Apparatus for automatically processing shellfish

Abstract

An apparatus for automatically processing shellfish according to the present invention comprises: a physical washer which becomes a passage through which shellfish products are moved and protrudes toward the shellfish products on the passage to wash the shellfish products; a fluid washer which is connected to the passage to be adjacent to the physical washer, and sprays a fluid toward the shellfish products to remove fine foreign substances from the shellfish products; a quality inspector which extends from the passage to be adjacent to the fluid washer, and determines whether the shellfish products are in good condition or defective condition and carries out a radiation test; and a sorter which is connected to the passage of the quality inspector and picks up and separately discharges the shellfish products according to preset sizes.

Description

Automatic shellfish processing equipment

[0001] The present invention relates to an automatic shellfish processing device, and more particularly, to a shellfish washing, inspection and classification device that automatically washes and inspects shellfish and classifies them by size.

[0002] In general, as the demand for automated processes related to aquatic products increases, the demand for automated processes using robots for processing, recognition, sorting, and packaging is also increasing.

[0003] Among aquatic products, oysters are a type of seafood with weak microbial and sanitary conditions, yet the adoption rate of automated processing equipment is significantly low, and concerns about biological contamination of oysters are serious.

[0004] The oyster production process can be broadly categorized into foreign material screening, size sorting, and panning. These are labor-intensive tasks requiring significant human resources. Despite this, the adoption of automated equipment and systems related to this process in the seafood manufacturing industry remains extremely low. Furthermore, oyster production is often considered undesirable, and the existing workforce is aging, leading to a serious shortage of production personnel. Therefore, automated oyster production equipment that can prevent oyster contamination while minimizing human resources is urgently needed.

[0005] The present invention is an invention devised to solve the problems of the above-described prior art, and provides an automatic shellfish processing device that can improve productivity by uniformizing production volume and shortening process time through an automated process line that automatically washes and classifies oysters.

[0006] The tasks of the present invention are not limited to the tasks mentioned above, and other tasks not mentioned will be clearly understood by those skilled in the art from the description below.

[0007] In order to achieve the above object, the automatic shellfish processing device of the present invention is an automatic shellfish processing device that automatically performs the processes of washing, inspecting, and packaging the shellfish aquatic product that is fed, and may include a physical washer that is protruded toward the shellfish aquatic product along a path along which the shellfish aquatic product moves and washes the shellfish aquatic product, a fluid washer that is adjacent to the physical washer and connected to the path and sprays a fluid toward the shellfish aquatic product to remove fine foreign substances attached to the shellfish aquatic product, a quality inspector that is adjacent to the fluid washer and extends from the path and determines whether the shellfish aquatic product is good or bad and inspects it for radioactivity, and a sorter that is connected to the path of the quality inspector and picks up the shellfish aquatic product by preset sizes and separates and discharges it.

[0008] And the physical washing machine may include a zigzag washing section in which a groove is formed to discharge foreign substances that fall off from the shellfish marine product, and in which the shellfish marine product moves together with the fluid generated at the end, and in which a section in which at least a portion is inclined downward and bent is formed, and a cylindrical washing section having a cylindrical shape and is rotatable about a central axis, and including a panel formed to protrude upward so that the shellfish marine product moved in the zigzag washing section rotates.

[0009] In addition, the physical washing machine may include a dewatering transport section having a slope that rises from the end of the zigzag washing section toward the cylindrical washing section and transporting the shellfish marine product from the zigzag washing section to the cylindrical washing section, and a unit transport section that is formed long and transports the shellfish marine product from the cylindrical washing section to the fluid washing machine.

[0010] And the bubble washing unit may include an air generating member that is formed long along the path, has a downward slope toward the center, and provides air toward the shellfish aquatic product, and a washing water generating member that has an upward slope at the end of the air generating member and moves the shellfish aquatic product, and a plurality of washing nozzles are arranged at a predetermined interval at the top.

[0011] In addition, the classifier may include a robot unit that is connected to the fluid cleaner and determines the size of the moved shellfish by comparing it with preset data and gripping and moving the shellfish, a sorting conveyor unit that is formed long in the form of a conveyor belt adjacent to the robot unit and moves the shellfish settled through the robot unit to an end, and a sorting conveyor unit that is provided in the form of a conveyor belt at the lower end of the sorting conveyor unit and transports the shellfish by accommodating them in an array tray that individually accommodates them.

[0012] And the quality inspector may include an image acquisition unit that acquires image data on the shellfish aquatic product transferred from the fluid washing machine, an image processing unit that stores the image data on the shellfish aquatic product and inputs it into an automatic classification model to generate classification information on good / defective products, and generates classification information capable of catching the shellfish aquatic product through image processing on the image data, and then maps the classification information and the classification information.

[0013] In addition, the robot unit can use the mapped classification information and phage information received from the image processing unit to phage the shellfish aquatic product and classify it into a classification conveyor unit assigned according to the classification information.

[0014] And the above classification conveyor section can be slidably connected to a classification tray having a plurality of grooves formed at the end to accommodate shellfish products moved to the lower part.

[0015] In addition, the above classification tray is formed to penetrate along the direction in which the shellfish aquatic product is inserted, and a door that can be opened and closed at the bottom is slidably coupled so that when the door is opened, the shellfish aquatic product located on the arrangement tray can fall into the arrangement tray.

[0016] And the robot part is formed by combining a plurality of arm members with a head that rotates around a rotation axis, and a vision camera is provided adjacent to the robot part to detect the position and size of the shellfish marine product in order to grasp the shellfish marine product.

[0017] According to the present invention, the automatic shellfish processing device includes a physical washer, a fluid washer, and a sorter, automatically sorting and processing shellfish products. Therefore, it has the advantage of being unaffected by the skill level of the operator, thereby ensuring uniform production volume.

[0018] And since oysters are automatically sorted and processed, labor costs are reduced and the oyster washing and sorting process time is shortened, which has the advantage of improving productivity.

[0019] The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the claims.

[0020] FIG. 1 is a drawing for explaining the overall configuration of an automatic shellfish processing device according to one embodiment of the present invention;

[0021] FIG. 2 is a drawing for explaining a zigzag washing unit in an automatic shellfish processing device according to one embodiment of the present invention;

[0022] FIG. 3 is a drawing for explaining a cylindrical washing unit in an automatic shellfish processing device according to one embodiment of the present invention;

[0023] FIG. 4 is a drawing for explaining a vortex washing unit in an automatic shellfish processing device according to one embodiment of the present invention;

[0024] FIG. 5 is a drawing for explaining a bubble washing unit in an automatic shellfish processing device according to one embodiment of the present invention;

[0025] FIG. 6 is a drawing for explaining the structure of a quality inspector in an automatic shellfish processing device according to one embodiment of the present invention;

[0026] FIG. 7 is a drawing for explaining the system configuration of a quality inspector in an automatic shellfish processing device according to one embodiment of the present invention;

[0027] FIG. 8 is a flowchart showing a method of inspecting using a quality inspector in an automatic shellfish processing device according to one embodiment of the present invention;

[0028] FIG. 9 is a drawing for explaining a control unit in an automatic shellfish processing device according to one embodiment of the present invention;

[0029] FIG. 10 is a drawing for explaining an output image including classification information of a classification model in an automatic shellfish processing device according to one embodiment of the present invention;

[0030] FIG. 11 is a flowchart for explaining a method for creating a classification module (CM) in an automatic shellfish processing device according to one embodiment of the present invention;

[0031] FIG. 12 is a flowchart for explaining a method for classifying and processing shellfish aquatic products based on a learning model in an automatic shellfish processing device according to one embodiment of the present invention;

[0032] FIG. 13 is a drawing for explaining a classifier in an automatic shellfish processing device according to one embodiment of the present invention;

[0033] FIG. 14 and FIG. 15 are drawings for explaining a robot section, a classification conveyor section, and a sorting conveyor section in an automatic shellfish processing device according to one embodiment of the present invention;

[0034] FIG. 16 and FIG. 17 are drawings for explaining a robot part in an automatic shellfish processing device according to a first modified embodiment of the present invention;

[0035] FIG. 18 is a drawing for explaining a robot part in an automatic shellfish processing device according to a second modified embodiment of the present invention;

[0036] FIG. 19 is a drawing for explaining the support part and soft fingers of the robot part in the automatic shellfish processing device according to the second modified embodiment of the present invention;

[0037] FIG. 20 is a drawing for explaining the nail section of the robot section in the automatic shellfish processing device according to the second modified embodiment of the present invention.

