JP2015128384A - Adzuki bean separation method and adzuki bean separation equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an adzuki bean separation method and adzuki bean separation equipment which can separate non-defective beans from defective beans such as tough beans in the adzuki beans accurately.SOLUTION: An adzuki bean separation method comprises: a step S101 of soaking adzuki beans with water; a step S102 of taking in an image of the adzuki beans absorbed water by an imaging unit; a step S103 of extracting a target signal indicating at least either of the size or color of the adzuki bean from the image signal taken by the imaging unit; and a step S104 of determining a threshold by comparing the target signal with a reference signal set in advance, and selecting the adzuki beans on the basis of the determined results by a sorting unit.

Description

  The present invention relates to a red bean sorting method and a red bean sorting apparatus, and more particularly to a red bean sorting method and a red bean sorting apparatus that can reliably remove defective red beans such as stone beans from red beans.

  Azuki beans used as a raw material for red bean products such as grain bean may contain stone beans with poor water absorption. Stone beans do not swell even when immersed in water, and the hardness tends to remain. If such stone beans are mixed, it may cause boiled unevenness when the red beans are boiled, resulting in a decrease in taste and texture. It has been reported that a complaint of contamination is made, and countermeasures are eagerly desired as a raw material supplier and a product manufacturer. For this purpose, it is desirable to select and reliably remove defective red beans such as stone beans before processing the red beans.

  Conventionally, as a technique for sorting granular materials, a belt-type optical sorter described in Patent Document 1 and an automatic grain sorting device that sorts by grain color described in Patent Document 2 are disclosed. .

  In the belt type optical sorter described in Patent Document 1, light is irradiated to the raw material granular material conveyed by the belt conveyor, the reflected light is received, and the granular material is discriminated based on the received light data. Moreover, in the automatic grain sorter described in Patent Document 2, non-defective / defective products are sorted based on the color of the grain put into the apparatus.

JP 2009-090197 A JP 2004-181271 A

  However, in any of the techniques of Patent Documents 1 and 2, it is difficult to accurately determine a defective red bean such as a stone bean from among the red beans. In other words, in the case of dried beans, it is difficult to distinguish stone beans from good beans. For this reason, it is very difficult to accurately discriminate good quality beans and stone beans from reflected light and color.

  The present invention provides a red bean sorting method and a red bean sorting apparatus capable of accurately sorting good products and bad red beans such as stone beans from among the red beans.

  In order to solve the above-mentioned problem, the red bean sorting method according to the present invention includes a step of adding water to the red beans, a step of capturing an image of the red beans containing water by the imaging unit, and a signal of the video captured by the imaging unit. Extracting the target signal representing at least one of the size and the color of the object by the extracting unit, comparing the reference signal set in advance with the target signal, performing threshold determination, and selecting the red beans based on the determination result And a step of sorting by.

  According to such a configuration, by adding moisture to the red beans, changes in the size due to the swelling of the red beans and the color of the surface of the red beans appear depending on the state of the red beans. For this reason, the image of the red beans containing moisture is taken in by the imaging unit, and the red beans can be selected with high accuracy by comparing the extracted target signal with a preset reference signal and determining the threshold value.

  The red bean sorting device according to the present invention includes an immersion tank, an imaging unit, an extraction unit, a determination unit, and a sorting unit. The dipping tank is a tank for adding water to the red beans. The imaging unit captures an image of the red beans taken out from the immersion tank. The extraction unit extracts a target signal representing at least one of the size and the color of the red beans from the video signal captured by the imaging unit. The determination unit compares the reference signal set in advance with the target signal to determine the threshold value. The sorting unit sorts the red beans according to the threshold judgment result of the judging unit.

  According to such a configuration, the image of the red beans containing moisture is captured by the imaging unit in a state where the red beans are included in the immersion tank. Then, it is possible to accurately select the red beans by determining a threshold value with respect to the reference signal based on at least one of the size and the color of the red beans from the image of the red beans containing moisture.

  In the red bean sorting apparatus according to the present invention, a transport unit that transports the red beans from the immersion tank to the imaging unit may be further provided.

  According to such a configuration, it is possible to continuously and automatically select the red beans by capturing the video of the red beans sequentially transported by the transport unit with the image pickup unit and selecting them according to the result of the threshold determination.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to provide the red bean sorting method and the red bean sorting apparatus which can sort out poor red beans like stone beans and good red beans from red beans with high accuracy.

