Image acquisition device for quality evaluation of pig carcass, and grade determination method using same

Abstract

The present invention provides a device and a grade determination method using same, the device being capable of acquiring and analyzing multi-view images of various parts of a split carcass during hanging, on a trolley that travels along a rail installed on the ceiling of a slaughterhouse, and transferring the split carcass divided into left and right parts by using an A-type rotary gambrel (hook-type ring for hanging a pig carcass).

Description

Image acquisition device for quality evaluation of pig carcasses and grading method using the same

[0001] The present invention relates to an image acquisition device for quality evaluation of pig carcasses and a grading method using the same. More specifically, the invention relates to an image acquisition device for quality evaluation of pig carcasses and a grading method using the same, which collects accurate still image data of various parts of a pig carcass, synthesizes it into a single image, analyzes it based on artificial intelligence, and improves the accuracy and speed of grading based on the same.

[0002] Today, the implementation of a livestock grading system that differentiates the quality of domestic cattle and pig carcasses (slaughtered, butchered, and divided in half) by assessing them according to objective standards is contributing to the improvement of livestock product quality and income stabilization, while also creating a transparent distribution structure by enhancing the linkage between grades in the distribution process.

[0003] For this livestock product grading system, the accuracy of carcass grading is crucial to produce and distribute high-quality livestock products through price differentiation based on grade, and to enable consumers to identify the quality of the products.

[0004] Meanwhile, as the speed of slaughtering has increased due to the modernization and scaling up of slaughterhouse facilities, more than 600 pigs are being slaughtered per hour, and accordingly, non-destructive image analysis methods are being used to increase the accuracy of pig carcass grading results.

[0005] Currently, the grading system for pig carcasses is based on standard pigs (Grade 1+; carcass weight 83–92 kg, back fat thickness 17–24 mm), and grades are determined based on the difference in carcass weight and back fat thickness compared to the standard pig criteria.

[0006] In other words, the grades of pig carcasses in Korea are determined as 1+, 1, 2, and ungraded by applying the results of the first and second grades.

[0007] The first grade determination is made by applying the total weight of the carcass measured and provided by the slaughterhouse manager and the average values ​​of the back fat thickness between the last thoracic vertebra and the first lumbar vertebra of the left carcass (a carcass split in half from the neck to the tail) and the back fat thickness between the first thoracic vertebra and the twelfth thoracic vertebra. Carcasses with a weight of 83–92 kg (based on scalded carcasses) and a back fat thickness of 17–24 mm are graded as Grade 1+, and grades are determined as Grade 1 or Grade 2 depending on the extent to which the weight and back fat of Grade 1+ are exceeded.

[0008] The second grade is determined by comprehensively evaluating the carcass's appearance (fat, pork belly, fat attachment status), meat quality (fat deposition, meat color, meat texture, fat color and fat content), and defects, and the lower of the first and second grade results is determined as the final carcass grade.

[0009] Generally, the final grade of a pig carcass is determined by carcass quality and meat quality using 21 parameters including backfat thickness (BFT), hot carcass weight (CWT), sex, appearance, meat quality, and defects.

[0010] Meanwhile, the VCS2000 (automatic pig carcass judgment machine and system) from E+V Technology GmbH of Germany, which has been introduced to large-scale pig slaughterhouses in Korea, uses images of pig carcasses captured by one black-and-white camera and two color cameras to estimate and analyze major cut sections, total meat volume, and meat yield.

[0011] That is, one black-and-white camera photographs the back of a pig in a half-carcass state (a form in which the head, internal organs, tail, etc. are removed and the pig is cut in half after slaughter) against a white background, one of the two color cameras photographs the top of the suspended left carcass, and the other color camera photographs the bottom of the left carcass, and then the photos of the top and bottom are combined on a computer monitor to appear as a single left carcass, and the total meat yield and meat ratio are automatically predicted through the analyzed data.

[0012] However, due to the nature of combining the upper and lower photos of the left conductor into one, the part where the two photos are connected is unclear, and there is a problem that errors in grade judgment occur due to photo errors.

[0013] In addition, since only the left carcass is photographed and the right carcass is not, only a portion of the pig carcass is evaluated, and there is a problem in that it is difficult to accurately determine the sex of the pig.

[0014] It is stated that the background technology or prior art described herein refers to information possessed by the inventor or acquired during the process of deriving and completing the present invention, and is specified merely to aid in understanding the technical significance of the present invention and to be useful for prior art search and examination, and does not mean technology that was generally known and widely used in the technical field to which the invention belongs prior to the filing of the present invention.

[0015] Accordingly, the inventors of the present invention have devised the present invention as a result of continuous research and strenuous efforts to develop a new image acquisition device for evaluating the quality of pig carcasses and a grading method using the same. This is achieved by comprehensively considering the aforementioned matters and, with the idea of ​​resolving the technical limitations and problems of existing automatic pig carcass grading machines and systems, collecting accurate carcass image data for various parts of the pig carcass obtained by photographing the left and right carcasses in a half-carcass state (a form in which the head, internal organs, tail, etc. are removed after slaughtering and the carcass is cut in half), synthesizing the data into a single image, and analyzing it based on artificial intelligence, thereby enhancing the accuracy and speed of grading.

[0016] Therefore, the technical problem and objective of the present invention is to provide an image acquisition device for quality evaluation of pig carcasses and a grading method using the same, which can increase the accuracy and speed of grading.

[0017] The technical problems and objectives that the present invention aims to solve are not limited to those mentioned above, and other unmentioned technical problems and objectives will be clearly understood by those skilled in the art from the description below.

