apparatus that discriminates grain quality such as rice grains, and more particularly to an optical detection unit.

Conventional grain quality discriminating devices convey the grain (hereinafter referred to as “rice grain”) of an object to be measured to an optical detection unit one by one and perform grain quality discrimination based on detected light reception data (imaging data). Known to do. Specifically, the conveying means is configured to rotate a disk in which a plurality of grooves containing rice grains are formed in the peripheral portion by a motor, and the rice grains are put into the grooves one by one and conveyed to the optical detection unit. To do. The optical detection unit irradiates each conveyed rice grain with light, acquires reflected light and transmitted light obtained from the rice grain as the received light data, and sends the received light data to the determination unit. The determination unit determines the quality of rice grains one by one based on the received light data. And this determination result is displayed on a display part, and also what constituted the selection part which sorts a rice grain for every judgment grade.

By the way, while the rice grain measurement is repeated many times, in the optical detection unit, a change in the temperature or the like in the vicinity of the measurement rice grain (near the optical detection position) (determination environment change) occurs due to radiant heat from an irradiation light source or the like. Sometimes. When this determination environment change occurs, the amount of irradiation light at the optical detection position changes so far and accurate quality determination cannot be performed, and conventionally, measures for lowering the determination accuracy due to this light amount change have been taken. This measure is to obtain a correction coefficient corresponding to the change in the light quantity, and correct the change in the light quantity by multiplying the obtained light reception data to prevent a decrease in discrimination accuracy. Specifically, a reference plate having an arbitrary concentration is installed in any one of the grooves of the disk, and light reception data of the reference plate when not used for measurement is stored as “initial data”. On the other hand, the light reception data of the reference plate is acquired as “measurement data” at an arbitrary time when the rice grain measurement is repeatedly performed. Then, the measurement data and the initial data are compared to obtain a brightness difference (density difference) between the measurement data and the initial data, and the correction coefficient is calculated based on the density difference.

Japanese Utility Model Publication No. 5-90353

However, the grain quality discrimination device has a problem due to contamination of the reference plate. This problem is that a rice grain retention part for supplying rice grains to each groove is formed on the peripheral edge (groove) of the disk, and the reference plate embedded in the groove also passes through the rice grain retention part. Become. For this reason, rice bran or the like adheres to the reference plate when it passes through the rice grain retention portion and becomes dirty. Of course, this reference plate is contaminated with a cleaning brush installed in the middle of the disk transport, but it is used in a common drying facility that repeatedly measures about 1000 grains at a time. However, it is difficult to completely remove the contamination of the reference plate, and even if the contamination state of the reference plate is checked, it is left to the operator, and thus the stability of the determination accuracy has been regarded as a problem.
In view of this problem, an object of the present invention is to provide a grain quality discriminating apparatus that can stably and accurately discriminate quality without being affected by dirt on the reference plate.

In order to solve the above problem, according to claim 1,
A grain transport unit that transports the grains one by one through a plurality of grooves provided on the periphery of the rotating disk;
An optical detection unit for obtaining light reception data by irradiating the grain conveyed by the grain conveyance unit;
A determination unit that performs grain quality determination based on optical data from the optical detection unit;
An adjustment unit that performs adjustment by light amount adjustment or correction coefficient calculation corresponding to a change in irradiation light amount of the optical detection unit based on measured light reception data and initial light reception data from a reference plate provided in the groove detected by the optical detection unit;
In the grain quality discrimination device with
A reference plate having a different brightness is provided in the groove, and the adjustment unit calculates a correlation coefficient from a measurement image of the reference plate captured by the optical detection unit and an initial image, and the correlation coefficient is a predetermined condition. When the condition is satisfied, the adjustment by the adjustment unit is performed. On the other hand, when the condition is not satisfied, an image estimated to be a darkly soiled portion of the reference plate until the condition is satisfied The technical means of deleting from the image, recalculating the correlation coefficient, and adjusting by the adjusting unit is taken.