[0038] <Explanation of symbols>

[0039] 100: Physical cleaner 200: Fluid cleaner

[0040] 300: Classifier 400: Quality Inspector

[0041] 1300,2300: Robot Department

[0042] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings. First, in adding reference numerals to components in each drawing, the present invention will be described in detail with reference to the same drawings.

[0043] It should be noted that components are given the same reference numerals as much as possible even if they are shown in different drawings. In addition, in describing the present invention, preferred embodiments of the present invention will be described in detail with reference to the attached drawings for related known configurations or functions. First, when adding reference numerals to components in each drawing, it should be noted that identical components are given the same reference numerals as much as possible even if they are shown in different drawings. In addition, in describing the present invention, if it is determined that a specific description of related known configurations or functions may obscure the gist of the present invention, the detailed description will be omitted.

[0044] Various aspects of the present invention are described below. It is to be understood that the inventions presented herein may be embodied in a wide variety of forms, and that any specific structure, function, or both presented herein are merely exemplary. Based on the inventions presented herein, one of ordinary skill in the art will appreciate that one aspect presented herein may be implemented independently of any other aspects, and that two or more of these aspects may be combined in various ways. For example, a device may be implemented or a method may be implemented using any number of the aspects described herein. Furthermore, such a device may be implemented or such a method may be implemented using structures, functions, or structures and functions other than or in addition to one or more of the aspects described herein.

[0045] The present invention will be described in detail with reference to the drawings.

[0046] FIG. 1 is a drawing for explaining the overall configuration of an automatic shellfish processing device according to one embodiment of the present invention.

[0047] As illustrated, the automatic shellfish processing device according to the present invention has a path formed along which shellfish products are moved, and the moving force can be moved together with a conveyor belt and a fluid flowing in one direction.

[0048] The configuration of the present invention can largely include a physical washer (100), a fluid washer (200), a quality inspector (400), and a classifier (300).

[0049] The physical cleaner (100) becomes a path along which shellfish aquatic products move, and protrudes a predetermined amount toward the shellfish aquatic products along the path to filter out foreign substances that fall from the shellfish aquatic products and remove large foreign substances.

[0050] A fluid washer (200) connected to a physical washer (100) can provide a stream of water toward shellfish to remove fine foreign substances attached to the shellfish.

[0051] Shellfish products that have been washed can be classified by size through a classifier (300) and moved to a cold storage for storage.

[0052] Let us explain the configuration of the present invention in detail while looking at the drawings.

[0053] Figures 2 and 3 are drawings showing a physical washing machine (100).

[0054] FIG. 2 is a drawing for explaining a zigzag washing unit (120) in an automatic shellfish processing device according to one embodiment of the present invention, and FIG. 3 is a drawing for explaining a cylindrical washing unit (160) in an automatic shellfish processing device according to one embodiment of the present invention.

[0055] The physical washing machine (100) places shellfish on a fluid generating member (122) for washing, and water is sprayed from the fluid generating member (122) to move the shellfish, and its configuration may include a zigzag washing unit (120), a dewatering transport unit (140), a cylindrical washing unit (160), and a unit transport unit (180).

[0056] First, the zigzag washing unit (120) includes a fluid generating member (122) and a bending path member (124). The fluid generating member (122) has an upper part that is open in the shape of a square frame (244), and a drainage net (162) having a grid structure with holes formed on the bottom surface can be provided.

[0057] At least one nozzle (not shown) is formed at the end of the fluid generating member (122) to spray a stream of water onto the shellfish aquatic product, and when the stream of water is sprayed through the nozzle, the shellfish aquatic product on the fluid generating member (122) can move to the next path.

[0058] The fluid generating member (122) is connected to the bending path member (124).

[0059] The bending path member (124) has a shape of a frame (244) that is formed long and has an open upper portion, and can have a shape that is bent at least once along the length direction.

[0060] And the bending path member (124) is formed on a thin panel (15) (PANNEL) along the path, and foreign substances that fall while the shellfish aquatic product moves along with the fluid can be filtered out.

[0061] Here, the panel (15) may be formed to be inclined along the path, and may have a shape inclined in the direction in which the fluid flows from the bottom to the top.

[0062] This may be to allow the shellfish seafood to come into contact with the panel (15) and sediment foreign substances without disturbing the flow of fluid.

[0063] And, since the bottom surface of the bending path member (124) is formed with a drainage net (162), the settled foreign substances can be drained together with the fluid.

[0064] The fluid used here may be water used to wash away foreign substances attached to shellfish products, such as washing water, alkaline water, or purified water.

[0065] In addition, the bending path member (124) is formed to be inclined from one side to the other, and specifically, the ground clearance is formed to gradually decrease, thereby allowing fluids and shellfish to move naturally.

[0066] A dehydration transport unit (140) may be connected to the end of the bending path member (124).

[0067] The dewatering transport unit (140) is in the form of a conveyor belt, and the bottom is formed with a drainage net (162), so that it can have an upward path to pull shellfish products settled on the upper side and move them upward.

[0068] A cylindrical washing section (160) may be provided at the lower end of the dehydration transport section (140).

[0069] The cylindrical washing unit (160) has a cylindrical shape and is rotatable around a central axis, and may have a shape in which a plurality of panels (15) are formed protruding toward the top.

[0070] The cylindrical washing unit (160) can rotate the settled shellfish and cause foreign substances to fall off by centrifugal force, and has an opening formed in a portion of the circumference so that the shellfish and marine products from which foreign substances have been fallen off can fall outward toward the opening.

[0071] In the cylindrical washing unit (160), a drainage net (162) is formed on the floor adjacent to the opening. Here, the drainage net (162) formed in the cylindrical washing unit (160) is formed to have a downward slope toward the unit transfer unit (180), so that it can be moved to the unit transfer unit (180) by its own weight and the fluid can be discharged to the outside.

[0072] Shellfish products that have escaped through the above opening can be moved by the unit transport unit (180).

[0073] The unit transport section (180) is in the form of a conveyor belt, and a drainage net (162) can be formed on the floor. In addition, a panel (15) is formed perpendicular to the floor along the length direction to divide the path.

[0074] This is to transport a certain amount of shellfish products that have been separated from the cylindrical washing unit (160), and the amount of shellfish products placed between the plurality of panels (15) can be controlled according to the speed at which the unit transport unit (180) transports them, thereby moving the shellfish products to the fluid washing machine (200).

[0075] Meanwhile, the fluid washer (200) may include a vortex washing unit (220) and a bubble washing unit (240).

[0076] A vortex washing unit (220) may be formed on the inner surface of a semicircular receiving unit (not shown) and a vortex nozzle (222) may be formed to provide a stream of water toward shellfish products.