[0018] A specific means according to an aspect of the present invention for effectively achieving a specific technical purpose while embodying a new concept for solving the technical problem of the present invention as described above presents a device for acquiring multi-faceted image of a divided carcass during the process of transporting a divided carcass by suspending it from left and right using an A-type rotary gamble (a hook-type hook for hanging pig carcasses) on a trolley traveling along a rail installed on the ceiling of a slaughterhouse.

[0019] Specifically, a chroma key screen fixedly installed at a certain distance below both the left and right sides of the rail; a first vision camera installed below the side of the rail and transmitting a conductor image obtained by photographing the belly portion of the bipart conductor located within the chroma key screen; a turn guide installed below the rail and guiding the bipart conductor, while suspended from the gamble, to rotate 90 degrees together with the gamble as it moves along the rail by the movement of the trolley; a spread guide installed below the rail and guiding the lower part of the bipart conductor, rotated 90 degrees by the turn guide, to spread out in both the left and right directions of the rail; a second vision camera installed below the right side of the rail and transmitting a conductor image obtained by photographing the conductor number of the right conductor among the bipart conductors located within the chroma key screen; a third vision camera installed below the right side of the rail and transmitting a conductor image obtained by photographing the side surface of the right conductor among the bipart conductors located within the chroma key screen; and the device installed below the rail and suspended from the gamble The present invention provides an image acquisition device for quality evaluation of a pig carcass, characterized by comprising: a pull rod that pulls the lower part of the right carcass of the bipartite moving along the rail by the movement of a trolley in the direction opposite to the direction of movement; and fourth to sixth vision cameras installed on the lower right side of the rail, which respectively transmit carcass images obtained by dividing and photographing the cross-section of the left carcass, which is revealed as the right carcass of the bipartite located within the chroma key screen is pulled in the direction opposite to the direction of movement by the pull rod, into the upper, central, and lower parts.

[0020] Thus, the present invention can collect and synthesize into one multi-faceted carcass image data for the entire pig carcass obtained by photographing the left and right carcasses in the state of half-carcass (a form in which the head, internal organs, tail, etc. are removed and the carcass is cut in half after slaughtering the pig).

[0021] In addition, a preferred aspect of the present invention is configured to further include a controller that performs overall control functions of the device, wherein the controller controls the first vision camera by means of a signal from a first sensor that detects the position of the bipartite conductor moving along the rail by the movement of the trolley while suspended from the gamble and converts it into an electrical signal and outputs it, controls the second vision camera by means of a signal from a second sensor that detects the position of the bipartite conductor moving through the first sensor and converts it into an electrical signal and outputs it, controls the pull load and the third vision camera by means of a signal from a third sensor that detects the position of the right conductor among the bipartite conductors moving through the second sensor and converts it into an electrical signal and outputs it, and controls the fourth to sixth vision cameras respectively by means of a signal from a fourth sensor that detects the position of the left conductor among the bipartite conductors moving through the third sensor and converts it into an electrical signal and outputs it.

[0022] In addition, in a preferred aspect of the present invention, the pull rod is configured to include a rotary actuator controlled by the controller and a lever arm that rotates at a constant angle by the rotary actuator to temporarily pull and hold the lower part of the right conductor of the bipart conductor in the direction opposite to the direction of movement, thereby enabling more stable imaging of the bipart conductor to obtain an accurate conductor image.

[0023] In addition, a preferred aspect of the present invention comprises a lighting module that brightly illuminates the bipart conductor, a guide roller installed at the tip of the spread guide that guides the bipart conductor, which moves along the rail while suspended from the gamble and driven by the trolley, to spread smoothly without shaking, and guard rails installed on both the left and right sides of the spread guide that guide the bipart conductor, which moves along the rail while suspended from the gamble and driven by the trolley, to prevent it from deviating outward, thereby enabling the bipart conductor to be photographed more stably and an accurate conductor image to be obtained.

[0024] In addition, a preferred aspect of the present invention is configured such that the fourth to sixth vision cameras are each mounted on a vertical bar so as to be movable up and down, and further comprises a setting board having a reference point for determining the angle of view adjustment and the vertical position of the fourth to sixth vision cameras in order to set the overlapping section between the image captured by the fourth vision camera and the fifth vision camera, which divides the cross-section of the left conductor of the bipartite conductor into upper, central, and lower parts for shooting, and the overlapping section between the image captured by the fifth vision camera and the sixth vision camera, thereby maintaining consistency in the overlapping section.

[0025] In addition, a preferred aspect of the present invention is configured to further include a reader that obtains unique identification information by tagging an RFID tag attached to the gamble on which the biparted carcass is suspended, and then sends it to a computing device, wherein the computing device receives carcass images of the biparted carcass, synthesizes them to extract feature points, analyzes them through an artificial intelligence model to determine the sex and whether the biparted carcass is a sow, measures back fat thickness, body length, and body width, verifies the traceability number, and determines the grade based on this.

[0026] In other words, collected conductor image data can be automatically measured and analyzed for quality evaluation items through an artificial intelligence model, and based on this, the accuracy and speed of grading can be improved.

[0027] In addition, as a preferred aspect of the present invention, the computing device comprises a gender determination area tracking unit that tracks a gender determination area from a conductor image of the bipartite conductor, a cropping unit that crops an area corresponding to the tracked gender determination area, a gender learning unit that learns a gender reading model through a conductor image of the cropped area, a vertebral segment tracking unit that tracks a vertebral segment from a conductor image of the bipartite conductor, a labeling unit that individually labels the tracked vertebral segments, a storage unit that stores the coordinates of the labeled vertebral segments in a database, and a conductor width calculation unit that calculates coordinates corresponding to the midpoint between the lumbar and thoracic vertebrae using the stored vertebral segment coordinates and calculates the conductor width by position using the calculated coordinates, thereby increasing the accuracy and speed of grade determination.