According to the adjustment unit, the reference plate (measurement image) that is being measured is heavily soiled depending on whether the correlation coefficient obtained from the measurement image of the reference plate and the initial image satisfies a predetermined condition (others) It is determined whether or not there is a part that is conspicuous and dirty than the part. The condition is, for example, whether or not the correlation coefficient exceeds a predetermined threshold value (for example, 0.9). If the condition is satisfied, adjustment by the adjustment unit (adjustment by light amount adjustment or correction coefficient calculation) is performed. . On the other hand, if the condition is not satisfied, it is assumed that a darkly soiled portion is generated on the reference plate. In this case, the darkly soiled portion generated on the reference plate is estimated until the correlation coefficient reaches a value satisfying the condition. Are excluded from the measurement image and the initial image, and the correlation coefficient is calculated. As a result, the adjustment by the adjustment unit is performed based on the measurement image and the initial image that do not include the dark and dirty part, so that the adjustment becomes accurate. Regarding the correlation coefficient, the fact that the correlation coefficient between the measured image and the initial image is large means that the initial image of the reference plate in an unused state (the device is new and free from dirt) and has already been measured many times. It shows that the measurement image of the reference plate when performing the same overall tendency (uniform density). That is, since it is assumed that there is no portion that is thinly soiled as a whole but partially darkly soiled, an accurate measurement image adjustment can be performed. On the other hand, when the value of the correlation coefficient is small, there will be a part that is partially dark and dirty, which causes the correlation coefficient value to be reduced. Until the condition is satisfied, the image corresponding to the darkly soiled portion is deleted from the measurement image and the initial image, and the correlation coefficient is calculated.

According to claim 2, the adjustment unit divides the measurement image and the initial image into sections each composed of a plurality of pixels, the correlation coefficient is calculated from the section pixel data of the measurement image and the section pixel data of the initial image, and A technical measure is taken to delete dark and dirty parts in units of sections. Thus, when the correlation coefficient is calculated by dividing the measurement image and the initial image into a plurality of sections, for example, since the calculation can be performed based on one or a plurality of pixel data in each section, The calculation based on the pixel data need not be performed. Therefore, the calculation efficiency is good.

According to claim 3, a technical means is provided in which the section is a plurality of specific sections. Thereby, since it is not necessary to perform the calculation of the correlation coefficient based on all the sections, the calculation efficiency is further improved. In addition, the deletion of a partition is efficient because not all the partitions are targeted.

According to a fourth aspect of the present invention, technical means is adopted in which the partition pixel data is set to an average density value of pixel data of the corresponding partition. As a result, all pixel data in the section, that is, RGB lightness average values, and the like can be taken into account, so that a more accurate correlation coefficient is obtained and the accuracy of detecting the dirty portion of the reference plate is improved.

According to claim 5, the section deletion performed by the adjustment unit is to calculate a correlation coefficient in a section excluding the own section, and to set the own section in the correlation coefficient having the largest value among the correlation coefficients first. Technical measures are taken to delete and then sequentially delete the self-partition at the next largest correlation coefficient in the second order. As a result, the sections can be specified in the order of dark stains, so that the sections can be accurately and efficiently deleted by following this order.

According to a sixth aspect of the present invention, there is provided a technical means for providing a notifying unit for notifying abnormality due to dirt on the reference plate when the number of sections to be deleted by the adjusting unit reaches a predetermined number. Thereby, the minimum level of the measurement image and the initial image necessary for obtaining an accurate correlation coefficient is notified. By this notification, the operator can stop the apparatus and clean the reference plate.

According to a seventh aspect of the present invention, the section takes a technical means in which the imaging pixels of the reference plate are divided vertically and horizontally by an arbitrary number of sections. Since the rotating disk of the grain transport unit is uneven in rotation, the pixels of the captured reference plate image are longer and shorter in the rotation direction, and the number of pixels that image the reference plate is different. Even in such a case, by dividing the imaging pixels of the reference plate vertically and horizontally by an arbitrary number of partitions, the corresponding partition positions of the initial image and the measurement image captured at any time can be made the same. The correlation coefficient value becomes more accurate.