It is a flowchart which illustrates the red bean sorting method which concerns on 1st Embodiment. It is a figure which illustrates the image | video before and after immersing a red bean in water. It is a figure which illustrates the binary image before and behind immersing a red bean in water. It is a schematic diagram which illustrates the red bean sorting apparatus which concerns on 2nd Embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same members are denoted by the same reference numerals, and the description of the members once described is omitted as appropriate. The drawings are schematic and are not necessarily the same as actual ones.

(First embodiment)
First, the red bean sorting method according to the first embodiment will be described.
As shown in the flowchart of FIG. 1, the red bean sorting method according to the present embodiment extracts a target image by a step of adding water to the red beans (step S <b> 101), a step of capturing a video of the red beans (step S <b> 102). A step (step S103) and a step of sorting red beans (step S104).

  In the process shown in step S101, water is contained in the red beans by immersing the red beans to be selected in water. For example, the red beans are soaked in normal temperature water for about 12 to 16 hours. In addition, although normal temperature water is used here, the water set to predetermined | prescribed temperature and the solution containing a certain solute may be sufficient. In the following description, these are referred to as “water”. By soaking the beans in water for about 12 hours, good beans will swell when they absorb water, and the color of the bean crust will change, but stone beans will not absorb water, so they will not swell and the color of the bean skin will not change. . It should be noted that the soaking for more than 24 hours here is not suitable for the production of grain bean paste because the red beans are excessively swollen and cause skin breakage.

  In the process shown in step S102, an image of red beans containing moisture is captured by a camera (imaging unit). As the camera, for example, a color imaging device that captures a visible light image and converts it into an image signal is used. The camera captures an image of the red beans to be selected (water-containing red beans) and outputs an image signal based on the video.

  In the process shown in step S103, the extraction unit extracts a target signal representing at least one of the size and color of the red beans from the image signal of the video captured by the camera. In the present embodiment, the target signal that represents the size of the red beans is referred to as a “first target signal”, and the signal that represents the color of the red beans is referred to as a “second target signal”. The extraction unit discriminates the red beans region from the image signal output from the camera, and extracts at least one of the first target signal and the second target signal from the image signal in the region.

  For example, the image signal output from the camera is binarized, and the red bean image area is extracted from the image signal. And a 1st object signal and a 2nd object signal are extracted from the signal of the image region of the extracted red beans.

  As a 1st object signal, it is a signal showing the pixel number of the image area of a red bean, for example. The larger the number of pixels, the larger the area of the red beans. The second target signal is, for example, a signal representing the brightness of the pixel in the red bean image area. The second target signal is, for example, a value obtained by averaging the brightness of each pixel in the red bean image area.

  In the process shown in step S104, a threshold value is determined by comparing a preset reference signal with a target signal, and red beans are selected by a selection unit based on the determination result. In the present embodiment, a signal that represents a reference for the size of the red beans is referred to as a “first reference signal”, and a reference that represents the reference for the color of the red beans is referred to as a “second reference signal”. In the threshold determination, at least one of comparison between the first reference signal and the first target signal and comparison between the second reference signal and the second target signal is performed.

  Then, based on the result of the threshold determination, the sorting unit sorts the red beans. That is, the red beans are selected depending on whether the target signal exceeds the reference signal. Examples of the selection of the red beans by the selection unit include (1) selection based on the size of the red beans, (2) selection based on the color of the red beans, and (3) selection based on both the size and the color of the red beans. Which of the sorting methods (1) to (3) is applied is appropriately determined depending on the kind of sorting beans, sorting accuracy, and the like.

  As described above, in the red bean sorting method according to the present embodiment, the red beans are accurately selected by utilizing the change in size and color of the red beans when the red beans to be sorted are swollen in water. Will be able to.

  In the present embodiment, particularly normal red beans and stone beans can be accurately selected. FIG. 2A shows the appearance of the red beans before they are immersed in water. FIG. 2B shows an appearance after the red beans shown in FIG. 2A are immersed in water. As shown in FIGS. 2 (a) and 2 (b), the red beans swell when containing water, and the surface color changes.

  Here, since normal red beans have higher water absorption than stone beans, they swell larger than stone beans. On the other hand, stone beans do not swell much even when immersed in water. In addition, normal red beans are lightly colored in bean hulls by containing moisture. On the other hand, stone beans have low water absorption, so that the color of the bean skin does not become so light even when immersed in water. In the present embodiment, the selection is performed with high accuracy by utilizing the difference in change between normal red beans and stone beans when the red beans are immersed in water.

