CLASSIFICADOR ÓPTICO

BR112022025849B1Active Publication Date: 2026-08-04SATAKE CORP
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
BR · BR
Patent Type
Patents
Current Assignee / Owner
SATAKE CORP
Filing Date
2021-06-14
Publication Date
2026-08-04

Smart Images

  • Figure 00000027_0000
    Figure 00000027_0000
  • Figure 00000028_0000
    Figure 00000028_0000
  • Figure 00000029_0000
    Figure 00000029_0000
Patent Text Reader

Abstract

OPTICAL CLASSIFIER. An optical classifier that includes: an inspection part that is configured to perform optical inspection on granular objects transferred by means of transfer; a judgment part that is configured to judge whether each granular object is a good article or a defective article based on the optical inspection performed by the inspection part; an ejector controller that includes multiple ejector nozzles and is configured to classify the granular objects into defective article and good article by ejecting compressed air onto the defective article from the ejector nozzles;and a classification intensity configuration section that is capable of defining a classification intensity in the ejector controller, wherein the classification intensity consists of several classification intensity levels, and the classification intensity configuration section is capable of predefining, for each of the classification intensity levels, various types of operating configuration parameters that control the operation of the ejector nozzles.
Need to check novelty before this filing date? Find Prior Art

Description

1 / 24 “OPTICAL CLASSIFIER FIELD OF TECHNIQUE

[001] The present invention relates to an optical classifier that allows the user to easily configure the operation of an ejector nozzle of the optical classifier. BACKGROUND OF THE TECHNIQUE

[002] To date, to remove a grain considered a defective article, an optical classifier removes the grain considered a defective article by ejecting compressed air onto it from an ejector nozzle. The compressed air needs to hit a defective article for a long period of time and over a wide range to reliably remove the defective article and obtain high-quality grains after classification. However, this causes a good article to be enveloped in the compressed air and removed along with it, resulting in a reduction in yield. To address this, in order to reduce the number of good articles enveloped in the compressed air as described above and to obtain a high yield, an optical classifier is configured so that the compressed air hits a defective article for a short period of time and over a narrow range (see Patent Literature 1, for example). CITATION LIST Patent Literature

[003] Patent Literature 1

[004] Patent No. JP 4206522 SUMMARY OF THE INVENTION PROBLEM WITH THE TECHNIQUE

[005] In the existing optical classifier described above, individual items such as a delay period and an ejection period can be set for the ejector nozzle operation configuration; however, a specialized maintenance engineer needs time to make settings when adjusting a throughput level between high and low. For example, depending on whether to set a high throughput or not and whether to set Petition 870220118813, dated 12 / 16 / 2022, pp. 74 / 110 2 / 24 normal or high-quality rating, a specialized maintenance engineer needs time to perform the injector nozzle operation configuration through trial and error. Thus, it is sometimes difficult for an average user to easily make the configurations.

[006] In this context, a problem to be solved by the present invention is to provide an optical classifier that allows the user to easily perform an operating configuration of an ejector nozzle of the optical classifier. SOLUTION TO THE PROBLEM

[007] (1) An optical classifier comprising: an inspection part which is configured to perform optical inspection on granular objects transferred by means of transfer; a judgment part which is configured to judge whether each granular object is a good article or a defective article based on the optical inspection performed by the inspection part; an ejector controller which includes multiple ejector nozzles and is configured to classify the granular objects into defective article and good article by ejecting compressed air onto the defective article from the ejector nozzles;and a classification intensity configuration section that is capable of defining a classification intensity in the ejector controller, wherein the classification intensity consists of several classification intensity levels, and the classification intensity configuration section is capable of predefining, for each of the classification intensity levels, several types of operating configuration parameters that control the operation of the ejector nozzles.

[008] (2) The optical classifier, according to claim 1, wherein the ejector controller is able to vary the compressed air ejection ranges by the multiple ejector nozzles according to the multiple classification intensity levels.

[009] (3) The optical classifier according to claim 1 or 2, wherein the ejector controller includes fault area identification means that are capable of identifying a fault area in the defective article and the Petition 870220118813, dated 12 / 16 / 2022, pp. 75 / 110 3 / 24 fault area identification methods are capable of performing, on continuously arranged fault area pixels in a lateral direction, thinning processing to leave a portion of the centerline of the fault area pixels as fault pixels.

[0010] (4) The optical classifier according to claim 3, wherein the multiple types of operating configuration parameters include at least one ejection period by the ejector nozzles and an overlap value indicating a width of an overlap area encompassing the ejection ranges of adjacent ejector nozzles while pressing a boundary line between adjacent ejector nozzles, and the ejector controller causes both adjacent ejector nozzles to perform an ejection operation if at least part of the failure area falls within the overlap area.

[0011] (5) The optical classifier, according to any one of claims 1 to 4, wherein the classification intensity setting part is connected to configuration input means, enabling a user of the optical classifier to select any of the classification intensity levels and is able to adjust a yield according to the classification intensity level thus selected.

[0012] (6) The optical classifier according to claim 5, wherein the configuration input means allows a pre-configuration operation of the operating configuration parameters for each of the classification intensity levels.