According to the present invention, the correlation coefficient between the measurement image and the initial image is obtained, and it is determined whether or not a darkly stained portion is generated on the reference plate depending on whether or not the correlation coefficient satisfies a predetermined condition. . At this time, if the correlation coefficient is equal to or greater than a predetermined threshold value, it is assumed that there is no darkly stained portion on the reference plate, and the adjustment unit responds to a change in the amount of light emitted from the optical detection unit based on the measurement image and the initial image. Adjustment based on light amount adjustment or correction coefficient calculation is performed. On the other hand, when the above condition is not satisfied, it is determined that a darkly soiled portion exists on the reference plate, images corresponding to this darkly soiled portion are excluded, and a correlation coefficient is calculated with the remaining images to calculate the above condition. Try to meet. Therefore, even when the presence of a darkly soiled portion is estimated on the reference plate, the measurement image and the initial image of this portion are deleted, so the correlation coefficient is calculated based on the accurate measurement image and the initial image. Thus, the adjustment by the adjustment unit becomes accurate, and thus the quality determination accuracy by the determination unit is stably improved.

Hereinafter, the grain quality discriminating apparatus 1 of the present invention constitutes a transport unit 3 that transports a grain (hereinafter referred to as “rice grain S”) of an object to be measured to the optical detection unit 2. The transport unit 3 is of a known disk transport system, and includes a disk 4 provided with a plurality of grooves 4a around which rice grains S enter one by one. The disc 4 is disposed on a base plate 6 supported in an inclined manner by a gantry 6b, and an output shaft 5a of a motor 5 fixed to the gantry 6b is attached to the center of the disc 4 so as to be rotatable. (See FIGS. 1 and 2). In this embodiment, the disc 4 is assumed to rotate in the clockwise direction (arrow Y). Each groove 4a has a bottom portion 4b made of a transparent material and a peripheral edge portion made of an open portion 4c. On the base plate 6, weir portions 6 a for preventing the rice grains S in the grooves 4 a from being released from the open portions 4 c are formed along the peripheral edge of the disc 4. Moreover, the rice grain supply part 7 is comprised in the inclination downward position of the disc 4. As shown in FIG. The rice grain supply unit 7 has a wide flat surface connected to the weir part 6a so that the rice grains S can stay in the upper part of the peripheral edge of the disc 4. In addition, when this rice grain supply part 7 is made into the conveyance start end side of the disc 4, the conveyance termination | terminus side is formed in the notch part 6c which notched the said dam part 6a, and the structure which makes the rice grain S which measurement completed fall from the groove | channel 4a It has become.

The optical detection unit 2 includes a first optical detection unit 2a and a second optical detection unit 2b, and is disposed at a position above the tilt of the disc 4 (see FIGS. 1 and 2). The first optical detection unit 2a is configured to optically detect (image) the upper and side surfaces of the rice grain S that has been conveyed (see FIG. 3). The first optical detection unit 2a has a condenser lens 8, a CCD linear sensor (light receiving sensor) 9, and red (R) / green (G) / blue (irradiation unit 10) above the groove 4a (rice grains S). A light-emitting diode B) is formed, and a condensing lens 11 and a CCD linear sensor (light-receiving sensor) 12 are formed on the side of the groove 4a. Further, below the groove 4a, an irradiation portion (light emitting diode) 13 for irradiating the rice grains S from below is formed in the internal space where the base plate 6 is recessed. In the first optical detector 2a, the weir 6a on the side of the groove 4a is made of a transparent material so that light from the rice grains S can enter the CCD linear sensor 12. The CCD linear sensors 9 and 12 are capable of receiving red (R), green (G), and blue (B) light and are orthogonal to the conveying direction of the rice grains S (the arrow Y). Are arranged to scan.

Next, regarding the configuration of the second optical detection unit 2b, the second optical detection unit 2b is configured to optically detect (image) the lower surface (bottom surface) of the rice grain S (see FIG. 4). The second optical detection unit 2b includes a condenser lens 14, a CCD linear sensor (light receiving sensor) 15, and red (R), green (G), A blue (B) light emitting diode is formed, and an irradiation part (light emitting diode) 17 for irradiating the rice grain S from above is formed above the groove 4a. The CCD linear sensor 15 is also capable of receiving red (R), green (G), and blue (B) light, and scans in a direction perpendicular to the conveying direction of the rice grains S (the arrow Y). It arrange | positions so that it may do.

Next, the characteristic configuration of the present invention will be described.