[0077] This will be explained in detail through Figure 4.

[0078] FIG. 4 is a drawing for explaining a vortex washing unit (220) in an automatic shellfish processing device according to one embodiment of the present invention.

[0079] As shown, the vortex nozzle (222) may be formed as a pair of unit nozzles facing each other on the inner surface of the receiving portion, and may be arranged in multiple units at a set interval along the inner surface of the receiving portion.

[0080] And the vortex nozzle (222) has holes formed eccentrically in the direction in which the shellfish is moved, so that it becomes the driving force for the shellfish to be moved, and foreign substances in the shellfish can be removed through the pressure of the shellfish stream through the vortex.

[0081] In addition, the vortex washing unit (220) has a groove formed on the lower surface, and a drainage net (162) is formed in the groove, so that sediments can be loaded on the upper surface of the drainage net (162).

[0082] And the drain net (162) is provided with a handle so that the worker can lift the drain net (162) to discharge foreign substances.

[0083] A circulation hole (224) may be formed continuously at the bottom of the vortex nozzle (222).

[0084] The circulation hole (224) is formed vertically below the vortex nozzle (222) and may have a structure connected to a separate pump to recover water inside the receiving portion and supply it back to the vortex nozzle (222).

[0085] The end of the vortex washing unit (220) can be connected to the bubble washing unit (240).

[0086] FIG. 5 is a drawing for explaining a bubble washing unit (240) in an automatic shellfish processing device according to one embodiment of the present invention.

[0087] As shown, the bubble washing unit (240) may include an air generating member (242) and a washing water generating member.

[0088] The air generating member (242) is formed long along the above path in the shape of a drainage net (162), has a downward slope toward the center of the bubble washing section (240), and can provide air toward shellfish seafood.

[0089] Already in the bubble washing section (240), shellfish products slide downward along the path of the air generating member (242), and at this time, the air generating member (242) generates air to wash the shellfish products.

[0090] Meanwhile, the washing water generating member includes a washing transport member (248) in the form of a conveyor belt, and the washing transport member (248) has an upward slope at the end of the bubble washing member (240) to move shellfish products, and a plurality of washing nozzles (246) are arranged at predetermined intervals at the top to spray washing water.

[0091] Specifically, the washing nozzles (246) may take the form of being arranged in multiple numbers on the frame (244), and the frame (244) may be formed long so that multiple washing nozzles (246) may be arranged along the width direction of the washing transport part (248).

[0092] And the frame (244) can be formed in multiple numbers along the direction in which the washing transport part (248) material rises, and three can be formed according to an embodiment of the present invention.

[0093] And the bottom surface of the washing transport unit (248) is formed with a drainage net (162) so that the washing water sprayed from the washing nozzle (246) can be drained.

[0094] As shown in FIG. 5, three frames (244) are formed along the length direction of the washing transport member (248), and the heights of the frames (244) are all the same, so that the washing water transport member has an upward path, and the frame (244) positioned adjacent to the end of the washing water transport member can be formed with a relatively narrower distance from the drainage net (162) of the washing water transport member than the other frames (244).

[0095] Specifically, if the washing water transport member has a path that rises from one end to the other end along the length direction, three frames (244) are formed in the direction from one end to the other end of the washing water transport member, and are sequentially referred to as a first frame (244), a second frame (244), and a third frame (244), and the washing nozzle (246) formed in the second frame (244) is relatively closer to the drainage net (162) of the washing transport member (248) than the washing nozzle (246) formed in the first frame (244), and the washing nozzle (246) formed in the third frame (244) and the drainage net (162) can be formed relatively closer than the washing nozzle (246) formed in the second frame (244).

[0096] Meanwhile, the quality of shellfish products that have been washed through a fluid washer (200) can be confirmed through inspection.

[0097] Specifically, the quality inspector (400) extends from the above path adjacent to the fluid washer (200) and can determine whether shellfish seafood is good or bad and conduct radioactivity tests.

[0098] The quality inspector (400) will be described with reference to FIGS. 6 to 8.

[0099] FIG. 6 is a drawing for explaining the structure of a quality inspector in an automatic shellfish processing device according to one embodiment of the present invention, FIG. 7 is a drawing for explaining the system configuration of a quality inspector in an automatic shellfish processing device according to one embodiment of the present invention, and FIG. 8 is a flowchart showing a method of inspecting using a quality inspector in an automatic shellfish processing device according to one embodiment of the present invention.

[0100]

[0101] The inspection target (410) can enter the quality inspection device (400) through the sample transfer table (420). The radiation detector (430) can detect radioactive materials present in the inspection target (410), and the scale (440) can measure the weight of the inspection target (410) and transfer it to the quality inspection device (400). Here, the inspection target (410) refers to shellfish marine products, and may refer to marine products including the shellfish marine products.

[0102] In another embodiment, the inspection object (410) may be provided to the quality inspector (400) in a preset quantity without the need to measure the weight with a scale (440), or weight information of the inspection object (410) may be provided to the quality inspector (400) from an upper computing device.

[0103] The quality inspector (400) can determine the radioactive nuclides and radioactivity concentrations present in the sample (100), as described later, and can be configured to display information including the detected nuclides and concentrations and output an alarm if the determined radioactivity concentration is above a set alarm level.

[0104] FIG. 7 is an exemplary block diagram showing a system configuration including a radiation inspection device according to one embodiment of the present disclosure.

[0105] As illustrated in FIG. 7, the quality inspection device (400) may include a radiation detector (430) and a control unit (540). The radiation detector (430) may be configured to measure the energy spectrum of radiation generated from the inspection target (410).

[0106] The control unit (540) may be configured to determine a radioactive nuclide (e.g., Cs-137, I-131, etc.) based on the measured energy spectrum, determine a region of interest to apply to the measured energy spectrum according to the determined nuclide, determine a count rate within the region of interest, and determine a radioactivity concentration of the inspection target based on the determined count rate within the region of interest.

[0107] The radiation detector (430) may include a plastic scintillator, a photomultiplier tube (PMT) (515), a summing circuit (520), and a multi-channel analyzer (MCA) (530).

[0108] The plastic scintillator may be, for example, a PVT scintillator (510), and may be configured to convert incident radiation generated from an inspection target (410) into light of a visible light wavelength. The PMT (515) may be configured to focus visible light from the PPVT scintillator (510) and convert it into an electrical signal. The summing circuit (520) may sum the signals received from the PMT (515) and perform noise removal and amplification functions. The MCA (530) may be configured to process the electrical signal from the summing circuit (520) to generate an energy spectrum of incident radiation from the inspection target and transmit it to the control unit (540).

[0109] The control unit (540) may be a computing device for analyzing the energy spectrum measured by the radiation detector (430), and may be implemented as a personal computer (PC), a dedicated computing device for radiation inspection, etc. In addition, the control unit (540) may be connected to a central control system (550) through a predetermined wired or wireless communication network, and may be configured to perform a radiation inspection of an inspection target according to a command from the central control system (550) and provide the inspection result (e.g., measurement completion signal, inspection pass / fail, radiation concentration, etc.) to the central control system (550).

[0110] In one embodiment, when the quality inspector (400) is implemented in manual mode, the quality inspector (400) may be equipped with a scale (440) for measuring the weight of the inspection target. In another embodiment, when the quality inspector (400) is implemented in automatic mode, the weight information of the inspection target sample may be provided from a central control system (550) or may be processed so that a preset quantity of the inspection target sample is input into the quality inspector (400).

[0111] FIG. 8 is an exemplary flowchart illustrating a method for examining radioactivity in an inspection subject according to one embodiment of the present disclosure.