[0028] And specific means according to another embodiment of the present invention include: (a) detecting the position of a bifurcated carcass transported by being suspended from a gambler using a trolley running along a rail installed on the ceiling of a pig carcass slaughterhouse; (b) photographing the belly portion of the bifurcated carcass located within a chroma key screen fixedly installed at a certain distance below both the left and right sides of the rail, and transmitting the carcass image to a computing device; (c) rotating the bifurcated carcass, which moves along the rail by the movement of the trolley while suspended from the gambler, 90 degrees together with the gambler; (d) photographing a carcass number marked on the hind leg portion of the bifurcated carcass rotated 90 degrees, and transmitting the carcass image to the computing device; (e) spreading the lower portion of the bifurcated carcass rotated 90 degrees in both the left and right directions of the rail; (f) photographing the side surface of the right carcass among the bifurcated carcasses located within the chroma key screen, and transmitting the carcass image to the computing device; and (g) moving the lower portion of the right carcass among the bifurcated carcasses. A method for determining a grade using an image acquisition device for quality evaluation of a pig carcass is presented, characterized by comprising the steps of: (a) pulling in a direction opposite to the direction of movement; (b) photographing the cross-section of the left carcass revealed as the right carcass among the biparted carcasses located within the chroma key screen is pulled in a direction opposite to the direction of movement, dividing the cross-section into upper, central, and lower sections, and transmitting the carcass images to the computing device; and (c) receiving the carcass images of the biparted carcass from the computing device, synthesizing them to extract feature points, analyzing them through an artificial intelligence model to determine the sex and whether the biparted carcass is a sow, measuring backfat thickness, body length, and body width, verifying the carcass number, and determining the grade based on the above.

[0029] Thus, the present invention can automatically measure and analyze quality evaluation items using an artificial intelligence model based on collected conductor image data, and improve the accuracy and speed of grade determination based on this.

[0030] In addition, in a preferred aspect of the present invention, the computing device of step (a) receives information regarding the carcass weight of the biparted carcass from the outside and can determine whether it is a sow based on the carcass weight, body length, and body width, along with a carcass image obtained by photographing the belly portion of the biparted carcass.

[0031] In addition, as a preferred aspect of the present invention, the artificial intelligence model can detect sex characteristic points based on carcass images of a standard pig (Grade 1+; carcass weight 83~92kg, back fat thickness 17~24mm) and generate a sex determination artificial intelligence model through deep learning training by classifying items for sex determination (penile ring, ischiocavernosus muscle, semimembranosus muscle, penile eyelid muscle).

[0032] According to an embodiment that implements the technical concept on which a unique solution means is based to solve the technical problem of the present invention, multi-faceted carcass image data for various parts of a pig carcass obtained by photographing the left and right carcasses in a half-carcass state (a form in which the head, internal organs, tail, etc. are removed after slaughtering a pig and the carcass is cut in half) can be collected and synthesized into one.

[0033] In addition, collected conductor image data can be automatically measured and analyzed using an artificial intelligence model based on predetermined quality evaluation criteria, thereby increasing the accuracy and speed of grading.

[0034] Therefore, the reliability of quality assessment and grading based on sex, backfat thickness, and carcass weight through cross-sectional image analysis of pig carcasses can be enhanced.

[0035] The effects of the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the description in the claims.

[0036] FIG. 1 is a schematic plan view of an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention.

[0037] FIGS. 2 to 5 are front view diagrams schematically showing an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention.

[0038] FIG. 6 is a perspective view showing a local part of the main elements constituting an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention.

[0039] FIG. 7 is a front view showing a setting board, which is one of the main elements constituting an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention.

[0040] FIG. 8 is a block diagram schematically showing a computing device among the main elements constituting an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention.

[0041] FIGS. 9 to 16 are photographs to help understand the process and explanation of analyzing a carcass image taken of a pig carcass using an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention.

[0042] Figure 17 is a photograph illustrating a computing device in a slaughterhouse office.

[0043] FIGS. 18 to 24 are photographs illustrating an artificial intelligence-based carcass grading system using an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention.

[0044] FIG. 25 is a vision software screen of a computing device constituting an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention.

[0045] Hereinafter, embodiments according to the present invention will be described in more detail with reference to the attached drawings.

[0046] Prior to this, it is specified that the terms described below are defined in consideration of their functions in the present invention, and should be interpreted in accordance with the concept consistent with the technical spirit of the present invention and the meaning commonly accepted or recognized in the relevant technical field.

[0047] In addition, if it is determined that a detailed description of known functions or configurations related to the present invention could obscure the essence of the present invention, such detailed description is omitted.

[0048] It is stated that the attached drawings may be partially exaggerated or simplified for the purpose of explaining the configuration, operation, and operating principles of the technology, as well as for ease of understanding and clarity of the technology, and that each component in the drawings does not exactly correspond to the actual size and shape.

[0049] In addition, the term "and / or" in this specification means a combination of multiple related described items or includes any of the multiple related described items, and when a part is said to include a certain component, it means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0050] In other words, terms such as "comprising" and "having" as set forth in this specification mean that there is a feature, number, step, process, operation, component, part, or combination thereof, and should be understood as not excluding the existence or addition of one or more other features, numbers, steps, processes, operations, components, parts, or combinations thereof.