Two kinds of reference plates having different brightness (density) are formed in the groove 4a of the disk 4. The reference plate was composed of a white first reference plate 18 and a semi-transparent second reference plate 19, and each was embedded in a continuous groove 4a (see FIG. 5). Next, an example of the signal processing unit (control unit) 20 that receives the light reception signals from the CCD linear sensors 9, 12, and 15 and performs signal processing will be described (see FIG. 6). The signal processing unit 20 has a central processing unit (hereinafter referred to as “CPU”) 21 as a central component, an input / output circuit (hereinafter referred to as “I / O”) 22 electrically connected to the CPU 21, and a read-only memory. Section (hereinafter referred to as “ROM”) 23 and a read / write storage section (hereinafter referred to as “RAM”) 24. The I / O 22 is electrically connected to the CCD linear sensors 9, 12, and 15 and is similarly connected to the display unit 25 and the setting start button 22a for displaying the result of the quality determination by the CPU 21. The ROM 23 stores an initial image acquisition program (program A), a measurement image acquisition / correction coefficient determination program (program B), and a lookup table, which will be described later. The signal processing unit 20 is also electrically connected to a circuit (not shown) for controlling the start-up of the motor 5 and the lighting output control of the irradiating units 10, 13, 16, 17 (per known means). Detailed explanation omitted).

The operation of the present invention will be described.

First, the initial image acquisition program (hereinafter referred to as “program A”) is executed (see FIG. 7). The execution purpose of the program A is to acquire the images (initial images) of the first and second reference plates 18 and 19 that are not dirty yet when the grain quality discriminating apparatus 1 of the present invention is in a new state. First, the main body is turned on (step 1) and the setting start button 22a is pushed (step 2). Next, in step 3, upon receiving the ON signal from the setting start button 22a, the CPU 21 starts rotation of the disk 4 by driving the motor 5, and irradiates the first and second optical detection units 2a and 2b. The parts 10, 13, 16, and 17 are turned on. Next, in step 4, the first optical detection unit 2a irradiates light from the irradiation units 10 and 13 to the first reference plate 18 (white color) reached by the rotation of the disc 4. The CCD linear sensor 9 scans and images the first reference plate 18 in a direction (arrow X) orthogonal to the traveling direction of the first reference plate 18, and sequentially sends image (imaging) data to the CPU 21.

Next, in step 5, the CPU 21 divides the image of the first reference plate 18 into specific sections G (1a, 2a,..., 23a) composed of a plurality of pixels K (FIGS. 8 and 9). reference). The CPU 21 obtains the red (R) image data average value (lightness) of the pixel K for each specific section G, and the calculated value (average value of red) of each specific section G is sequentially stored in the RAM 24 as “initial image. As a sequence of numbers in the RAM 24. Next, in Steps 6 and 7, as in Steps 4 and 5, image measurement of the second reference plate 19 (semi-transparent) is performed and every predetermined specific section G (1b, 2b,..., 23b). In addition, the red (R) image data average value (brightness) of the pixel K is obtained, and the calculated value (red average value) of each specific section G is sequentially stored in the “initial image” in the RAM 24. Thus, the program A is completed and the disk is stopped (step 8).

Next, a measurement image acquisition / correction coefficient determination program (hereinafter referred to as “program B”) will be described (see FIG. 10). The purpose of executing the program B is to obtain an accurate correction coefficient without being affected by a darkly dirty portion on the reference plate. The execution is performed in the setting stage before supplying the rice grain S to the rice grain supply unit 7.

In the description of the program B, it is hereinafter referred to as “STEP” for distinction from the “step” of the program A. In STEP1, the setting start button 22a is pressed with the main unit powered on. STEP 2 is performed in substantially the same manner as steps 3 to 7 described in the program A, that is, the disk 4 rotates and the irradiation units 10, 13, 16, and 17 are turned on. And the 1st reference board 18 (white) is imaged, and the average value (lightness) of red for every said specific division G (1a, 2a, ..., 23a) substantially equivalent to the rice grain S area from the acquired image. In the program B, the calculated value is stored as a “measurement image” in the RAM 24 as a numerical sequence. Similarly, the second reference plate 19 (translucent) is imaged, and the average value (brightness) of red for each specific section G (1b, 2b,..., 23b) is calculated from the acquired image, This calculated value is additionally stored in the “measurement image” in the RAM 24.