[0112] As illustrated in FIG. 8, when an inspection subject (410) enters a quality inspection device (400) (300), a radiation detector (430) measures radioactivity within the inspection subject (100) for a preset time (e.g., 1 minute) (305), and (in manual mode) a scale (440) can measure the weight of the inspection subject (410) (310).

[0113] The control unit (540) can determine the presence or absence of radioactive material based on the contamination suspicion determination criteria set based on the background values ​​measured by the radiation detector (430) (315). These contamination suspicion determination criteria can be expressed by the following formula:

[0114] [Mathematical Formula 1]

[0115]

[0116] Here, k can be set based on the false alarm rate presented by ANSI (American National Standards Institute).

[0117] In step 315, if the criteria according to mathematical formula 1 are not satisfied, the control unit (540) may determine that there is no radioactivity, output the inspection target (410), and terminate the radioactivity inspection. In step 315, if the criteria according to mathematical formula 1 are satisfied, the control unit (540) may perform an analysis on the measured energy spectrum as described below.

[0118] The control unit (540) can determine radioactive nuclides based on the measured energy spectrum (325). To this end, the control unit (540) can be configured to determine radioactive nuclides by determining the similarity between the reference energy spectrum and the measured energy spectrum for each nuclide using the Maximum Likelihood Estimation (MLE) technique.

[0119] For example, reference energy spectrum information for Cs-137, I-131, etc., which may be nuclides of interest for an inspection target (410) such as aquatic products, may be prepared in advance, and the control unit (540) may determine which nuclide, such as Cs-137 or I-131, is present in the inspection target (410) by comparing the measured energy spectrum with each of these reference energy spectra.

[0120] In one implementation example, reference energy spectrum information for a specific nuclide can be determined by applying a Monte Carlo computational simulation technique. For example, a quality inspector (400) can use a computational simulation utilizing the MCNP (Monte Carlo N-Particle Transport) code to obtain reference energy spectra for each nuclide, such as Cs-137 and I-131, when they exist in a test subject (e.g., aquatic food), and apply the obtained spectrum to determine radioactive nuclides.

[0121] Once the radioactive nuclide is identified in step 325, the control unit (540) can determine a region of interest to apply to the measured energy spectrum based on the identified nuclide (330). To this end, the control unit (540) can determine the region of interest based on the Compton edge energy, Compton maximum energy, and the resolution of the radiation detector (430) for the identified nuclide.

[0122] Compton edge energy refers to the highest energy that can be transmitted to the detector by complete backscattering during Compton scattering, and is a value that can be calculated according to a given formula using the peak energy of the corresponding nuclide.

[0123] Compton maximum energy can be determined to be identifiable depending on the resolution and energy range of the radiation detector (430). For example, when measuring high-energy gamma rays, the Compton maximum energy can be accurately identified in the measured energy spectrum. However, when measuring relatively low-energy gamma rays, the resolution may be low, making it difficult to accurately identify the Compton maximum energy. Accordingly, the present disclosure proposes a method for determining the Compton maximum energy for low-energy gamma rays based on the Compton maximum energy determined for high-energy gamma rays.

[0124] To this end, the quality inspector (400) can first determine the Compton maximum energy in high-energy gamma rays for a specific nuclide through computer simulation, and then use this to define a functional relationship to represent the Compton maximum energy in the low-energy region. The Compton maximum energy is a value that varies depending on the resolution of the radiation detector (430), and the resolution of the radiation detector (430) can be determined according to the following resolution function:

[0125] [Equation 2]

[0126]

[0127] Here, E represents the Compton edge energy or the gamma-ray energy absorbed by the radiation detector (430), FWHM (Full Width at Half Maximum) represents the resolution of the radiation detector (430) at the gamma-ray energy E, and a, b, and c may represent the degree of spread of a peak by the corresponding gamma-ray energy within the spectrum. a, b, and c may be determined through Monte Carlo computational simulation for the quality inspector (400).

[0128] The quality checker (400) can determine the fitted energy to Compton maximum energy relationship based on this resolution function.

[0129] After the radioactivity test for shellfish seafood is completed through the above process, the shellfish seafood goes through a process of judging whether it is good or bad through images.

[0130] The control unit (540) described for this purpose controls the operations of the image acquisition unit (710), the image processing unit (not shown), and the robot unit (320).

[0131] FIG. 9 is a diagram for explaining a control unit in an automatic shellfish processing device according to an embodiment of the present invention, FIG. 10 is a diagram for explaining an output image including classification information of a classification model in an automatic shellfish processing device according to an embodiment of the present invention, FIG. 11 is a flowchart for explaining a method for generating a classification module (CM) in an automatic shellfish processing device according to an embodiment of the present invention, and FIG. 12 is a flowchart for explaining a method for classifying and processing shellfish aquatic products based on a learning model in an automatic shellfish processing device according to an embodiment of the present invention.

[0132] Figure 9 is a drawing for explaining the control process of the control unit (540).

[0133] When explaining in FIGS. 9 to 12, the shellfish aquatic product of the present invention will be explained with oysters as the target.

[0134] The control unit (540) causes image data acquired by the image acquisition unit (710) including the vision camera (VISION-CAMERA) to be input into the classification module (820), so that classification information is generated by the classification module (820).

[0135] Then, the control unit (540) causes the image data to be input into the phage information generation module (230), so that the location (xy coordinates) of at least one shellfish aquatic product included in the image data is generated as phage information by the phage information generation module (230), and then the classification information and the phage information are mapped.

[0136] The control unit (540) transmits the mapped classification information and phage information to the robot unit (320).

[0137] In addition, the control unit (540) receives movement amount information of the sorting conveyor unit (340) from the DAQ (Data acquisition) controller that controls the movement of the sorting conveyor unit (340) described later and transmits it to the robot unit (320), so that the gripper of the robot unit (320) can grip shellfish seafood.

[0138] Data transmission and reception between the control unit (540) and the remaining components (image acquisition unit (710), image processing unit (not shown), robot unit (320)) can be performed using an Ethernet communication protocol or a GPIO (General Purpose Input Output) communication protocol.

[0139] Specifically, communication between the control unit (540) and the image processing unit (not shown), the control unit (540) and the robot unit (320), the control unit (540) and the DAQ controller (A), the image acquisition unit (710) and the image processing unit (not shown) is performed using an Ethernet communication protocol, and communication between the sorting conveyor unit (340) and the DAQ controller (A), the DAQ controller (A) and the robot unit (320), and the robot unit (320) and the gripper can be performed using a GPIO communication protocol.

[0140] Next, an automatic classification model for shellfish aquatic products will be described with reference to FIG. 10. The automatic classification model for shellfish aquatic products (CM, hereinafter referred to as the "classification model") is programmed into a classification module (820) and determines whether shellfish aquatic products are good or bad.

[0141] The control unit (540) allows the image data acquired by the image acquisition unit (710) to be input into the classification module (820), thereby training the classification model (CM).

[0142] The image data used in the learning process is called the original image. The original image is analyzed using a classification model (CM) to derive classification information, which is used to classify good and bad products.

[0143] Referring to Figure 10, the classification model (CM) is a learning model generated through learning (deep learning / machine learning) using learning data. The classification model (CM) includes multiple layers, each of which performs multiple operations.

[0144] Additionally, multiple layers are connected through weights. In other words, the computational results of one layer are weighted and input to the computation of the next layer. In other words, the classification model (CM) performs multiple operations connected by weights between multiple layers. For convenience of explanation, the multiple operations connected by weights between multiple layers of the classification model (CM) will be referred to as "weighted operations."