[0013] (7) The optical classifier according to claim 6, wherein, in response to the pre-configuration operation of the operation definition parameters at one of the classification intensity levels, the definition input means allow the definition of the operation definition parameters at other classification intensity levels without the pre-definition operation of the same. ADVANTAGEOUS EFFECTS OF THE INVENTION

[0014] According to the invention specified in item (1) above, it is Petition 870220118813, dated 12 / 16 / 2022, pp. 76 / 110 4 / 24 It is possible to remove the granular object, which is the defective item, using the multiple ejector nozzles that are controlled according to the classification intensity level defined by the classification intensity configuration section. Furthermore, the classification intensity levels described above are configured in such a way that the various types of operating configuration parameters that control the operation of the ejector nozzles can be predefined for each of the various classification intensity levels. Consequently, the user can easily configure the optical classifier according to the user's desired output and quality without having to spend time configuring the ejector nozzles as in the past.

[0015] According to the invention specified in item (2) above, it is possible to adjust the compressed air ejection ranges by the multiple ejector nozzles so that they vary according to the multiple levels of sorting intensity. This allows a defective item to be sorted and removed appropriately according to the yield and quality required by the user.

[0016] According to the invention specified in item (3) above, it is possible to identify the fault area in the defective article and perform, in the area of ​​fault pixels arranged continuously in the lateral direction, the thinning process to leave the central line part of the fault pixel area as fault pixels. This process allows narrowing the compressed air ejection bands by the multiple ejector nozzles and improving the yield.

[0017] According to the invention, as specified in item (4) above, the configurable operating parameters for the ejector nozzles include at least the ejection period by the ejector nozzles and the overlap value indicating the width of the overlap area that extends the ejection intervals of the adjacent ejector nozzles, while sandwiching the boundary line between the adjacent ejector nozzles. The optical classifier is configured in such a way that both of the two adjacent ejector nozzles Petition 870220118813, dated 12 / 16 / 2022, pp. 77 / 110 5 / 24 performs the ejection operation if at least part of the defective area on the granular object falls within the overlap area. This allows adjusting the compressed air ejection range by adjusting the overlap value and possibly sorting and removing the defective item according to the yield and quality requested by the user.

[0018] According to the invention, as specified in item (5) above, the user of the optical classifier can select the classification intensity level using the configuration input means. For example, through a console screen as shown in Figure 4, the user can very easily select the classification intensity level according to the yield and quality requested by the user.

[0019] According to the invention, as specified in item (6) above, through a console screen, as shown in Figure 10, for example, an operator such as a maintenance engineer can easily predefine the operation configuration parameters for each of the classification intensity levels.

[0020] According to the invention, as specified in item (7) above, the optical classifier is configured in such a way that, by predefining the operation configuration parameters at one of the classification intensity levels, the operation configuration parameters at other classification intensity levels are automatically defined based on a predetermined configuration rule. This configuration allows saving time and effort required to predefine the operation configuration parameters individually for all classification levels. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] [Figure 1] Figure 1 shows a schematic longitudinal sectional view illustrating an internal structure of an optical classifier in one embodiment of the present invention.

[0022] [Figure 2] Figure 2 shows a block diagram of the control processing means of an optical classifier in a Petition 870220118813, dated 12 / 16 / 2022, pp. 78 / 110 6 / 24 embodiment of the present invention.

[0023] [Figure 3] Figure 3 shows a diagram illustrating a sketch of a classification intensity level in an embodiment of the present invention.

[0024] [Figure 4] Figure 4 shows a diagram illustrating how a classification intensity level adjustment screen is displayed on a liquid crystal display in one embodiment of the present invention.

[0025] [Figure 5] Figure 5 shows a diagram illustrating the grain image processing performed at the time of determining an ejection area in an embodiment of the present invention.

[0026] [Figure 6] Figure 6 shows a conceptual diagram illustrating various operating configuration parameters and ejection areas in primary classification using the optical classifier, in one embodiment of the present invention (classification intensity levels 5 to 3).

[0027] [Figure 7] Figure 7 shows a conceptual diagram illustrating various operating configuration parameters and ejection areas in primary classification using the optical classifier, in one embodiment of the present invention (classification intensity levels 2 and 1).

[0028] [Figure 8] Figure 8 shows a conceptual diagram illustrating various operating configuration parameters and ejection areas in secondary classification and tertiary classification using the optical classifier, in one embodiment of the present invention (classification intensity levels 5 to 3).

[0029] [Figure 9] Figure 9 shows a conceptual diagram illustrating various operating configuration parameters and ejection areas in secondary classification and tertiary classification using the optical classifier, in one embodiment of the present invention (classification intensity levels 2 and 1).