  Next, in STEP 3, the CPU 21 calculates the “correlation coefficient A” between the “initial image” and the “measurement image” in the specific section (46 sections) acquired in the previous STEP. The calculation is performed by reading the initial image and the measurement image stored in the RAM 24 and using a known calculation formula (see FIG. 11). In the present invention, two types of reference plates having different brightness (density), that is, for example, a white first reference plate 18 and a translucent second reference plate 19 are used. The data is distributed in a band shape in a certain linear direction in the scatter diagram, so that an accurate correlation coefficient between the two can be obtained. This is because, when the brightness is one type of reference plate, a band-shaped data distribution cannot be obtained and a correlation coefficient cannot be obtained.

  Next, in STEP 4, “correlation coefficient A> 0.9” is determined based on the correlation coefficient A obtained in STEP 3. If YES at this time, since the correlation between the initial image and the measurement image is high, it is determined that there are no darkly contaminated portions on the first and second reference plates 18 and 19, and the process proceeds to STEP 10 (described later). On the other hand, if NO, since the correlation is low, it is recognized that there is a dark and dirty part on the reference plate, and the process proceeds to STEP5.

In STEP 5 and later, a dark and dirty portion is specified, and this portion is sequentially excluded and the correlation coefficient is recalculated. The specific procedure is to calculate a correlation coefficient in a section other than the self section for the specific section constituting the initial image and the measurement image. For example, the section 1a in the initial image and the measurement image is set as a self section, and the correlation coefficient is calculated based on the other 45 sections (2a,..., 23a, 1b,. Is “correlation coefficient B”. Subsequently, in the same manner, the correlation coefficient B is calculated with the remaining sections excluding the section 2a as the self section, and then the correlation coefficient B is calculated with the section 3a as the self section. Perform sequential calculations. Next, go to STEP6.

  In STEP 6, the correlation coefficients B for each of the 46 sections obtained in STEP 5 are first arranged in descending order. After the rearrangement, the largest correlation coefficient B is stored as the correlation coefficient A, and the self section in the correlation coefficient B is recognized as a “dark and dirty section” and the section is set as an “exclusion section”. . Next, proceed to STEP7.

  In STEP7, the excluded section recognized in STEP5 is excluded from the specific section. Thereby, the remaining 45 sections become specific sections. Next, the process proceeds to STEP8.

  In STEP 8, it is determined whether or not the condition of “excluded section = 8” is satisfied. This is to check whether the number of sections excluded so far has reached eight sections, and the number of the first and second reference plates 18 and 19 is too large because there are too many dark sections (excluded sections). The purpose is to check whether or not a measurement image (section) for accurate correlation coefficient calculation can be secured by reducing the area. If the exclusion zone = 8 (if YES), the first and second reference plates 18 and 19 are too dirty (there are many exclusion zones), and the display section 25 reads “Clean the reference plate. "Is displayed and the rotation of the disk 4 is stopped. On the other hand, if NO, the process returns to STEP 4 and the determination of “correlation coefficient A> 0.9” is made. If the determination is NO again, it is determined that there is still a dark and dirty section in the specific section, and the above-described STEPs 5, 6, 7, and 8 are performed. As a result, one other dark and dirty section is excluded from the specific section and the correlation coefficient A is obtained again based on the remaining specific section, and the process returns to STEP 4 to determine “correlation coefficient A> 0.9”. In this way, until the condition “correlation coefficient A> 0.9” is satisfied, until the “excluded section = 8”, the dark and dirty sections are identified and sequentially deleted, and the correlation coefficient by recalculation is determined. The comparison with A is repeated.

  Next, in STEP 10, the specific section that has not been excluded in the previous STEP is recognized as the “final section”.

  Next, in STEP 11, the data (red lightness average value) of each final section of the initial image and the measurement image is read, and each lightness average value of the initial image and the measurement image is obtained. That is, the average red brightness value Y in the final section of the initial image and the average red brightness value X in the final section of the measurement image are obtained.

  Next, in STEP 12, the correction coefficient is determined. The correction coefficient is determined by the preset lookup table (correspondence table). The lookup table is obtained in advance with correction coefficients for combinations of values of the lightness average value Y (initial image) and the lightness average value X (measurement image). According to the lookup table of STEP 12, the correction coefficient for the lightness average value Y and the lightness average value X can be immediately determined without performing arithmetic processing or the like.