[0145] The multiple layers of the classification model (CM) may include a convolution layer (CL), a deconvolution layer (DL), a pooling layer (PL), and at least one fully-connected layer (FL). The convolution layer (CL) performs a convolution operation that applies weights using a predetermined filter (kernel) to the input original image or the feature map (FM) of the previous layer, and generates a feature map (FM) or feature image by performing an operation using an activation function.

[0146] In addition, the deconvolution layer (DL) performs a deconvolution operation that applies weights using a predetermined filter (kernel) to the feature map (FM) of the previous layer, and performs an operation using an activation function to generate a feature map (FM) or a feature image. The pooling layer (PL) performs a downsampling operation or an upsampling operation using a predetermined filter (kernel) to generate a feature map (FM) or a feature image.

[0147] A fully connected layer (FL) performs operations on multiple inputs from the previous layer using an activation function to produce at least one output. Here, the filter (kernel) can be a matrix where each element is a weight. Examples of activation functions include the sigmoid, hyperbolic tangent (tanh), exponential linear unit (ELU), rectified linear unit (ReLU), leaky ReLU, maxout, minout, and softmax.

[0148] As illustrated in Fig. 10, when an original image (TI) is input, the classification model (CM) performs multiple operations in which multiple layers of weights are applied to the input original image (TI) to derive object information [(x, y, w, h), C01=0.888, C02=0.101, C03=0.011] including a bounding box (BB) (x, y, w, h) and a classification probability (C01=0.888, C02=0.101, C03=0.011), and outputs an output image (TO) including the derived object information. The bounding box (BB) defines the area occupied by the learned object.

[0149] That is, the bounding box (BB) defines the area occupied by the learned object through the center coordinate (x, y), width (w), and height (h). The classification probability represents the conditional probability that an object within the bounding box (BB) is classified into the class of the learned object. For example, the classes of the learned objects are assumed to be large shellfish (C01), medium shellfish (C02), and small shellfish (C03). The bounding box (BB) (x, y, w, h) defines the area occupied by one of the objects, i.e., large shellfish (C01), medium shellfish (C02), and small shellfish (C03), in the original image. Small shellfish refers to those that are 20px*20px or less in pixel size, large shellfish refers to those that are 40px*40px or more in pixel size, and medium shellfish refers to those that are 20px*20px to 40px*40px in pixel size.

[0150] The classification probability represents the probability that an object within the bounding box (BB) is a large shellfish (C01=0.888), a medium shellfish (C02=0.101), and a small shellfish (C03=0.011).

[0151] For example, as shown in Fig. 10, if the object information output by the classification model (CM) is “(x, y, w, h), C01=0.888, C02=0.101, C03=0.011”, it indicates that the probability that the object within the area box (BB) is a large shellfish aquatic product is 89%, the probability that it is a medium shellfish aquatic product is 10%, and the probability that it is a small shellfish aquatic product is 1%.

[0152] The classification module (820) classifies the size of the object according to the classification probability of the object information, and outputs classification information including an indicator indicating the classified object and an area box (x, y, w, h) of the classified object. The output classification information is provided to the phage information generation module (230). For example, as in the above-described assumption, if the object information output by the classification model (CM) is “(x, y, w, h), C01=0.888, C02=0.101, C03=0.011”, the classification module (820) determines that the object in the area box (BB) is a large shellfish aquatic product because the probability that the object in the area box (BB) is a large shellfish aquatic product is the highest at 89%, and outputs an indicator indicating a large shellfish aquatic product and the corresponding area box (x, y, w, h) as classification information.

[0153] The classification module (820) generates classification information by distinguishing between good and bad shellfish products based on the size of the shellfish products, which are objects classified by the classification model (CM), and the ratio of foreign substances contained in the shellfish products. Good shellfish products are shellfish products that contain foreign substances below the standard, and bad shellfish products are shellfish products that contain foreign substances above the standard. Here, foreign substances refer to shellfish shells, sand, etc. that are introduced during the shelling process (the process of peeling shellfish) during the shellfish processing process. The standard may be separately set by the user according to the quality grade of the shellfish product. Meanwhile, the classification model (CM) can distinguish foreign substances contained in image data using separate images of foreign substances such as shellfish shells and sand as learning data, and calculate the ratio of foreign substances contained in image data of photographed shellfish products.

[0154] Again, the classification model (CM) according to an embodiment of the present invention can be divided into a plurality of modules, namely, a backbone module (Backbone), a neck module (Neck), and a head module (Head). Each of the backbone module (Backbone), the neck module (Neck), and the head module (Head) includes one or more layers among the aforementioned multiple layers.

[0155] The backbone module performs weighted operations (multiple operations where weights are applied between multiple layers) on the input image, i.e., the original image, to generate a feature map (FM) that compresses the features of the original image.

[0156] The head module (Head) performs weight calculations on the feature map (FM) generated by the backbone module (Backbone), i.e., the feature map (FM) in which the features of the original image are compressed, to detect the box (BB) and classification probability. The head module (Head) can be divided into a dense prediction module (Dense Prediction) and a sparse prediction module (Sparse Prediction). The dense prediction module (Dense Prediction) detects the box (BB) and classification probability using the same layer, and the sparse prediction module (Sparse Prediction) detects the box (BB) and classification probability using a separate layer.

[0157] The neck module connects the backbone module and the head module.

[0158] In particular, the neck module (Neck) performs a weight operation on the feature map (FM) generated by the backbone module (Backbone), that is, the feature map (FM) in which the features of the original image are compressed, thereby refining and reconfiguring the feature map (FM), thereby restoring the compressed feature map (FM) to the size of the original image, and merges the box (BB) and classification probability derived by the head module (Head) into the restored original image and outputs it.

[0159] Next, a method for training a classification model (CM) according to an embodiment of the present invention will be described. Fig. 11 is a flowchart illustrating a method for generating a classification model (CM) according to an embodiment of the present invention.

[0160] Referring to Fig. 11, the learning module (210) prepares learning data in step S110. It includes a learning source image and a label corresponding to the learning source image. The learning source image (TI) means an image in which an object to be learned, such as a large oyster (C01), a medium oyster (C02), a small oyster (C03), etc., is captured. An example of such a screen is illustrated in Fig. 6. The label includes a ground-truth box (GT) (dx, dy, w, h) indicating an area occupied by an object in the learning source image (TI) and a hard code indicating the class of the object in the ground-truth box (GT). For example, as shown in Fig. 4, if the object of the real-world box (GT) is a large oyster (C01), the hard code can be set to a one-hot vector through one-hot encoding, such as “(C01, C02, C03) = (1, 0, 0)”. As another example, if the object of the real-world box (GT) is a small oyster (C03), the hard code can be set to “(C01, C02, C03) = (0, 0, 1)”.

[0161] Next, the learning module (210) inputs the original image (TI) for learning into the classification model (CM) having weights for which learning has not been completed in step S120. Then, the classification model (CM) performs multiple operations to which multiple inter-layer weights are applied to the original image (TI) for learning in step S130, thereby dividing the original image (TI) for learning into multiple cells, and calculates object information [(dx, dy, w, h), (C01, C02, C03)=(0.700, 0.190, 0.110)] including a bounding box (BB) (dx, dy, w, h) that predicts the area occupied by an object and a classification probability of an object within the bounding box (BB) “(C01, C02, C03)=(0.700, 0.190, 0.110)”, and derives a learning output image (TO) including the calculated object information. An example screen of such a learning output image (TO) is shown in Fig. 7. At this time, the classification model (CM) divides the learning original image (TI) into multiple cells, and then, based on the divided cells, creates a region box (BB) ( , , , ) and the conditional probability that an object within the area box (BB) is classified into the class of the learned object [(C01, C02, C03) = (0.700, 0.190, 0.110)] is calculated.