[0030] [Figure 10] Figure 10 shows a diagram illustrating how a configuration screen for maintenance engineers is displayed in Petition 870220118813, dated 12 / 16 / 2022, pp. 79 / 110 7 / 24 a liquid crystal display in one embodiment of the present invention. DESCRIPTION OF THE MODALITIES

[0031] One embodiment of an optical classifier of the present invention will be described below based on the drawings. Figure 1 shows a schematic longitudinal sectional view illustrating an internal structure of an optical classifier 1 in this embodiment. Figure 2 shows a block diagram illustrating a configuration of the control processing means of the optical classifier 1. Optical Classifier Configuration

[0032] As shown in Figure 1, the optical classifier 1 includes, in a machine structure 2: a chute 4 that inclines at an angle of approximately 60 degrees relative to the horizontal position and is configured to transfer grains 100 as a means of transferring the grains 100; a reservoir tank 5 that is designed to hold the grains 100; a vibratory feeder 6 that is configured to transport the grains 100 from the reservoir tank 5 to the chute 4; optical detectors 7a and 7b that are provided so as to vertically sandwich the falling trajectories of the grains 100 falling from the lower end of the chute 4; an ejector 8 that is provided below the optical detectors 7a and 7b; a collecting channel for good articles 9 that is located below the ejector 8 in a position on the same slope line as the chute 4 and designed to receive the grains 100 that fall along the fall paths without receiving a jet of air from the ejector nozzles 80;a defective items collection channel 10 which is arranged side by side with the good items collection channel 9 and designed to collect defective grains that have received a jet of air from the ejector nozzles 80; and an auxiliary defective items collection channel 11 which is designed to collect defective grains that have failed to receive a jet of air from the ejector nozzles 80, have struck a peripheral member and ricocheted back.

[0033] The trough 4 is formed to have the shape of a flat plate without a groove portion so as to allow the grains 100 to slide in a Petition 870220118813, dated 12 / 16 / 2022, pp. 80 / 110 8 / 24 trajectory in the form of a band with a large width. In order to prevent the grains 100 from overflowing from the chute 4 and to prevent the grains 100 from floating off a lower surface of the chute 4 during sliding in the chute 4, a chute cover 4a is provided with a predetermined interval from the lower surface.

[0034] The vibratory feeder 6 has a feeder trough 6a supported on a support part 6b and is configured to be able to feed the chute 4 with the grains 100 by driving a vibratory member, such as an electromagnetismally driven coil 6c. Note that the chute 4 above may have a shape with a grooved portion and cross-sectional shapes, such as a U-shaped cross-section, a V-shaped cross-section and a concave cross-section may be employed as appropriate as the cross-sectional shape of the grooved part.

[0035] The optical detectors 7a and 7b, which serve as part of the protection for carrying out optical controls, are respectively enclosed by box bodies 12a and 12b. The box body 12a, located ahead of the grain fall trajectories 100, is equipped with: a CCD camera 13a for visible light; a NIR camera 14 for near-infrared light; visible light sources 15a and 15b consisting of fluorescent lamps and the like; a near-infrared light source 16a consisting of a halogen lamp and the like; and a back 17a with its face turned towards the optical detector 7b.

[0036] The box body 12b, located behind the grain fall paths 100, is provided with: a CCD camera 13b for visible light; visible light sources 15c and 15d consisting of fluorescent lamps and the like; a near-infrared light source 16b consisting of a halogen lamp and the like; and bottoms 17b and 17c having faces turned towards the optical detector 7a. The window elements 18a and 18b made of transparent glass are fitted into the box bodies 12a and 12b on their sides near the grain fall paths 100.

[0037] Ejector 8 of optical classifier 1 in this embodiment is Petition 870220118813, dated 12 / 16 / 2022, pp. 81 / 110 9 / 24 supplied with 48 ejector nozzles 80 arranged in line at equal intervals in the same direction as the width of the trough 4. Each ejector nozzle 80 includes an ejector solenoid valve 81 which is configured to switch on / off an ejection of compressed air. Here, the number of ejector nozzles 80 can be changed as appropriate according to factors such as the dimensions of the width of the trough 4. The ejector 8 of this embodiment is capable of storing compressed air and is configured to be supplied with air from an air compressor (not shown) via an air supply pipe 22. Note that one or more sub-tanks (not shown) may be provided between the ejector 8 and the air compressor to temporarily store air. This configuration eliminates the risk of air shortage even when the amount of air ejected by the ejector nozzles 80 is large.

[0038] A front door 24 that is vertically hinged by an air cylinder 23 is provided in an inclined front wall of the machine frame 2. This can facilitate maintenance work, such as cleaning. Below the front door 24, a liquid crystal display 25 that functions as a configuration input means in which a console panel consisting of a touch panel and a monitor are combined and a power switch (not shown) is provided. Such a configuration in which the liquid crystal display 25 and the on / off button are arranged at the operator's eye level can facilitate the operation of the machine.

[0039] As shown in Figure 1, a defective article receptacle 27, a good article receptacle 28 and an auxiliary defective article receptacle 29 are provided as receptacles for the separated grains 100. In addition, a sample collection part 30 is provided above the defective article receptacle 27.