  The correction coefficient is determined as described above, and thereafter, the disk 4 is automatically stopped and the program B is terminated (STEP 13).

  After the correction coefficient is determined in this way, the next measurement is the rice grain S. In measuring rice grains, first, rice grains S are supplied to the rice grain supply unit 7 and a measurement start button (not shown) is turned ON. Then, the disk 4 rotates, and the rice grains S of the rice grain supply unit 7 enter the groove 4a one by one and are conveyed to the first and second optical detection units 2a and 2b. In the first optical detection unit 2a, the irradiated light hits the rice grains S from the irradiation units 10 and 13, and the rice linear S is scanned by the CCD linear sensors 9 and 12. Thereby, a plane image and a side image of the rice grain S are obtained, and each image data is sequentially stored in the RAM 24. Rice grains S that have passed through the first optical detection unit 2a are imaged by the second optical detection unit 2b. That is, in the second optical detection unit 2b, the irradiated light hits the rice grain S from the irradiation units 16 and 17, and the rice linear S 15 is scanned by the CCD linear sensor 15. Thereby, a back image of the rice grain S is obtained, and the image data is sequentially stored in the RAM 24. The CPU 21 sequentially reads the image data of each rice grain S stored in the RAM 24 to determine the quality, and stores the determination data in the RAM 24. When the imaging by the first and second optical detection units 2a and 2b and the quality determination by the CPU 21 are completed, the CPU 21 totals the quality determination data stored in the RAM 24 and causes the display unit 25 to display the total result. Thus, the quality discrimination of one lot of rice grains is completed, the next lot of rice grains is supplied again, and the quality discrimination is started. Note that the execution time of the program B of the present invention (correction coefficient calculation) may be executed after measurement of a predetermined lot, or may be executed every several hours.

  The features of the present invention will be further described. Regarding the execution of the program B, it is possible to make it less likely to be adversely affected by uneven rotation of the disk 4 by the following method. The rotation speed of the disc 4 is not always constant, and when the first and second reference plates 18 and 19 are imaged by the CCD linear sensor 9, the width of one pixel in the rotation direction Y actually differs. This is because the pixel width changes depending on how fast the disks (first and second reference plates 18, 19) pass during one scan. Specifically, when the rotation speed is high, the pixel width becomes short, and when the rotation speed is low, the pixel width becomes long. In the present invention, each pixel in the Y direction and X direction of the groove 4a is equally divided by the captured images of the first and second reference plates 18 and 19 (see FIG. 8). That is, the number of pixels in the Y direction and the X direction in one section is not determined in advance, but the pixels in the measured groove image are equally divided. Thereby, every time the program B is executed (correction coefficient calculation), the position of the specific section (1a,..., 1b,...) Is always the same without moving back and forth in the Y direction. For this reason, even if rotation unevenness occurs, an accurate correlation coefficient can be obtained, and there is an effect that the correction coefficient does not vary.

  Hereinafter, modifications of the embodiment of the present invention will be described. In the above description, when the correlation coefficient A is calculated, only the average value of red in a specific section is used. However, the present invention is not limited to this, and the average value of green and / or blue By using an average value, a more accurate correction coefficient can be obtained. In order to obtain a more accurate correction coefficient, the CCD linear sensor 15 may also be used, and the images on the back side of the first and second reference plates 18 and 19 may be used.

  In the reference plate of the present invention, the first and second reference plates having different brightness levels are provided in the separate grooves 4a in the above embodiment. However, the present invention is not limited to this. For example, half of the reference plate is the first reference plate. The remaining half of the plate 26a may be a second reference plate 26b having a lightness different from that of the first reference plate 26a, and may be a single reference plate 26 having different lightness portions (see FIG. 12).

  Further, regarding the adjustment unit, in the above description, the correction coefficient calculation corresponding to the change in the amount of light emitted from the optical detection unit is calculated based on the measurement light reception data (measurement image) and the initial light reception data (initial image) from the reference plate. The optical detection data from each rice grain is corrected by the calculated value. However, the adjustment unit is not limited to the adjustment by the correction coefficient, and may adjust the light amount itself based on the density difference between the initial image and the measurement image. That is, the output of the irradiation unit is changed based on the density difference to change the amount of light irradiated to the rice grains.