[0162] Next, the learning module (210) calculates a loss representing the difference between the object information of the learning output image (TO) and the label through a loss function in step S140. At this time, the learning module (210) can calculate the loss through the loss function of the following equation (3).

[0163] Equation (3):

[0164]

[0165] S represents the number of cells, and B represents the number of area boxes (BB) within a cell. Also, , represents the center coordinate of the area box (BB), and and represent the width and height of the area box (BB), respectively. dx, dy are the center coordinates of the real area box (GT), and w and h are the width and height of the real area box (GT). C is the confidence according to the real area box (GT). is the predicted confidence indicating the probability that an object exists in the j-th area box of the i-th cell. pi(c) is a hard code indicating the class of the object in the real area box (GT), and can be, for example, “(C01, C02, C03)=(1, 0, 0)”. represents the conditional probability that an object within a bounding box is classified into the class of the learned object. Here, c is an index representing the class. For example, it can be calculated as (C01, C02, C03) = (0.700, 0.190, 0.110). Here, i is the index representing the cell where the object exists, and j is the index representing the predicted bounding box (BB). and is a hyperparameter. is a parameter to further reflect the parameters for the area box, and is a parameter for balancing the loss for the coordinates (dx, dy, w, h) of the area box with other losses. is to weight the values ​​of the parameters of the area box where the object is located and to reflect less the values ​​of the parameters of the area where the object is not located. That is, is a parameter for the balance between the area boxes with objects and the area boxes without objects. Indicates that an object exists in cell i. represents the area box j of cell i.

[0166] The first and second terms of equation (3) are the coordinates of the area box (BB) predicted by the classification model (CM). , , , ) and the coordinate loss that represents the difference between the coordinates (dx, dy, w, h) of the ground truth box (GT) to be learned. In addition, the third term of equation (3) is to calculate the confidence loss that represents the difference between the presence or absence of an object in the ground truth box (BB) for each cell and the ground truth box (GT). Finally, the last term of equation (3) is to calculate the classification loss that represents the difference between the classification probability, which is the conditional probability that an object in the ground truth box (BB) is classified into the class of the learned object, and the hard code representing the class of the object in the ground truth box (GT).

[0167] Next, the learning module (210) performs optimization to modify the weights of the classification model (CM) so that the loss derived through the loss function in step S150 is minimized. That is, the learning module (210) performs optimization to modify the weights of the classification model (CM) so that the loss including the coordinate loss, the confidence loss, and the classification loss derived through the loss function of Equation (3) is minimized.

[0168] Steps S120 to S150 described above repeatedly update the weights of the classification model (CM) using a plurality of different training source images or a batch of a plurality of different training source images to complete training. This repetition is repeated until a predetermined condition is satisfied. For example, the condition for completing training may be that the coordinate loss, the confidence loss, and the classification loss are each below a predetermined threshold, and the total loss (coordinate loss + confidence loss + classification loss) converges. As another example, the condition for completing training may be that the learning rate is preset and the preset learning rate is reached.

[0169] Next, a method for classifying and processing aquatic products based on a classification model according to an embodiment of the present invention will be described. Figure 12 is a flowchart illustrating a method for classifying and processing aquatic products based on a classification model according to an embodiment of the present invention.

[0170] Referring to FIG. 8, the control unit (540) receives the original image captured through the image acquisition unit (710) at step S210. This original image is provided to the classification module (820) and the phage information generation module (830).

[0171] The classification module (820) inputs the original image captured in step S220 into the learned classification model (CM). That is, the classification module (820) inputs the original image into the learned classification model (CM) according to the same method as the embodiment described with reference to FIG. 5.

[0172] Then, the classification model (CM) performs multiple operations in which multiple inter-layer weights are applied to the original image in step S230, thereby deriving object information [e.g., (x, y, w, h), C01=0.888, C02=0.101, C03=0.011] including a bounding box (BB) (x, y, w, h) and a classification probability (e.g., C01=0.888, C02=0.101, C03=0.011), as illustrated in FIG. 10, and outputs an output image (TO) including the derived object information.

[0173] The classification probability represents the conditional probability that an object within a BB is classified into a class of a learned object, and the classification module (820) generates classification information by distinguishing between good and bad oysters based on the size of the oyster, which is an object classified by the classification model (CM), and the proportion of foreign substances contained in the oyster. Good oysters are oysters that contain foreign substances below the standard, and bad oysters are oysters that contain foreign substances above the standard. Here, foreign substances refer to oyster shells, sand, etc. that are introduced during the oyster shucking process (oyster shucking process) during the oyster processing. The standard can be separately set by the user according to the quality grade of the oyster.

[0174] Accordingly, the classification module (820) generates classification information including whether the oyster is a good oyster or a bad oyster according to the size of the oyster and the ratio of foreign substances in the area box according to the classification probability in step S230, and provides the generated classification information to the phage information generation module (830).

[0175] For example, the classification module (820) may determine that the object is a good-quality large oyster based on the classification probability, and may determine that the area occupied by the large oyster is (x, y, w, h) based on the area box. Accordingly, classification information including an indicator indicating a good-quality large oyster and an area box (x, y, w, h) indicating the area occupied by the large oyster may be provided to the phage information generation module (830).

[0176] Meanwhile, when receiving the original image, the phage information generation module (830) corrects the distortion of the original image caused by the image acquisition unit (710) in step S310. Step S310 is described in more detail as follows.

[0177] In order to obtain position information, the grip information generation module (830) must convert the object's position information, which is composed of pixel coordinates, into actual physical coordinates. Accordingly, the grip information generation module (830) corrects radial distortion caused by the convex lens of the industrial image acquisition unit (710). The grip information generation module (830) can obtain internal camera parameters for distortion correction and then convert the distorted original image into a corrected image.

[0178] As above, the shellfish washing unit that has completed the radioactivity test and quality test through the quality tester (400) can be moved to the classifier (300).

[0179] FIG. 13 is a drawing for explaining a classifier in an automatic shellfish processing device according to one embodiment of the present invention, and FIGS. 14 and 15 are drawings for explaining a robot section, a classification conveyor section, and a sorting conveyor section in an automatic shellfish processing device according to one embodiment of the present invention.

[0180] Shellfish products that have passed through the fluid washing machine (200) can be classified by size by the classifier (300) after all washing has been completed.

[0181] The classifier (300) is connected to the fluid cleaner (200) and can compare the size of the moved shellfish with preset data to determine the size of the shellfish currently being photographed and move it.

[0182] Shellfish products are moved to a conveyor-type sorting conveyor unit (340), and the robot unit (320) can pick up and move the shellfish products.

[0183] The robot part (320) is formed so that a plurality of arm members (322) are combined with a head that rotates around a rotation axis, and a vision camera (not shown) is provided at an adjacent location to detect the size of shellfish seafood.

[0184] The robot unit (320) has an electronic control device such as a signal processing unit and a communication unit installed inside, so that the size of shellfish aquatic products photographed through the vision camera can be compared with previously stored data and classified.

[0185] In one embodiment of the present invention, the stored data may classify shellfish into three sizes, and the currently photographed shellfish may be identified by matching the size category with information about the size of the currently photographed shellfish.

[0186] The robot part (320) includes a plurality of arm members (322), and a gripping member (not shown) is provided at the end of the arm member (322) to grip shellfish seafood.

[0187] The above-mentioned phage member can move the phage-processed shellfish products to the sorting conveyor unit (340).