  FIG. 3A shows a binary image of the red beans before being immersed in water. FIG. 3B shows a binary image after the red beans shown in FIG. 3A are immersed in water. In the present embodiment, after the red beans are soaked in water, an image of the red beans is captured by the camera. Then, the captured image signal is binarized to obtain a first target signal representing the size of the red beans.

  As explained above, the red beans swell and increase in appearance when they contain moisture. For this reason, the area of the binarized image of the red beans increases as shown in FIG. 3 (a) to FIG. 3 (b). Here, a normal red bean swells greatly compared with a stone bean by containing a water | moisture content.

  In the present embodiment, an image after the red beans are soaked in water is captured, and based on the difference in appearance between the normal red beans and the stone beans, the normal red beans and the stone beans are accurately selected using threshold determination. Specifically, at least one of a signal representing the size (area) of the red beans as shown in FIG. 3B and a signal representing the color of the red beans as shown in FIG. By comparing with the standard, the normal red beans and stone beans are accurately selected. This makes it possible to accurately remove stone beans that were difficult to sort by conventional methods.

(Second Embodiment)
Next, a red bean sorting apparatus according to the second embodiment will be described.
As shown in the schematic diagram of FIG. 4, the red bean sorting device 1 according to this embodiment includes an immersion tank 10, a camera 20, an extraction unit 31, a determination unit 32, and a sorting unit 40.

  The immersion tank 10 has a container 11 in which water can be stored. Red beans are put into the container 11 in a state where water is stored in the container 11. A discharge path 11 a is provided in the bottom portion of the container 11. An on-off valve 12 is provided at the tip of the discharge path 11a. In addition, the immersion liquid of 11c can be used as subsequent stewed juice by carrying out primary stock. The control computer 30 to be described later may control the opening / closing operation of the opening / closing valve 12.

  The camera 20 is an imaging unit that captures an image of red beans taken out from the immersion tank 10. As the camera 20, for example, a color imaging device that takes a visible light image and converts it into an image signal is used. The camera 20 takes in an image of red beans containing moisture and outputs an image signal based on the image. It is desirable for the camera 20 to capture an image of each red bean. For example, the camera 20 outputs an image signal as shown in FIG.

  The extraction unit 31 extracts a target signal representing at least one of the size and color of the red beans from the video signal captured by the camera 20. The extraction unit 31 is a calculation unit that processes an image signal. The images captured by the camera 20 include background images in addition to red beans. The extraction unit 31 extracts a red bean image area from the image signal based on the video, and extracts a target signal representing at least one of the size and color of the red bean from the extracted image area signal.

  As illustrated in FIG. 3B, the extraction unit 31 binarizes the image signal and extracts an image region of red beans. The extraction unit 31 may extract a red bean image region by extracting the outline of the red beans from the image signal. The extraction unit 31 extracts a first target signal representing the size of the red beans and a second target signal representing the color of the red beans from the signal of the image region of the red beans.

  The first target signal is, for example, a signal that represents the number of pixels in the image area of the red beans, and corresponds to the area of the red beans. For example, when the image signal of the red beans is extracted by binarizing the image signal, the first target signal is obtained by counting the number of pixels including the extracted binary signal.

  The second target signal is, for example, a signal that represents the brightness of each pixel in the image region of red beans. For example, the second target signal is calculated by calculating the brightness from the color signals (for example, R (red), G (green), and B (blue)) of each pixel in the image area of the red beans, and averaging the brightness of each pixel. Desired. Here, brightness is used as the second target signal, but values of other color components such as saturation may be used. The extraction unit 31 is realized by program processing of the control computer 30. The extraction unit 31 may be configured by hardware.

  The determination unit 32 compares the reference signal set in advance with the target signal (at least one of the first target signal and the second target signal) extracted by the extraction unit 31 and performs threshold determination. In the present embodiment, a first reference signal that represents a reference for the size of the red beans and a second reference signal that represents the reference for the color of the red beans are set as the reference signals.

  The determination unit 32 determines whether or not the target signal exceeds the reference signal. When the first target signal is used, the determination unit 32 performs threshold determination by comparing the first target signal with the first reference signal. When the second target signal is used, the determination unit 32 performs threshold determination by comparing the second target signal with the second reference signal. The determination unit 32 is realized by program processing of the control computer 30. The determination unit 32 may be configured by hardware.