  In the above, the condition of the correlation coefficient is compared with correlation coefficient A> 0.9. However, the value to be compared can be changed as appropriate, and a positive value closer to 1 is measured with the initial image. The correlation of the image is strong.

The longitudinal cross-sectional view of the grain quality discrimination | determination apparatus of this invention is shown. The top view of the grain quality discrimination device of the present invention is shown. FIG. 3 is a longitudinal sectional view of a first optical detection unit of the present invention. The longitudinal cross-sectional view of the 2nd optical detection part of this invention is shown. The reference board of this invention is shown. The signal processing part of this invention is shown. The flow of the program A of this invention is shown. The 1st reference board and 2nd reference board in this invention are shown. The division in this invention is shown. The flow of the program B of this invention is shown. It is explanatory drawing which calculates a correlation coefficient from an initial image and a measurement image in this invention. The modification of the reference | standard board in this invention is shown.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Grain quality discrimination | determination apparatus 2 Optical detection part 2a 1st optical detection part 2b 2nd optical detection part 3 Conveyance part 4 Disk 4a Groove 4b Bottom part 4c Opening part 5 Motor 5a Output shaft 6 Base board 6a Weir part 6b Base 6c Notch Unit 7 Rice grain supply unit 8 Condensing lens 9 CCD linear sensor 10 Irradiating unit 11 Condensing lens 12 CCD linear sensor 13 Irradiating unit 14 Condensing lens 15 CCD linear sensor 16 Irradiating unit 17 Irradiating unit 18 First reference plate (white)
19 Second reference plate (translucent)
20 signal processing unit 21 central processing unit (CPU)
22 Input / output circuit (I / O)
22a Setting start button 23 Read-only memory (ROM)
24 Read / write memory (RAM)
25 Display unit 26 Reference plate 26a First reference plate 26b Second reference plate G Section K Pixel S Rice grain (grain)

Claims (7)

A grain transport unit that transports the grains one by one through a plurality of grooves provided on the periphery of the rotating disk;
An optical detection unit for obtaining light reception data by irradiating the grain conveyed by the grain conveyance unit;
A determination unit that performs grain quality determination based on optical data from the optical detection unit;
An adjustment unit that performs adjustment by light amount adjustment or correction coefficient calculation corresponding to a change in irradiation light amount of the optical detection unit based on measured light reception data and initial light reception data from a reference plate provided in the groove detected by the optical detection unit;
In the grain quality discrimination device with
A reference plate having a different brightness is provided in the groove, and the adjustment unit calculates a correlation coefficient from a measurement image of the reference plate captured by the optical detection unit and an initial image, and the correlation coefficient is a predetermined value. When the condition is satisfied, the adjustment by the adjustment unit is performed. On the other hand, when the condition is not satisfied, an image that is estimated to be a darkly soiled portion of the reference plate is satisfied until the condition is satisfied. A grain quality discriminating apparatus which is deleted from an initial image, recalculates a correlation coefficient, and performs adjustment by an adjustment unit. The adjustment unit divides the measurement image and the initial image into sections each composed of a plurality of pixels, and the correlation coefficient is calculated from the section pixel data of the measurement image and the section pixel data of the initial image. 2. The grain quality discrimination device according to claim 1, wherein the deletion is performed in units of sections. The grain quality discriminating apparatus according to claim 2, wherein the section is a plurality of specific sections. The grain quality determination device according to claim 2 or 3, wherein the partition pixel data is an average density value of pixel data of a corresponding partition. The section deletion performed by the adjustment unit is to calculate a correlation coefficient in a section excluding the section itself, delete the section in the correlation coefficient having the largest value among the correlation coefficients, and then The grain quality discrimination device according to claim 3 or 4, which is sequentially performed so as to delete the self-compartment in the correlation coefficient that is larger than the second. The grain quality determination according to claim 2 or 5, further comprising a notifying unit for notifying abnormality due to dirt on the reference plate when the number of sections to be deleted by the adjusting unit reaches a predetermined number. apparatus. The grain quality discriminating apparatus according to any one of claims 2 to 6, wherein the sections are obtained by dividing the imaging pixels of the reference plate vertically and horizontally by an arbitrary number of sections.

Images