[0051] Furthermore, each process and step may occur differently from the specified order unless the context clearly indicates a specific sequence. That is, each process and step may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

[0052] Meanwhile, the terms "part" and "unit" used in the present invention refer to a unit or module form that performs a role in processing at least one intended function or a certain operation in a device or system, and this can be implemented through means such as hardware, software, or a combination of hardware and software, or a device or assembly capable of performing independent operations.

[0053] Furthermore, the term "module" as used in the present invention may refer to a unit comprising one or more combinations of hardware, software, or firmware, and may be interchangeably used with terms such as unit, logic, logical block, component, or circuit; it may be the smallest unit or part thereof of a component formed integrally, or the smallest unit or part thereof that performs one or more functions, and may be implemented mechanically or electronically.

[0054] Furthermore, terms such as top, bottom, upper surface, lower surface, or upper, lower, upper side, lower side, front / rear, left / right, etc. used in the present invention are used for convenience to distinguish relative positions or explain directions of movement for each component. For example, the upper part of a drawing may be named or referred to as the upper part and the lower part as the lower part, and the length direction may be named or referred to as the front / rear direction and the width direction as the left / right direction.

[0055] In addition, terms such as "first," "second," etc. used in the present invention may be used to describe various components. That is, terms such as "first," "second," etc. may be used solely for the purpose of distinguishing one component from another.

[0056] [Best mode for carrying out the invention]

[0057] The image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention is a device for acquiring multi-faceted images of a divided carcass (C) by suspending it from a trolley (T) that travels along a rail (R) installed on the ceiling of a slaughterhouse using an A-type rotary gamble (H) (a hook-type hook for hanging pig carcasses) and transporting it, as shown in FIGS. 1 to 7. The main elements constituting the device include a chroma key screen (10), first to sixth vision cameras (11)(12)(13)(14)(15)(16), a turn guide (20), first to fourth sensors (21)(22)(23)(24), a spread guide (30), a pull load (40), a controller (60), a reader (70), a setting board (80), and a computing device (90).

[0058] The chroma key screen (10) is fixedly installed at a certain distance below both the left and right sides of the rail (R).

[0059] That is, the chroma key screen (10) is suspended below both the left and right sides of the rail (R) with a blue or green background to facilitate the synthesis and editing of multiple conductor images obtained by photographing the bipartite conductor (C) with the first to sixth vision cameras (11)(12)(13)(14)(15)(16) and to process fine outlines.

[0060] In addition, the second to sixth vision cameras (12)(13)(14)(15)(16) are installed in front of the chroma key screen (10) positioned on the lower left side of the rail (R), and the first vision camera (11) is installed in front of the chroma key screen (10) positioned on the lower right side of the rail (R).

[0061] Additionally, when the bipartite conductor (C) is photographed by the first to sixth vision cameras (11)(12)(13)(14)(15)(16), a plurality of lighting modules (19) may be installed to brightly illuminate the bipartite conductor (C) and the chroma key screen (10) to illuminate the dark parts.

[0062] Here, an LED tube light that shines in the front direction can be used as the lighting module (19).

[0063] The first to sixth vision cameras (11)(12)(13)(14)(15)(16) are installed at a certain position below the side of the rail (R) to face the chroma key screen (10).

[0064] That is, when the first vision camera (11) detects a bipartite conductor (C) located inside the chroma key screen (10) and converts it into an electrical signal and outputs it to the controller (60), the first sensor (21) of the first vision camera (11) transmits a color conductor image obtained by photographing the belly portion of the bipartite conductor (C) under the control of the controller (60) to the computing device (100).

[0065] When the second vision camera (12) detects a bipartite conductor (C) located within the chroma key screen (10) and converts it into an electrical signal to output to the controller (60), the second sensor (22) captures the conductor number of the right conductor among the bipartite conductors (C) under the control of the controller (60) and transmits the obtained color conductor image to the computing device (100).

[0066] When the third vision camera (13) detects a bipartite conductor (C) located within the chroma key screen (10) and converts it into an electrical signal and outputs it to the controller (60), the third sensor (23) transmits a color conductor image obtained by photographing the side of the right conductor of the bipartite conductor (C) under the control of the controller (60) to the computing device (100).

[0067] When the fourth vision camera (14) detects a bipartite conductor (C) located within the chroma key screen (10) and converts it into an electrical signal to output to the controller (60), the fourth sensor (24) transmits a color conductor image obtained by photographing the upper cross-section of the left conductor of the bipartite conductor (C) under the control of the controller (60) to the computing device (100).

[0068] When the fourth sensor (24) detects a bipartite conductor (C) located within the chroma key screen (10) and converts it into an electrical signal to output to the controller (60), the fifth vision camera (15) transmits a color conductor image obtained by photographing the center of the cross-section of the left conductor of the bipartite conductor (C) under the control of the controller (60) to the computing device (100).

[0069] When the fourth sensor (24) detects a bipartite conductor (C) located within the chroma key screen (10) and converts it into an electrical signal to output to the controller (60), the sixth vision camera (16) transmits a color conductor image obtained by photographing the lower cross-section of the left conductor of the bipartite conductor (C) under the control of the controller (60) to the computing device (100).

[0070] Among these, the 4th to 6th vision cameras (14)(15)(16) are each mounted on a vertical bar (B) so as to be movable up and down.

[0071] That is, the 4th to 6th vision cameras (14)(15)(16) can adjust their mounting position relative to the vertical bar (B).

[0072] Here, it is preferable to install the first to sixth vision cameras (11)(12)(13)(14)(15)(16) together with the main camera and backup camera in a single protective housing to prepare for defects or failures.

[0073] In addition, an air curtain may be installed at the opening of the housing where the first to sixth vision cameras (11)(12)(13)(14)(15)(16) are exposed, to create an air film by blowing compressed air from top to bottom to block the entry of foreign substances into the housing.