[0040] The configuration of the control processing means of optical classifier 1 will be described with reference to Figure 2. The CCD cameras 13a and 13b for visible light and the NIR camera 14 for near-infrared light are electrically connected to a signal processor part 31 for binarization. Petition 870220118813, dated 12 / 16 / 2022, pp. 82 / 110 10 / 24 images acquired. The signal processing section 31 is connected to a memory section 32 that stores images to perform the necessary processing on them. The liquid crystal display 25 is electrically connected to the memory section 32.

[0041] The signal processing part 31 includes at least: an image data acquisition part 33 that temporarily stores image data; a threshold data storage memory 34 that stores threshold data to determine whether the acquired image data is good or defective; a binarization calculation mechanism 35 to binarize the acquired image data; a good / defective judgment part 36 that functions as a judgment part configured to judge whether a target article is good or defective; and an ejector controller 37 that is configured to control the operation of the ejector 8.

[0042] The CCD cameras 13a and 13b for visible light and the NIR camera 14 for near-infrared light are each connected via an I / V converter (not shown) which is configured to convert the detected light intensity value into a voltage value, to an amplifier configured to amplify the voltage value. Then, based on the amplified voltage values ​​(detection signals), the good / defective judgment part 36 compares them with the threshold values ​​stored in the threshold data storage memory 34 to judge whether a target item is good or defective.

[0043] The memory part 32 includes at least: an image data storage memory 38 that stores data from the image data acquisition part 33 as needed; a threshold data calculation mechanism 39 that is configured to calculate a threshold value to judge whether a grain is good or defective based on the image data stored in the image data storage memory 38; a classification intensity configuration part 40 that is configured to perform an ejector operation configuration 8; a receiving mechanism Petition 870220118813, dated 12 / 16 / 2022, pp. 83 / 110 11 / 24 of handling signal 41 to receive a signal generated by a touch operation on the liquid crystal display 25 and send the processed image data to the liquid crystal display 25.

[0044] The good / defective judgment part 36 in the signal processing part 31 is electrically connected to the ejector controller 37, and the ejector controller 37 controls an ejector drive circuit 42 in order to remove a defective grain 100 based on the ejector operating conditions 8 defined in the classification intensity setting part 40 of the memory part 32.

[0045] As shown in Figure 2, the ejector drive circuit 42 above is electrically connected to each of the ejector solenoid valves 81 that switch on / off an ejection of compressed air through the respective ejector nozzles 80. The optical classifier is configured in such a way that the ejector drive circuit 42 inserts a drive signal for each of the ejector solenoid valves 81 to open or close the corresponding ejector solenoid valve 81. As described earlier, the ejector 8 in this embodiment includes the 48 ejector nozzles 80 and the ejector nozzles 80 include, respectively, the ejector solenoid valves 81. Note that, of the 48 ejector solenoid valves 81, Figure 2 shows a first ejector solenoid valve 81a, a second ejector solenoid valve 81b, and a third valve ejector solenoid 81c which are adjacent to each other and can eject compressed air into the entire body of a single grain 100. Classification Processing Configuration

[0046] The classification processing of optical classifier 1 of this embodiment will be described in detail below.

[0047] The optical classifier 1 of this embodiment can predefine the classification intensity at the moment of removal of the grains 100 judged as a defective article by the ejector 8. In other words, as shown in Figure 3, the optical classifier can define the classification intensity in five levels, that is, from level 5 to level 1. Level 1 has Petition 870220118813, dated 12 / 16 / 2022, pp. 84 / 110 Level 1 has a high sorting intensity and can reliably separate defective items, and therefore some good items may be caught in the compressed air ejected from the ejector nozzles 80 and removed along with it. Consequently, Level 1 results in a low yield but can achieve high-quality grains 100. On the other hand, Level 5 has a low sorting intensity and therefore some defective items may not be removed by the ejector 8. Consequently, Level 5 results in a high yield, but the quality of the resulting grains 100 is low. Note that how the ejector 8 operates at each of the sorting intensities will be described in detail later.

[0048] Figure 4 shows an example of a display mode on the liquid crystal display 25. The optical classifier 1 is configured in such a way that a classification intensity level adjustment screen as shown in the drawing is displayed once a user using the optical classifier 1 touches an ejector setting icon 253 with their finger. Then, by touching an up level icon 251 and a down level icon 252, a desired ejector classification intensity level 8 can be easily selected. Note that the classification intensity level can alternatively be selected by touching a round setting level icon 255 shown in the drawing and moving it to a desired level position.

[0049] Based on Figure 5, a description will be given of how a grain 100 being a defective article is identified when the ejector 8 removes the defective article. In step 1, the good / defective judgment part 36 acquires an object image of the grain 100. Then, in step 2, a defective area in the grain 100 is identified by a calibration dose based on a light intensity detection value obtained from the object image. Then, in step 3, a flaw area that meets a defective size to be judged as defective is identified based on a predetermined threshold value defined in advance. Then, in step 4, the flaw pixels of the flaw area arranged continuously in the lateral direction are Petition 870220118813, dated 12 / 16 / 2022, pp. 85 / 110 13 / 24 are subjected to a roughing process to leave only the center line as the failure area. As will be described in detail later, this roughing process can improve performance. Then, in step 5, based on the failure area thus identified, the compressed air ejection intervals by the ejector nozzles 80 are determined according to the previously defined classification intensity level (levels 5 to 1). Note that, as shown in Figure 5, the roughing process in step 4 is not essential, and the ejection range setting in step 5 can be performed without the roughing process as needed.