  The sorting unit 40 sorts the red beans according to the threshold determination result of the determination unit 32. The sorting unit 40 includes, for example, a nozzle 41 that ejects air. The selection unit 40 receives a signal based on the result of the threshold determination of the determination unit 32 from the control computer 30 and injects air from the nozzle 41. When the target signal exceeds the reference signal (exceeds the threshold) as a result of the threshold determination by the determination unit 32, or the target signal does not exceed the reference signal (does not exceed the threshold), the control computer 30 selects the selection unit. The signal for injecting air to 40 is sent. When air is jetted from the nozzle 41, the red beans are blown out and sorted by the force of the air.

  The red bean sorting device 1 according to this embodiment further includes a hopper 51 and a belt conveyor 52. The hopper 51 is a container that receives and temporarily stores red beans taken out from the immersion tank 10. The hopper 51 has a funnel shape, and the red beans are discharged onto the belt conveyor 52 one by one.

  The belt conveyor 52 is a transport unit that transports the red beans to the imaging range of the camera 20. The red beans placed on the rotating belt are conveyed one by one and sent to the imaging range of the camera 20. A first container 61 and a second container 62 are disposed on the downstream side of the belt conveyor 52. The red beans that have been transported to the end by the belt conveyor 52 usually fall naturally and are stored in the first container 61. On the other hand, when air is injected from the nozzle 41, the trajectory of the natural fall of the red beans is changed by the force of the air. As a result, the red beans are stored in the second container 62.

Next, the operation of the red bean sorting device 1 according to this embodiment will be described.
First, the on / off valve 12 of the immersion tank 10 is closed by the control computer 30, and water is stored in the container 11. Next, the red beans to be selected are stored in the container 11 and immersed in water. The red beans are soaked in water at room temperature for about 12 to 16 hours, for example. The control computer 30 starts counting the dipping time from the start of the introduction of the red beans, opens the drainage line 11c after the above time has elapsed, and then opens the on-off valve 12.

  When the on-off valve 12 is opened, the red beans stored in the immersion tank 10 are put into the hopper 51. Then, red beans are discharged from the hopper 51 one by one onto the belt conveyor 52. When the red beans are conveyed to the imaging range of the camera 20 by the belt conveyor 52, the camera 20 takes in the video of the red beans and sends an image signal to the extraction unit 31. The camera 20 takes in an image of each red bean sequentially conveyed by the belt conveyor 52 and sends an image signal to the extraction unit 31.

  The target signal of red beans extracted by the extraction unit 31 is sent to the determination unit 32, and a threshold is determined. A signal corresponding to the result of the threshold determination is sent to the selection unit 40. The control computer 30 calculates the timing at which the red beans captured by the camera 20 pass in front of the nozzles 41 based on the timing of capturing the video by the camera 20 and the speed of conveyance of the red beans by the belt conveyor 52. Is sent to the sorting unit 40.

  As a result of the threshold determination, when it is determined that the red beans are normal, the red beans are naturally dropped from the end of the belt conveyor 52 and stored in the first container 61. On the other hand, as a result of the threshold determination, if it is determined that it is a defective red bean such as stone beans, for example, the control computer 30 sends a signal to the sorting unit 40 in accordance with the conveyance timing of the defective red beans. Upon receiving this signal, the sorting unit 40 ejects air from the nozzle 41. As a result, the defective red beans are blown away by the force of air and stored in the second container 62. By performing this operation for each red bean conveyed by the belt conveyor 52, normal red beans are automatically sorted into the first container 61, and defective red beans are automatically sorted into the second container 62.

  The normal red beans selected in the first container 61 are sent to the processing process of the red bean product. In this embodiment, since the red beans are soaked in water, the process immediately proceeds to processing of the red beans product. For example, when producing grain bean, it is not necessary to return the red beans with water again. In addition, the red beans are stored in a vacuum pack or frozen as necessary.

Here, a specific selection example will be described.
The red beans A1 to A10 shown in FIG. 2A are a group of red beans considered normal. Moreover, red beans B1-B10 are a group of red beans considered to be stone beans. When these red beans A1 to A10 and B1 to B10 are immersed in water, they swell as shown in FIG.

  When the image signals of red beans A1 to A10 and B1 to B10 shown in FIGS. 2 (a) and 2 (b) are binarized, they are as shown in FIGS. 3 (a) and 3 (b). In this embodiment, normal azuki beans and stone beans are determined by thresholding the swollen red bean color shown in FIG. 2 (b) and the swollen red bean size (area) shown in FIG. 3 (b). Sort out.