[0188] The classification conveyor unit (340) is adjacent to the robot unit (320) and is formed in a long shape in the form of a conveyor belt, and can be a path for moving shellfish marine products that are moved and settled through the robot unit (320).

[0189] The sorting conveyor unit (360) is provided in the form of a conveyor belt at the lower end of the sorting conveyor unit (340) and can transport shellfish products by individually accommodating them in an array tray (40).

[0190] As illustrated in FIG. 7, the classification conveyor unit may be provided in multiple units, and may be provided with a first classification member (342), a second classification member (344), and a third classification member (346) according to one embodiment of the present invention.

[0191] The robot section can move shellfish of a first size to a first sorting member (342), shellfish of a second size to a second sorting member (344), and shellfish of a third size to a third sorting member (346) according to the size of the shellfish.

[0192] The first size, the second size, and the third size are intended to represent different sizes, such that the first size may be a relatively smallest size, and the third size may be a relatively largest size.

[0193] And the sorting conveyor unit (360) may have a first sorting member (362) positioned at the lower end of the first sorting member (342), a second sorting member (364) positioned at the lower end of the second sorting member (344), and a third sorting member (366) positioned at the lower end of the third sorting member (346).

[0194] The sorting conveyor unit (340) can be slidably connected to a sorting tray (20) having a plurality of grooves formed at the lower end to accommodate shellfish products moved thereto.

[0195] In addition, the lower surface of the classification tray (20) is formed to be openable so that shellfish products received in the groove can fall to the array tray (40) when the lower surface is opened.

[0196] And, an array tray (40) is placed under the sorting tray (20) so that shellfish products falling from the sorting tray (20) can be individually contained one by one.

[0197] The array tray (40) is placed below the classification tray (20), and a groove corresponding to the shape of the shellfish seafood may be formed.

[0198] The array tray (40) is placed below the classification tray (20) and is characterized by having a number of grooves arranged in accordance with the shape of shellfish seafood.

[0199] Additionally, the array tray (40) may be formed with grooves of different sizes corresponding to the sizes of shellfish products moved to the first alignment member (362), the second alignment member (364), and the third alignment member (366).

[0200] Through this, shellfish products arranged on the sorting tray are dropped into the arrangement tray (40), and the arrangement tray (40) is moved to the end of the sorting conveyor unit (360) and can be stored frozen.

[0201] The robot part (320) according to one embodiment of the present invention can be transformed into two forms.

[0202] Let us check the modified embodiment of the robot part (320) through the drawing.

[0203] FIG. 16 and FIG. 17 are drawings for explaining a robot part in an automatic shellfish processing device according to a first modified embodiment of the present invention.

[0204] The robot unit (1300) may include a second transport line (L2), which is a line through which defective shellfish products are discharged from among the shellfish products that have passed the quality inspection machine (400), and a third transport line (L3), which is a line through which shellfish products selected as having excellent quality are transported.

[0205] In the description of the present invention, shellfish aquatic products are replaced with 'oysters'.

[0206] The robot unit (1300) is configured to sort clean oysters into a plurality of preset sizes and discharge them separately. The clean oysters separated and discharged by the robot unit (1300) can be moved to the third transport line (L3). Here, the plurality of preset sizes can be, for example, three sizes: S (Small), M (Midium), and L (Large). In this case, the clean oysters classified by size can be classified into the respective sizes of S, M, and L and discharged separately.

[0207] The robot unit (1300) may be equipped with a vision camera (not shown) (here, the vision camera is a separate configuration from the vision camera included in the image acquisition unit (710)) that recognizes and sorts clean oysters by a plurality of preset sizes.

[0208] The robot unit (1300) may be configured with a delta robot structure. The robot unit (1300) configured with a delta robot structure may be equipped with a picker (P) that picks up and separates oysters by size based on the results selected by a vision camera (not shown).

[0209] In addition, the robot part (1300) of the delta robot structure may include a base (1310), an arm (1320), a link arm (1330), and an effector (1340). A plurality of arms (1320) are radially connected to the base (1310). The plurality of arms (1320) may be provided so that their rotation angles can be adjusted based on the base (1310). Each of the plurality of arms (1320) has an individual driving motor on the base (1310), so that the rotation angle adjustments can be individually controlled. One end of the link arm (1330) is connected to each arm (1320) in pairs. At this time, one end of the link arm (1330) may be connected to the arm (1320) by a universal joint (1350). An effector (1340) is connected to the other end of each pair of link arms (1330). At this time, the other end of each pair of link arms (330) and the effect (1340) may be connected by a universal joint (1350). In addition, a picker (P) may be placed on the lower side of the effect (1340). The picker (P) may be configured to pick up oysters by digging them up or by blackening them. When the robot unit (1300) is equipped with the delta robot structure as described above, oysters can be picked up at a high speed in various directions.

[0210] A second modified embodiment of the robot part (320) will be described with reference to FIGS. 18 to 20.

[0211] FIG. 18 is a drawing for explaining a robot part in an automatic shellfish processing device according to a second modified embodiment of the present invention, FIG. 19 is a drawing for explaining a support part and a soft finger of a robot part in an automatic shellfish processing device according to a second modified embodiment of the present invention, and FIG. 20 is a drawing for explaining a nail part of a robot part in an automatic shellfish processing device according to a second modified embodiment of the present invention.

[0212] As illustrated in FIG. 18, a robot part (2300) according to one embodiment of the present invention includes a body (2310), a support part (2320), a soft finger (2330), and a nail part (2340). The body (2310) is formed as a circular plate having a predetermined thickness. The body (2310) supports the entire robot part (2300), and serves to be fixed to a shaft or the like when the robot part (2300) is mounted on a robot arm.

[0213] A body (2310) is equipped with a plurality of air intake pipes (2311). An opening is formed on the body (2310), and the air intake pipes (2311) are connected through the opening. The body (2310) may be formed of a rigid material such as metal or reinforced plastic.

[0214] The support (2320) is connected to the lower surface of the body (2310). The support (2320) extends downward from the perimeter of the lower surface of the body (2310). The support (2320) may have a cylindrical shape with a space formed therein.

[0215] Meanwhile, the support member (2320) may have an air inlet pipe (2321) and a second inclined surface (2322). The air inlet pipe (2321) and the second inclined surface (2322) are connected to each other and have different inclinations. The air inlet pipe (2321) may extend vertically from the periphery of the lower surface of the body (2310), or may have a diameter that slightly increases downward from an end fixed to the lower surface of the body (2310). The second inclined surface (2322) may have a diameter that decreases downward from an end connected to the air inlet pipe (2321). That is, the support member (2320) may have a cylindrical shape with a convex center.

[0216] The inclination angle of the air inlet pipe (2321) may be 0 to 5° with respect to the vertical direction. The inclination angle of the second inclined surface (2322) may be 10 to 30° with respect to the vertical direction.

[0217] The support member (2320) is formed of a material with high rigidity. The support member (2320) may be formed of metal or reinforced plastic. The robot member (2300) can secure horizontal rigidity by the support member (2320).

[0218] The soft finger (2330) is the part that grips objects. The soft finger (2330) uses a soft material to grip soft objects such as oysters, and instead of directly applying pressure to the soft object, it indirectly wraps and lifts it.

[0219] As illustrated in Fig. 19, the soft finger (2330) is connected to the lower part of the body so that the upper part is positioned within the support part (2320). The lower part of the soft finger (2330) is exposed outside the support part (2320). The soft finger (2330) is shaped like a pocket, and an opening is formed at the upper part through which air is sucked in or discharged from the air inlet tube (2311).