[0050] Next, a description of the operating configuration parameters will be given for each of the classification intensity levels (Levels 5 to 1) in this mode.

[0051] Figures 6 and 7 show conceptual diagrams illustrating various operating configuration parameters that are predefined for each of the classification intensity levels (levels 5 to 1), compressed air ejection areas for the defective grain 100, and the like. Note that Figures 6 and 7 show a primary classification mode performed by the optical classifier 1 equipped with the flat plate-shaped chute 4. Figures 8 and 9 show a secondary and tertiary classification mode performed by the optical classifier 1 equipped with the U-shaped chute 4. The conceptual diagrams in Figures 6 to 9 show, of the 48 ejector nozzles 80 described above, a first ejector nozzle 80a, a second ejector nozzle 80b, and a third ejector nozzle 80c that are adjacent to each other and can eject compressed air to the entire body of a single grain 100.

[0052] In this mode, as shown in Figures 6 to 9, the following operating items are predefined as the operating configuration parameters for each of the classification intensity levels (levels 5 to 1): (1) a compressed air ejection period; (2) an overlap value that allows adjacent ejector nozzles 80 to be Petition 870220118813, dated 12 / 16 / 2022, pp. 86 / 110 14 / 24 activated when a failure area exists in an overlapping area that extends across the ejection lanes of both ejector nozzles 80 while pressing the boundary between these nozzles; and (3) an ON / OFF adjustment of the roughing process that leaves, as a failure area, only the center line of the failure pixels of the defective article’s failure area arranged continuously in the lateral direction.

[0053] Subsequently, a description will be given of how ejector 8 operates on defective grain 100 in these operating configuration parameters. The Level 5 configuration shown in Figure 6 is the configuration that results in high throughput and is preset to 10 for the ejection period, 0 for the overlap value, and ON for roughing processing. Note that the ejection period 10 is not a value in time units, but an arbitrarily defined value and, in fact, a period in which compressed air hits a falling grain 100 within a range illustrated by a double-headed arrow in the drawing. Needless to say, this value can be set in time units.

[0054] In the Level 5 configuration shown in Figure 6, the overlap value is set to 0. In this case, only the second ejector nozzle 80b, which can eject compressed air to an identified flaw area, operates. Furthermore, in the Level 5 configuration, since the roughing process is set to ON, a linear flaw area is identified, as shown in the drawing, and this area is smaller than the actual flaw area.

[0055] Next, the Level 4 setting shown in Figure 6 has a higher grading intensity and results in a slightly lower yield, but can achieve slightly higher quality 100 grains than the Level 5 setting described above. The ejection period is preset to 15, which is about 1.5 times longer than that of Level 5, the overlap value is preset to 4, and the roughing process is preset to ON. Note that the overlap value of 4 does not have a specific unit, but is an arbitrarily defined value. For example, Petition 870220118813, dated 12 / 16 / 2022, pp. 87 / 110 15 / 24 The overlap value can be the number of pixels in the width direction of an image sensor, which is a value equivalent to the length in the width direction of the overlap area shown in the drawing. Needless to say, this overlap value can be defined as a value in units of distance (dimensions).

[0056] In the Level 4 configuration shown in Figure 6, the overlap value is set to 4. Consequently, as shown in the drawing, the overlap areas encompass the ejection lanes of the first ejector nozzle 80a and the second ejector nozzle 80b adjacent to each other, and the ejection lanes of the second ejector nozzle 80b and the third ejector nozzle 80c adjacent to each other, while the boundary line between the first and second ejector nozzles 80a and 80b and the boundary line between the second and third ejector nozzles 80b and 80c, respectively, are defined. In the embodiment shown in the drawing, only the second ejector nozzle 80b operates, since the failure area subjected to roughing processing falls within the ejection lane of the second ejector nozzle 80b without entering the overlap areas.

[0057] Next, the Level 3 configuration shown in Figure 6 has a higher classification intensity than the Level 4 configuration described above. The ejection period is preset to 20, which is twice that of Level 5, the overlap value is preset to 8, which is twice that of Level 4, and the roughing processing is preset to ON. In the Level 3 configuration, the overlap value is set to 8 and the width of each overlap area is twice that of Level 4. However, in the embodiment shown in the drawing, only the second ejector nozzle 80b operates, since the flaw area subjected to roughing processing falls within the ejection range of the second ejector nozzle 80b without entering the overlap areas.

[0058] Next, the Level 2 setting shown in Figure 7 has a higher classification intensity than the Level 3 setting described above. The ejection period is preset to 25, which is 2.5 times Petition 870220118813, dated 12 / 16 / 2022, pp. 88 / 110 16 / 24 greater than that of Level 5, the overlap value is preset to 14, which is approximately twice that of Level 3, and the roughing process is preset to OFF.