  Regarding the size (area) of the red beans, a threshold value (first reference signal) of the size (area) is set in advance. When the size (area) of the red beans A1 to A10 and B1 to B10 shown in FIG. 3B is compared with the first reference signal, the size of the red beans A1 and A2 among the red beans A1 to A10 considered normal. The size (area) does not exceed the first reference signal. In addition, among the red beans B1 to B10 that are considered to be stone beans, the size (area) of the red beans B9 exceeds the first reference signal.

  As for the color of the red beans, a threshold value (second reference signal) of a value (for example, brightness) representing the color is set in advance. When the lightness representing the colors of red beans A1 to A10 and B1 to B10 shown in FIG. 2B is compared with the second reference signal, the lightness of red beans A1 and A4 among the red beans A1 to A10 considered normal is the first. 1 Reference signal is not exceeded. That is, about red beans A1 and A4, even if immersed in water, it does not swell so much.

  Further, among the red beans B1 to B10 that are considered to be stone beans, the lightness of the red beans B7 and B9 exceeds the second reference signal. That is, for the red beans B7 and B9, the color change of the surface after being immersed in water is large.

  In the present embodiment, normal red beans and stone beans are selected using these results. For example, if the size (area) of the selection target red beans does not exceed the first reference signal and the color (lightness) of the red beans does not exceed the second reference signal, Judge that there is. In addition, when satisfying both the size (area) of the selection target red beans does not exceed the first reference signal and the color (lightness) of the red beans does not exceed the second reference signal, it is a stone bean. You may judge.

  The determination part 32 determines whether it is a stone bean by logical product or logical sum. For example, as the first threshold determination, a case where the size (area) of the red beans does not exceed the first reference signal is true (= 1), a case where the red beans exceed the first reference signal is false (= 0), and a red beans is used as the second threshold determination. The case where the color (lightness) does not exceed the second reference signal is true (= 1), and the case where the color exceeds the second reference signal is false (= 0). The determination unit 32 determines whether or not it is a stone bean by a logical product or logical sum of a result of the first threshold determination and a result of the second threshold determination.

  For example, when the logical product is “1” (when both the first and second threshold determinations are true (= 1)), the determination unit 32 determines that the bean is stone beans. Further, the determination unit 32 may determine that the logical sum is “1” (when at least one of the first and second threshold determinations is true (= 1)), it is a stone bean.

  The determination unit 32 may be configured to be able to switch a logical operation method as to whether or not it is a stone bean. Thereby, the determination method can be switched according to the type of the red beans, the sorting accuracy, etc., and the stone beans can be removed more accurately.

As described above, according to the red bean sorting method and the red bean sorting apparatus according to the present embodiment, it is possible to accurately sort defective red beans such as stone beans and good red beans from among the red beans. In addition, beans selected as defective red beans including stone beans can be used as raw beans for non-grain products.

  Although the present embodiment and specific examples have been described above, the present invention is not limited to these examples. For example, even if a person skilled in the art appropriately added, deleted, or changed a design for each of the above-described embodiments and specific examples, the scope of the present invention is included as long as the gist of the present invention is included. Contained in

  DESCRIPTION OF SYMBOLS 1 ... Azuki bean sorter, 10 ... Immersion tank, 11 ... Container, 11a ... Discharge path, 11b ... Water supply line, 11c ... Drainage line (primary stock of immersion liquid), 12 ... Open / close valve, 20 ... Camera, DESCRIPTION OF SYMBOLS 30 ... Control computer, 31 ... Extraction part, 32 ... Determination part, 40 ... Sorting part, 41 ... Nozzle, 51 ... Hopper, 52 ... Belt conveyor, 61 ... 1st container (good quality red bean collection container), 62 ...・ Second container (Ishizu beans red beans collection container)

Claims (3)

Adding water to the red beans;
Capturing an image of the red beans containing moisture with an imaging unit;
A step of extracting a target signal representing at least one of the size and color of the red beans from the video signal captured by the imaging unit;
Comparing the reference signal set in advance with the target signal to determine a threshold, and selecting the red beans by the selection unit based on the determination result;
A method for sorting red beans. A soaking tank for adding moisture to the red beans;
An image capturing unit for capturing an image of the red beans taken out of the immersion tank;
An extraction unit that extracts a target signal representing at least one of the size and color of the red beans from the image signal captured by the imaging unit;
A determination unit that determines a threshold value by comparing a reference signal set in advance with the target signal;
A sorting unit that sorts the red beans according to the result of the threshold judgment of the judging unit;
A red bean sorting device comprising:   The red bean sorting apparatus according to claim 2, further comprising a transport unit configured to transport the red beans from the immersion tank to the imaging unit.