[0074] The turn guide (20) is installed in the form of a stainless steel structure under the rail (R) to induce the bipart conductor (C) to naturally rotate 90 degrees along with the gamble (H) as it moves along the rail (R) by the movement of the trolley (T) while suspended from the A-type rotary gamble (H).

[0075] And the turn guide (20) serves to support the first vision camera (11) so that it does not shake when the first vision camera (11) photographs the belly portion of the bipartite conductor (C) under the control of the controller (60).

[0076] The first to fourth sensors (21)(22)(23)(24) are installed adjacent to the chroma key screen (10) and emit light signals such as infrared rays and receive reflected signals, thereby detecting the presence of a bipartite conductor (C) located inside the chroma key screen (10) by moving along the rail (R) by the movement of the trolley (T) while suspended from the gamble (H).

[0077] That is, the first sensor (21) detects a bipartite conductor (C) located inside the chroma key screen (10), converts it into an electrical signal, and outputs it to the controller (60).

[0078] The second sensor (22) detects the bipartite conductor (C) located inside the chroma key screen (10) after passing through the first sensor (21), converts it into an electrical signal, and outputs it to the controller (60).

[0079] The third sensor (23) detects the position of the right conductor among the two conductors (C) located inside the chroma key screen (10) after passing through the second sensor (22), converts it into an electrical signal, and outputs it to the controller (60).

[0080] The fourth sensor (24) detects the position of the left conductor among the two conductors (C) located inside the chroma key screen (10) after passing through the third sensor (23), converts it into an electrical signal, and outputs it to the controller (60).

[0081] Here, the first to fourth sensors (21)(22)(23)(24) may be non-contact type sensors such as a Doppler sensor that detects reflected waves from a bipolar conductor (C) using ultrasound or microwaves, a reflective type that detects reflected light from a bipolar conductor (C) by combining a light source and infrared rays, or a light-blocking type that detects when the front of the receiver is obscured by the bipolar conductor (C).

[0082] The spread guide (30) is installed in the form of a stainless steel structure under the rail (R) to naturally spread the lower part of the bipartite conductor (C), which moves in a state rotated 90 degrees by the turn guide (20), in both left and right directions of the rail (R).

[0083] And at the leading end of the spread guide (30), a plurality of guide rollers (31) are installed to guide the bipart conductor (C), which moves along the rail (R) while suspended from the gamble (H) by the movement of the trolley (T), so that it spreads smoothly without shaking.

[0084] In addition, guard rails (32) are installed at a certain distance on both the left and right sides of the spread guide (30) to guide the bipartite conductor (C), which moves along the rail (R) while suspended from the gamble (H) by the movement of the trolley (T), so that it does not deviate outward.

[0085] The pull rod (40) is installed under the rail (R) to temporarily pull the lower part of the right conductor of the two conductors (C) moving along the rail (R) by the movement of the trolley (T) while suspended from the gamble (H) in the opposite direction of movement.

[0086] Specifically, the pull rod (40) is composed of a rotary actuator (41) installed in the front center part of the spread guide (30) and controlled by a controller (60), and a lever arm (42) that rotates clockwise by 90 degrees by the rotary actuator (41) to temporarily pull and hold the lower part of the right conductor of the two conductors (C) in the opposite direction to the direction of movement.

[0087] The controller (60) performs overall control functions.

[0088] For example, the controller (60) controls the first vision camera (11) by the signal of the first sensor (21), which detects the position of the bipartite conductor (C) moving along the rail (R) by the movement of the trolley (T) while suspended from the gamble (H) and converts it into an electrical signal to output.

[0089] Then, the pull load (40) and the third vision camera (13) are simultaneously controlled by the signal from the second sensor (22), which detects the position of the bipartite conductor (C) moving through the first sensor (21) and converts it into an electrical signal to output it, and the signal from the third sensor (23), which detects the position of the right conductor among the bipartite conductors (C) moving through the second sensor (22) and converts it into an electrical signal to output it.

[0090] In addition, the fourth to sixth vision cameras (14)(15)(16) are controlled by the signal of the fourth sensor (24), which detects the position of the left conductor among the bipartite conductors (C) moving through the third sensor (23), converts it into an electrical signal, and outputs it.

[0091] A reader (70) is mounted on a rail (R) to obtain unique identification information by tagging an RFID tag attached to or embedded in a gamble (H) on which a bipartite conductor (C) is suspended, and then sending it to a computing device (100).

[0092] The setting board (80) serves to determine the angle of view adjustment criteria and upper / lower positions of the 4th to 6th vision cameras (14)(15)(16) in order to set the overlapping section between the image captured by the 4th vision camera (14) and the 5th vision camera (15), which divide the cross-section of the left conductor of the bipartite conductor (C) into upper, central, and lower parts and capture it, and the overlapping section between the image captured by the 5th vision camera (15) and the 6th vision camera (16).

[0093] That is, on the front of the setting board (80), reference points (81) or reference lines for determining the angle of view adjustment criteria and upper / lower positions of the 4th to 6th vision cameras (14)(15)(16) are formed in a certain pattern.

[0094] The computing device (90) receives carcass images of the biparted carcass (C), synthesizes them to extract feature points, analyzes them through an artificial intelligence model to determine the sex and whether the biparted carcass (C) is a sow, measures the bone segments, back fat thickness per bone segment, body length and width, neck condition and fat thickness, and determines the grade based on this.

[0095] And the computing device (90) uses an artificial intelligence model to determine and label the item-specific parts (gender classification, multifidus muscle thickness, vertebral segments, back fat thickness / area, body length / body width, outline, and pelvis) of the bipartite body (C), generates an artificial intelligence model for each labeling unit through artificial intelligence deep learning, and stores the analysis results in a database.