[0059] In the Level 2 configuration shown in Figure 7, the overlap value is set to 14 and the width of each overlap area is approximately twice that of Level 3 as shown in the drawing. In the mode shown in the drawing, an upper right portion of the failure area (roughing process: OFF) falls within the overlap area defined between the second ejector nozzle 80b and the third ejector nozzle 80c. This causes the third ejector nozzle 80c to operate beyond the second ejector nozzle 80b to eject compressed air to the 100 grit within the range of the ejection areas shown in the drawing.

[0060] Next, the Level 1 configuration shown in Figure 7 has a higher classification intensity than the Level 2 configuration described above. The ejection period is preset to 30, which is three times greater than that of Level 5, the overlap value is preset to 20, which is approximately 1.5 times greater than that of Level 4, and the processing thinning is preset to OFF.

[0061] In the Level 1 configuration shown in Figure 7, the overlap value is set to 20 and the width of each overlap area is approximately 1.5 times greater than that of Level 2 as shown in the drawing. In the embodiment shown in the drawing, an upper right portion and a lower left portion of the failure area (roughing process: OFF) fall within the defined overlap area so as to encompass the boundary line between the second ejector nozzle 80b and the third ejector nozzle 80c and the overlap area defined between the second ejector nozzle 80b and the first ejector nozzle 80a, respectively. This causes the third ejector nozzle 80c and the first ejector nozzle 80a to operate beyond the second ejector nozzle 80b to eject compressed air to the 100 grit within the range of the ejection areas shown in the drawing. Petition 870220118813, dated 12 / 16 / 2022, pp. 89 / 110 17 / 24

[0062] Consequently, in the Level 1 configuration, the ejection areas of the ejector nozzles 80 are wider than the outer shape of the target grain 100. However, as the grains 100 flow continuously in a band shape while being close to each other, some good articles near the defective article are caught in the compressed air ejected from the ejector nozzles 80 and removed together. Therefore, Level 1 results in a low yield, but can achieve high-quality grains 100 because defective articles can be safely removed.

[0063] Subsequently, a description will be given of how the ejector 8 operates on the defective grain 100 in the operating configuration parameters in secondary and tertiary classification, which are shown in Figures 8 and 9. Here, secondary and tertiary classification refers to a classification process of feeding the grains 100, obtained after the primary classification performed by the optical classifier 1, into the optical classifier 1 to again perform secondary and tertiary classification processes repeatedly. Furthermore, the secondary and tertiary classification described above uses the chute 4 equipped with multiple U-shaped grooves, each with a width sufficient for a grain 100 to flow down. As shown by the schematic diagrams in Figures 8 and 9, the compressed air ejection area of ​​an ejector nozzle 80 has approximately the same width as the width of each grain 100.

[0064] The Level 5 configuration shown in Figure 8 is the configuration that results in high throughput, and is pre-configured to 10 for the ejection period, 0 for the overlap value, and ON for roughing processing. In the Level 5 configuration, the overlap value is set to 0. In this case, only the second ejector nozzle 80b, which can eject compressed air to an identified flaw area, operates. Furthermore, in the Level 5 configuration, since the roughing processing is set to ON, a linear flaw area is identified, as shown in the drawing, and this area is smaller than the actual flaw area. Petition 870220118813, dated 12 / 16 / 2022, pp. 90 / 110 18 / 24

[0065] Next, the Level 4 setting shown in Figure 8 has a higher grading intensity and results in a slightly lower yield, but can achieve slightly higher quality 100 grains than the Level 5 setting described above. The ejection period is preset to 15 which is about 1.5 times longer than that of Level 5, the overlap value is preset to 0 and the roughing process is preset to ON.

[0066] Next, the Level 3 setting shown in Figure 8 has a higher grading intensity and results in a lower yield, but can achieve slightly higher quality 100 grains than the Level 4 setting described above. The ejection period is preset to 20, which is twice that of Level 5, the overlap value is preset to 0, and the roughing process is preset to ON.

[0067] Next, the Level 2 setting shown in Figure 9 has a higher grading intensity and results in an even lower yield, but can achieve 100 higher quality grains than the Level 3 setting described above. The ejection period is preset to 25 which is 2.5 times longer than that of Level 5, the overlap value is preset to 8 and the roughing process is preset to OFF.

[0068] In the Level 2 configuration shown in Figure 9, the overlap value is set to 8. Consequently, as shown in the drawing, the overlap areas are defined between the first ejector nozzle 80a and the second ejector nozzle 80b adjacent to each other and between the second ejector nozzle 80b and the third ejector nozzle 80c adjacent to each other, respectively. In the mode shown in the drawing, only the second ejector nozzle 80b operates, since the failure area (roughing process: OFF) falls within the ejection range of the second ejector nozzle 80b without entering the overlap areas. Petition 870220118813, dated 12 / 16 / 2022, pp. 91 / 110 19 / 24

[0069] Next, the Level 1 configuration shown in Figure 9 has a higher classification intensity than the Level 2 configuration described above. The ejection period is preset to 30, which is three times greater than that of Level 5, the overlap value is preset to 16, which is twice that of Level 2, and the roughing process is preset to OFF.