[0220] The soft finger (2330) is the part that comes into direct contact with objects. Hygiene management is crucial, and it is preferable to use food-grade, non-toxic materials. Furthermore, to grip an object, the soft finger (2330) is deformed in the desired direction and shape to generate the desired gripping motion. To achieve this, the present invention utilizes the characteristics of flexible materials such as silicone and the differences in the shapes of each part.

[0221] The soft finger (2330) has an elastic portion (2331) and an inflatable portion (2332). The soft finger (2330) is formed of a flexible, elastic, and stretchable material, such as silicone or rubber. In the present embodiment, the soft finger (2330) is formed of a silicone material, and the elastic portion (2331) and the inflatable portion (2332) are formed integrally, and are in the form of a flat pocket with an empty interior.

[0222] An opening (2333) is formed at the top of the soft finger (2330). An air inlet pipe (2311) is connected to the opening (2333), and air is introduced into the soft finger (2330) through the opening (2333).

[0223] The elastic portion (2331) is located on one side of the soft finger (2330) and has a constant thickness. The elastic portion (2331) may face the center of the support portion (2320). The elastic portion (2331) is a portion that directly contacts the object when the soft object is gripped, and has a rigidity that does not deform its shape due to the weight of the object. The elastic portion (2331) is formed with a predetermined thickness to secure rigidity, and is formed with the same thickness throughout the elastic portion. The elastic portion (2331) is formed of a flexible and human-safe material, so that the object can be gripped safely without being pressed or damaged by strong pressure.

[0224] The expansion portion (2332) is located on the other side of the soft finger (2330) and has a thinner thickness than the elastic portion (2331). The expansion portion (2332) can contact the inner surface of the support portion (2320). A folding surface is formed on the expansion portion (2332). The folding surface is a surface formed by folding a portion of the expansion portion (2332), and is a portion that is mainly expanded when air is injected into the soft finger (2330).

[0225] When air is introduced into the soft finger (2330) through the air inlet tube (2311), the expansion portion (2332) of the soft finger (2330) expands. The soft finger (2330) may have a shape that is bent inward toward the support portion (2320) to facilitate gripping of an object.

[0226] As illustrated in FIG. 20, the nail portion (2340) is connected to the lower end of the soft finger (2330). The nail portion (2340) may be in the shape of a thin plate. The soft finger (2330) expands and moves toward the center of the support portion (2320). If an object is placed flat on the floor, it may not be gripped well or may be difficult to grip in its entirety. The nail portion (2340) is formed to be thin and sharp, so that it can easily lift objects made of soft materials. The soft finger (2330) and the nail portion (2340) may be formed integrally from the same material, and the nail portion (2340) may be formed in a flat shape without an internal space like the soft finger (2330). When gripping an object, the nail portion (2340) can concentrate force on the bottom of the object to lift the object placed on the conveyor as if scraping it from the floor. In addition, it can act as a support after gripping an object to prevent the gripped object from slipping off the soft finger (2330).

[0227] A slope (2341) may be formed at the end of the nail portion (2340) to enable good scratching of soft objects.

[0228] As described above, preferred embodiments of the present invention have been described. It will be apparent to those skilled in the art that the present invention can be embodied in other specific forms, in addition to the embodiments described above, without departing from the spirit or scope thereof. Therefore, the above-described embodiments should be considered illustrative rather than restrictive, and accordingly, the present invention is not limited to the above description, but may be modified within the scope of the appended claims and their equivalents.

Claims

1. An automatic shellfish processing device that automatically performs the processes of washing, inspecting, and packaging shellfish products. A physical washing machine that becomes a path along which the shellfish aquatic product moves and protrudes toward the shellfish aquatic product along the path to wash the shellfish aquatic product; A fluid washer adjacent to the physical washer and connected to the path, which sprays fluid toward the shellfish aquatic product to remove fine foreign substances attached to the shellfish aquatic product; A quality inspection device extending from the path adjacent to the fluid washing machine and determining whether the shellfish seafood is good or bad and performing a radioactivity test; and An automatic shellfish processing device, which is connected to the path of the quality inspection device and includes a classifier that picks up and separates shellfish products according to a preset size.

2. In the first paragraph, the physical washing machine, An automatic shellfish processing device comprising: a zigzag washing section in which a groove is formed to discharge foreign substances falling off the shellfish and in which the shellfish moves together with the fluid generated at the end, and in which a section in which at least a portion is inclined downward and bent is formed; and a cylindrical washing section having a cylindrical shape and being rotatable about a central axis, and including a panel formed to protrude upward so as to rotate the shellfish moved in the zigzag washing section.

3. In the second paragraph, the physical washing machine, An automatic shellfish processing device comprising: a dewatering transport section having an upward slope from the end of the zigzag washing section toward the cylindrical washing section, for transporting the shellfish from the zigzag washing section to the cylindrical washing section; and a unit transport section formed long and transporting the shellfish from the cylindrical washing section to the fluid washing machine.

4. In the first paragraph, the fluid washer, An automatic shellfish processing device comprising: a vortex washing unit formed on the inner side of a semicircular receiving portion and having a vortex nozzle for discharging fluid toward the shellfish; and a bubble washing unit connected to the vortex washing unit and having a washing nozzle for discharging fluid from the upper portion of the shellfish moved in the vortex washing unit.

5. In the fourth paragraph, the bubble washing unit, An automatic shellfish processing device comprising: an air generating member formed long along the above path, having a downward slope toward the center, and providing air toward the shellfish; and a washing water generating member having an upward slope at the end of the air generating member, causing the shellfish to move, and having a plurality of washing nozzles arranged at a predetermined interval at the upper portion.

6. In the first paragraph, the classifier, An automatic shellfish processing device comprising: a robot unit connected to the fluid cleaner, determining the size of the moved shellfish by comparing it with preset data, and gripping and moving the shellfish; a sorting conveyor unit formed in a long conveyor belt shape adjacent to the robot unit, and moving the shellfish settled through the robot unit to an end; and a sorting conveyor unit provided in the shape of a conveyor belt at the lower end of the sorting conveyor unit, and transporting the shellfish by accommodating it in an array tray that individually accommodates the shellfish.

7. In paragraph 6, the quality inspector, An automatic shellfish processing device, comprising: an image acquisition unit for acquiring image data on the shellfish aquatic product transported from the fluid washing machine; and an image processing unit for storing the image data on the shellfish aquatic product, inputting the image data into an automatic classification model, and generating classification information on good / defective products, and generating classification information capable of catching the shellfish aquatic product through image processing on the image data, and then mapping the classification information and the classification information.

8. In paragraph 7, the robot part, An automatic shellfish processing device characterized in that the shellfish aquatic product is crushed using the mapped classification information and crushing information received from the image processing unit, and is classified into a classification conveyor unit allocated according to the classification information.

9. In paragraph 6, the classification conveyor unit, An automatic shellfish processing device characterized in that a plurality of sorting trays having grooves formed at the lower end for receiving shellfish products moved to the lower end are slidably connected.

10. In the 9th paragraph, the classification tray, An automatic shellfish processing device characterized in that the shellfish is formed to penetrate along the direction in which the shellfish is inserted, and a door that can be opened and closed at the bottom is slidably coupled, so that when the door is opened, the shellfish located on the array tray falls onto the array tray.

11. In paragraph 6, the robot part, An automatic shellfish processing device characterized in that a plurality of arm members are formed to be combined with a head that rotates around a rotation axis, and a vision camera is provided adjacent to the robot section to detect the position and size of the shellfish marine product in order to grip the shellfish marine product.

Images