[0096] That is, the computing device (90) labels each vertebral segment consisting of 7 lumbar segments, 14 thoracic segments, and 5 to 7 cervical segments from the conductor image, labels the back fat extending from 7 lumbar segments to 1 cervical segment, labels the outline of the entire left conductor, labels the location of the multifidus muscle region, and labels the location of the pelvic region.

[0097] In addition, the artificial intelligence model of the computing device (90) can detect sex characteristic points based on carcass images of standard pigs (1+ grade; carcass weight 83~92 kg, back fat thickness 17~24 mm) and generate an artificial intelligence model for sex determination of sows, boars, castrated meat, etc. through deep learning learning by classification by items of sex determination target (penile ring, ischiocavernosus muscle, semimembranosus muscle, penile eyelid muscle).

[0098] And the quality evaluation and analysis program of the computing device (90) can increase the measurement rate and accuracy of the analysis items through applications that apply artificial intelligence models such as image signal processing, machine learning, and deep learning, and image learning algorithms such as CNN and yolact.

[0099] Here, the computing device (90) may include a laptop, desktop, laptop, tablet PC, slate PC, etc. equipped with a web browser.

[0100] Additionally, the computing device (90) may include an application server, computing server, database server, file server, game server, mail server, proxy server, and web server that communicate with an external device to process information, or may receive and store an artificial intelligence model optimized by an external server from an external server.

[0101] In addition, the computing device (90) can receive weight information regarding the weight of the bipartite (C) from the outside, and can infer multiple parameters regarding the meat based on the image of the carcass, the weight of the carcass, and the grade of the determined carcass through an artificial intelligence model.

[0102] For example, multiple parameters may include measurement items such as total meat yield, meat yield rate, meat yield of major cuts, and predicted values ​​for specific cuts.

[0103] Additionally, the computing device (90) can compare the grade of the conductor included in the log data of the previously recorded grade measurement system with the grade of the conductor determined through the artificial intelligence model, and verify and retrain the processing result of the artificial intelligence model based on the result.

[0104] Meanwhile, the computing device (90) is configured to include a gender determination part tracking unit (91), a cropping unit (92), a gender learning unit (93), a vertebral segment tracking unit (94), a labeling unit (95), a storage unit (96), a conductor width calculation unit (97), a conductor length calculation unit (98), and a control unit (99), as shown in FIG. 8.

[0105] The gender determination area tracking unit (91) tracks the gender determination area from the conductor image of the bipartite conductor (C).

[0106] If the bipartite conductor (C) is accurately suspended from the gamble (H), the gender determination part is located without deviating significantly from the preset area, and the gender determination part tracking part (91) tracks the location.

[0107] The cropping unit (92) crops the area corresponding to the tracked gender determination part when the gender determination part is tracked by the gender determination part tracking unit (91).

[0108] The gender learning unit (93) learns a gender reading model through a conductor image of an area cropped by the cropping unit (92).

[0109] That is, the accuracy of gender reading improves as the learning action of the gender learning unit (93) is repeated.

[0110] The vertebral segment tracking unit (94) tracks the vertebral segments from the conductor image of the bipartite conductor (C).

[0111] The labeling unit (95) individually labels the vertebrae tracked by the vertebra tracking unit (94).

[0112] The storage unit (96) stores all information necessary for the operation of the computing device (90).

[0113] For example, the coordinates of the vertebrae labeled by the labeling unit (95) are stored in a database.

[0114] And the storage unit (96) includes a database.

[0115] The body width calculation unit (97) calculates the coordinates corresponding to the midpoint between the lumbar and thoracic vertebrae using the coordinates of the vertebrae, and calculates the body width for each position using the calculated coordinates.

[0116] Here, the carcass width by location includes the carcass width including the fat of the 4th and 5th thoracic vertebrae of the bipartite carcass (C), the carcass width excluding the fat of the 4th and 5th thoracic vertebrae, and the carcass width including the fat of the 4th and 5th lumbar vertebrae.

[0117] In addition, the conductor width calculation unit (97) can calculate the conductor width for each position using the following mathematical formula 1.

[0118] [Mathematical Formula 1]

[0119] W adepth = Length a × rate

[0120] W bdepth = Length b × rate

[0121] W cdepth = Length c × rate

[0122] Here, W adepth , Wbdepth , W cdepth is the conductor width and Length for each of the above positions a is the carcass width including the fat of the 4th and 5th thoracic vertebrae mentioned above, Length b is the carcass width excluding the fat of the 4th and 5th thoracic vertebrae mentioned above, Length c is the width of the body including the fat of the 4th and 5th lumbar vertebrae mentioned above, and rate is the ratio between the actual body and the body image.

[0123] The carcass length calculation unit (98) calculates the carcass length by position using the coordinates of the vertebrae and the coordinates of the carcass's half-bones.

[0124] Here, the length of the carcass by position includes the length from the top of the lumbar spine to the 1st cervical vertebra, the length from the bottom of the lumbar spine to the 1st cervical vertebra, the length from the 7th lumbar vertebra to the top of the 1st thoracic vertebra, the length from the top of the lumbar spine to the 1st thoracic vertebra, and the length from the bottom of the lumbar spine to the 1st thoracic vertebra of the bipartite carcass (C).

[0125] In addition, the conductor length calculation unit (98) can calculate the conductor length for each location using the following mathematical formula 2.