[0070] In the Level 1 configuration shown in Figure 9, the overlap value is set to 16 and the width of each overlap area is twice that of Level 2 as shown in the drawing. In the mode shown in the drawing, a left portion of the failure area (roughing process: OFF) falls within the defined overlap area so as to encompass the boundary line between the second ejector nozzle 80b and the first ejector nozzle 80a. This causes the first ejector nozzle 80a to operate beyond the second ejector nozzle 80b to eject compressed air to the 100 grit within the range of the ejection areas shown in the drawing.

[0071] Consequently, in the Level 1 configuration, the ejection areas of the ejector nozzles 80 are wider than the outer shape of the target grain 100. However, since the grains 100 flow continuously in a band shape while being close to each other, some good articles near the defective article are sometimes enveloped in compressed air ejected from the ejector nozzles 80 and removed together. Therefore, Level 1 results in a low yield, but can achieve high-quality grains 100 because defective articles can be safely removed.

[0072] An example of the operating configuration parameters at each of the classification intensity levels in this mode has been described above. Of the 48 ejector nozzles 80, the ejector nozzles 80 located at both of their end portions each have the adjacent ejector nozzle 80 on only one side. Consequently, if the overlap value is set, these ejector nozzles 80 operate with the overlap area defined only on the side where the adjacent ejector nozzle 80 exists. Furthermore, Petition 870220118813, dated 12 / 16 / 2022, pp. 92 / 110 20 / 24 The operating configuration parameters for the ejection period, overlap, and roughing processing described above are predefined by a maintenance engineer, and therefore a user of optical classifier 1 only needs to specify the classification intensity level, as described above, on a configuration screen as shown in Figure 4. Thus, it is possible to significantly reduce the time and effort that were previously required for a maintenance engineer to frequently set the operating configuration parameters and spend time performing the ejector 8 operating configuration by trial and error.

[0073] Note that Figure 10 shows an example of a configuration screen for maintenance engineers and, by tapping a detailed configuration icon 254 in the upper right corner of the screen, a maintenance engineer can proceed to a detailed configuration screen (not illustrated) and set the operating configuration parameters for the ejection period, overlap, and roughing processing at each of the classification intensity levels. Furthermore, once the maintenance engineer sets the operating configuration parameters at one of the levels 5 to 1 when setting the operating configuration parameters, the operating configuration parameters at other classification intensity levels are automatically set according to the level.

[0074] According to this configuration, it is possible to save the time and effort required to enter the operation configuration parameters for all other classification intensity levels when defining and changing the operation configuration parameters. Other modalities

[0075] One embodiment of the optical classifier of the present invention has been described above. However, the present invention is not necessarily limited to the embodiment above, but also includes the following examples of modification, for example.

[0076] For example, although the method of classifying grains Petition 870220118813, dated 12 / 16 / 2022, pp. 93 / 110 21 / 24 Although 100 has been described in the embodiment above, a classification target is not necessarily limited to grains 100. The optical classifier of the present invention can be effectively applied to the case of classifying granular objects, including pieces of resin, such as pellets and beads, and fine articles, such as legumes, medicines, ores and white bait.

[0077] Meanwhile, although the intensity level classification has been defined in five levels from Level 5 to Level 1 in the modality above, the intensity level classification does not necessarily have to be defined in five levels, but can be defined in any number of levels.

[0078] Meanwhile, in the above embodiment, the three types of configuration items—ejection period, overlap, and roughing processing—were predefined at each of the classification intensity levels. Since these configuration items have a significant influence on yield and quality, it is possible to effectively adjust yield and quality by making these items pre-adjustable. However, configuration items are not necessarily limited to these items, but many other types of configuration items can be predefined.

[0079] Meanwhile, although the optical classifier of the above modality has been configured to be capable of level adjustment from low to high throughput as shown in Figure 4, level adjustment is not necessarily limited to such a configuration. For example, the operation configuration parameters can be set for each of the classification intensity levels with emphasis on the quality of the granular objects to be obtained, so as to allow level adjustment from low to high quality. Then, on the level selection screen, as shown in Figure 4, a user can select levels from low to high quality, in addition to the selection of low to high throughput levels.

[0080] Meanwhile, the optical sorter can still be equipped with switching means to ON / OFF the selection function. Petition 870220118813, dated 12 / 16 / 2022, pp. 94 / 110 22 / 24 of the intensity level classification, as shown in Figure 4. This configuration makes it possible to turn off the selection function in response to a user request and to precisely set each operating configuration parameter individually.