[0126] [Mathematical Formula 2]

[0127] W alength =Length d ×rate

[0128] W blength =Length e ×rate

[0129] W clength =Length f ×rate

[0130] W dlength =Length g ×rate

[0131] W elength =Length h ×rate

[0132] Here, W alenth , W blenth , W clenth , W dlenth , Welenth is the conductor length by the above location, Length d is the length from the top of the helix to the 1st cervical vertebra, Length e is the length from the lower end of the helix to the 1st cervical vertebra, Length f is the length from the 7th lumbar vertebra to the top of the 1st thoracic vertebra. g is the length from the top of the helix to the 1st thoracic vertebra, Length h is the length from the lower end of the thorax to the 1st thoracic vertebra, and rate is the ratio between the actual body and the body image.

[0133] The control unit (99) performs the overall control function of the computing device (90).

[0134] That is, the control unit (99) controls signal input and output between the gender determination part tracking unit (91), cropping unit (92), gender learning unit (93), vertebral segment tracking unit (94), labeling unit (95), storage unit (96), conductor width calculation unit (97), and conductor length calculation unit (98).

[0135] Hereinafter, the process of analyzing a carcass image taken of a pig carcass using an image acquisition device and a computing device (90) for quality evaluation of a pig carcass according to an embodiment of the present invention will be explained in detail with reference to FIGS. 9 to 16.

[0136] Figure 9 shows an image of the vertebral segments and back fat region of a bipartite body (C).

[0137] Here, the average back fat thickness is measured from between the 1st lumbar vertebra and the 14th thoracic vertebra to between the 11th thoracic vertebra and the 12th thoracic vertebra.

[0138] In addition, the fat thickness of the multifidus muscle refers to the fat thickness in the center of the multifidus muscle.

[0139] Figure 10 shows an image of an area for measuring conductor width and conductor length.

[0140] Here, carcass width a is the carcass width including fat of the 4th and 5th thoracic vertebrae, carcass width b is the carcass width excluding fat of the 4th and 5th thoracic vertebrae, and carcass width c is the carcass width including fat of the 4th and 5th lumbar vertebrae.

[0141] In addition, carcass length a is the length from the top of the pelvis to the 1st cervical vertebra, carcass length b is the length from the bottom of the pelvis to the 1st cervical vertebra, carcass length c is the length from the 7th lumbar vertebra to the top of the 1st thoracic vertebra, carcass length d is the length from the top of the pelvis to the 1st thoracic vertebra, and carcass length e is the length from the bottom of the pelvis to the 1st thoracic vertebra.

[0142] Figure 11 shows an image with each part marked to train an artificial intelligence model. Each part is marked directly within the conductor image, and the artificial intelligence model undergoes continuous learning through the conductor image, and as learning is repeated, the analysis performance of the conductor image improves.

[0143] Here, the conductor image includes a mechanical rod for fixing the bipartite conductor (C), and the artificial intelligence model can detect the position of the rod and delete it by determining it as an unnecessary image.

[0144] FIG. 12 shows a comparison of two types of conductor images for gender determination. To determine gender using the conductor images, the gender determination area tracking unit (91) tracks the gender determination area, and the cropping unit (92) crops the area corresponding to the gender determination area.

[0145] Afterwards, the gender learning unit (93) trains a gender identification model using images of the cropped area.

[0146] In this case, the VGG-19 algorithm can be applied to the gender identification model. Images classified by the gender identification model are stored in a database.

[0147] In Fig. 12, the area (A1) corresponding to the sex determination area of ​​(a) shows female genitalia, and the area (A2) corresponding to the sex determination area of ​​(b) shows male genitalia in a castrated state.

[0148] FIG. 13 shows an image of a bipartite conductor (C) with the vertebrae extracted. Coordinates are assigned to each vertebra. The number of vertebrae varies depending on the conductor, ranging from 26 to 28. Therefore, the vertebrae tracking unit (94) can track the vertebrae from the conductor image to determine the number of vertebrae, and using this result, the labeling unit (95) individually labels the vertebrae and assigns coordinates.

[0149] Figure 14 shows the result of tracking the isofat region of a bipartite conductor (C). The operation of tracking the isofat region can be performed by the Yolact algorithm. Once the isofat region is tracked, coordinates are assigned to the isofat region, and the corresponding image is stored in a database.

[0150] FIG. 15 shows the result of detecting the entire area of ​​the bipart conductor (C). The coordinates corresponding to the entire bipart conductor (C) can be stored in a database. The conductor width can be calculated by detecting the entire bipart conductor (C). The operation of detecting the entire area of ​​the bipart conductor (C) can be performed by the Yolact algorithm.

[0151] Figure 16 shows the results of the segmentation processing of the back fat region by vertebral segment. The back fat is divided into 26 to 28 masked regions corresponding to the area of ​​each vertebral segment, and the number of all vertical pixels in the segmented back fat masked regions is measured to calculate the average number of pixels.

[0152] In addition, the back fat thickness corresponding to each vertebral segment is calculated by multiplying the calculated number of pixels by the length per pixel. Finally, the back fat thickness corresponding to the contact points of each vertebral segment is calculated to measure the carcass width.

[0153] In this way, the parts and sizes of each bipartite conductor (C) can be accurately analyzed and calculated through the conductor image.

[0154] Meanwhile, a method for determining a grade using an image acquisition device for quality evaluation of a pig carcass according to an embodiment of the present invention is described as follows.

[0155] (a) A dividing carcass (C) is suspended from a trolley (T) that travels along a rail (R) installed on the ceiling of a pig carcass slaughterhouse using a rotary gamble (H), and the dividing carcass (C) moves and detects whether it is located inside a chroma key screen (10) that is fixedly installed at a certain distance below both the left and right sides of the rail (R).