[0081] The embodiment of the present invention and some examples of modifications thereof have been described above. However, the above embodiment of the present invention is intended to facilitate understanding of the present invention and is not intended to limit the present invention. The present invention may be modified and improved without departing from its essence, and equivalents thereof are also included in the present invention. Furthermore, the constituents stated in the scope of the claims and in the description may be combined or omitted, provided that such alteration may solve at least some of the above problems or provided that such alteration may bring about at least some of the above effects. LIST OF REFERENCE SYMBOLS OPTICAL CLASSIFIER MACHINE FRAME gutter 4th GUTTER COVER RESERVOIR TANK VIBRATORY FEEDER 6th FEED CHANNEL 6b SUPPORT PART 6c Electromagnetic Driven Coil 7a Optical Detector 7b Optical Detector EJECTOR GOOD ITEM COLLECTION CHANNEL DEFECTIVE ITEM COLLECTION CHANNEL Auxiliary channel for collecting items. Petition 870220118813, dated 12 / 16 / 2022, pp. 95-110 23 / 24 DEFECTIVE 12a BOX BODY 12b BOX BODY 13th CCD Camera 13b CCD CAMERA NIR camera 15th VISIBLE LIGHT SOURCE 15b VISIBLE LIGHT SOURCE 15c VISIBLE LIGHT SOURCE 15d VISIBLE LIGHT SOURCE 16th NEAR-INFRARED LIGHT SOURCE 16b NEAR-INFRARED LIGHT SOURCE 17a FACE BACKGROUND 17b FACE BACKGROUND 17c FACE BACKGROUND 18th WINDOW MEMBER 18b WINDOW MEMBER AIR SUPPLY PIPE FRONT DOOR LIQUID CRYSTAL DISPLAY RECEPTACLE FOR DEFECTIVE ITEM RECEPTACLE FOR GOOD ARTICLE Auxiliary receptacle for defective article. SAMPLE COLLECTION PART SIGNAL PROCESSING PART PART OF MEMORY IMAGE DATA ACQUISITION PART THRESHOLD DATA STORAGE MEMORY BINARIZATION CALCULATION MECHANISM GOOD / FLAWFUL PART OF JUDGMENT Petition 870220118813, dated 12 / 16 / 2022, pp. 96 / 110 24 / 24 Ejector controller IMAGE DATA STORAGE MEMORY THRESHOLD DATA CALCULATION MECHANISM Classification Intensity Configuration Part MANIPULATION SIGNAL RECEPTION MECHANISM Ejector drive circuit INJECTOR NOZZLE 80a FIRST INJECTOR NOZZLE 80b SECOND INJECTOR NOZZLE 80c THIRD INJECTOR NOZZLE Ejector solenoid valve 81a FIRST EJECTOR SOLENOID VALVE 81b SECOND EJECTOR SOLENOID VALVE 81c THIRD EJECTOR SOLENOID VALVE 100 GRAINS 251 LEVEL INCREASE ICON 252 LEVEL REDUCTION ICON 253 Ejector Configuration Icon 254 DETAILED CONFIGURATION ICON 255 CONFIGURATION LEVEL ICON Petition 870220118813, dated 12 / 16 / 2022, pp. 97 / 110

Claims

1 / 2 CLAIMS 1. Optical classifier characterized in that it comprises: an inspection part that is configured to perform optical inspection on granular objects transferred by means of transfer; a judgment part that is configured to judge whether each granular object is a good article or a defective article based on the optical inspection performed by the inspection part; an ejector controller that includes multiple ejector nozzles and is configured to classify the granular objects into defective article and good article by ejecting compressed air onto the defective article from the ejector nozzles;and a classification intensity configuration section that is capable of defining a classification intensity in the ejector controller, where the classification intensity consists of several classification intensity levels, and the classification intensity configuration section is capable of predefining, for each of the classification intensity levels, several types of operating configuration parameters that control the operation of the ejector nozzles.

2. Optical classifier, according to claim 1, characterized in that the ejector controller is capable of varying the compressed air ejection ranges through the various ejector nozzles according to the multiple levels of classification intensity.

3. Optical classifier, according to claim 1 or 2, characterized in that the ejector controller includes fault area identification means that are capable of identifying a fault area in the defective article, and the fault area identification means is capable of performing, on continuously arranged fault area pixels in a lateral direction, Petition 870220118813, dated 12 / 16 / 2022, p. 98 / 110 2 / 2 roughing processing to leave a centerline portion of the fault area pixels as fault pixels.

4. Optical classifier, according to claim 3, characterized in that the various types of operating configuration parameters include at least one ejection period by the ejector nozzles and an overlap value indicating a width of an overlap area that spans the ejection lanes of adjacent ejector nozzles, while sandwiching a boundary line between adjacent ejector nozzles, and the ejector controller causes both adjacent ejector nozzles to perform an ejection operation if at least part of the failure area falls within the overlap area.

5. Optical classifier, according to any one of claims 1 to 4, characterized in that the classification intensity setting part is connected to configuration input means, which allows a user of the optical classifier to select any of the classification intensity levels and is able to adjust a yield according to the classification intensity level thus selected.

6. Optical classifier, according to claim 5, characterized in that the definition input means allow a pre-configuration operation of the operating definition parameters for each of the classification intensity levels.

7. Optical classifier, according to claim 6, characterized in that, in response to the pre-configuration operation of the operating definition parameters at one of the classification intensity levels, the definition input means allow the definition of the operating definition parameters at other classification intensity levels without the pre-definition operation. Petition 870220118813, dated 12 / 16 / 2022, pp. 99 / 110