[0156] (b) When the bipart conductor (C) is positioned inside the chroma key screen (10), the first sensor (21) detects this and converts it into an electrical signal, which is then output to the controller (60). The first vision camera (11) then photographs the belly portion of the bipart conductor (C) under the control of the controller (60) and transmits the resulting color conductor image to the computing device (90).

[0157] (c) The bipartite conductor (C), which moves along the rail (R) while suspended from the gamble (H) by the movement of the trolley (T), is induced to change direction by the turn guide (20) through contact interference and rotated 90 degrees together with the gamble (H).

[0158] (d) When the second sensor (22) detects a bipartite conductor (C) moving along the rail (R) while rotated 90 degrees by the turn guide (20), it converts the signal into an electrical signal and outputs it to the controller (60). The second vision camera (12), under the control of the controller (60), photographs the conductor number displayed on the rear leg portion of the bipartite conductor (C) located within the chroma key screen (10) and transmits the resulting color conductor image to the computing device (90).

[0159] (e) The spread guide (30) guides the bipart conductor (C), which moves along the rail (R) while rotated 90 degrees by the turn guide (20), to spread out to the left and right sides through contact interference, thereby spreading the lower part of the bipart conductor (C) in both left and right directions of the rail (R).

[0160] (f) When the third sensor (23) detects a bipartite conductor (C) moving through the second sensor (22), it converts the signal into an electrical signal and outputs it to the controller (60). The third vision camera (13), under the control of the controller (60), captures the side of the right conductor of the bipartite conductor (C) located within the chroma key screen (10) and transmits the resulting color conductor image to the computing device (90).

[0161] (d) When the fourth sensor (24) detects the position of the right conductor among the bipart conductors (C) moving through the third sensor (23), it converts the signal into an electrical signal and outputs it to the controller (60), and the pull load (40) temporarily pulls the lower part of the right conductor among the bipart conductors (C) in the opposite direction to the direction of movement under the control of the controller (60).

[0162] (a) In addition, the right conductor of the bipartite conductor (C) located inside the chroma key screen (10) is pulled in the direction opposite to the direction of movement by the control of the controller (60), and the cross-section of the left conductor that is revealed is captured by the 4th to 6th vision cameras (14)(15)(16) by dividing it into upper, central, and lower sections and transmitting the conductor images obtained to the computing device (90).

[0163] (i) Afterwards, images of the carcass of the split carcass (C) are received from the computing device (90), synthesized to extract feature points, and analyzed through an artificial intelligence model to determine the sex and whether the split carcass (C) is a sow, and the back fat thickness, body length and width, neck condition and fat thickness are measured, and the grade is determined based on this.

[0164] In addition, the computing device (90) receives information regarding the weight of the carcass of the split carcass (C) from an external source such as a reader (70), and determines whether it is a sow based on the carcass image obtained by photographing the belly portion of the split carcass (C), the weight of the carcass, the length of the carcass, and the width of the carcass.

[0165] In this process, the gender determination part tracking unit (91) of the computing device (90) tracks the gender determination part using a conductor image, the cropping unit (92) crops the area corresponding to the gender determination part, and determines the gender by a gender reading model.

[0166] In addition, conductor images are used to identify each part and assign coordinates, and the information identified and labeled by the artificial intelligence model is stored in a database.

[0167] In addition, a linear regression algorithm is used to learn pig carcass yield data and predict and determine the yield.

[0168] Therefore, by automatically measuring quality evaluation items of pig carcasses, it is possible not only to accurately determine the final meat quality grade at a rapid pace, but also to measure back fat and fat content within the carcass, calculate LMP (Lean Meat Percentage), and even measure the expected yield for each part of the carcass.

[0169] Meanwhile, as shown in Fig. 17, the computing device (slaughterhouse ERP PC) in the slaughterhouse office can receive and store information regarding the grade of pig carcasses determined from a cloud server.

[0170] As shown in Fig. 18, the AI-based conductor grading system can monitor conductor images captured in real time.

[0171] That is, ① it displays the slaughter date and carcass number, ② the status of carcass photography and analysis, ③ shows the carcass analysis values ​​measured by AI, and ⑤ shows the carcass image being captured in real time.

[0172] As shown in Fig. 19, the AI-based conductor grading system can monitor CCTV video and still images captured in real time.

[0173] That is, ④ the scene of the pig carcass moving can be viewed as real-time CCTV video, and ⑤ internal still images of the wired and left carcasses are displayed.

[0174] As shown in Fig. 20, the artificial intelligence-based conductor grading system can verify the grading results.

[0175] In other words, ① you can search for a pig carcass by entering the slaughter date and carcass number and pressing the search button, ② you can view the AI ​​measurement analysis value for the searched carcass number, and ③ you can output an image of the searched slaughter date and carcass number.

[0176] And as shown in Fig. 21, clicking on the image for the searched slaughter date and carcass number can open an image popup.

[0177] In addition, as shown in Fig. 22, the image can be enlarged and reduced using a mouse in the image popup, and the distance can be measured by clicking two points in the image.

[0178] As shown in FIGS. 23 and 24, the input screen of the AI-based conductor grading system can view statistical data, conductor information, and conductor images in horizontal or vertical orientation.

[0179] Meanwhile, it is obvious to those skilled in the art that the present invention is not limited by the embodiments described above and the attached drawings, and that it can be modified and applied in various ways not exemplified within the scope of the technical concept of the present invention, as well as widely applied by substituting each component and changing to equivalent alternative embodiments.

[0180] Therefore, content related to modifying and applying the technical features of the present invention should be interpreted as being included within the technical concept and scope of the present invention.

[0181] The image acquisition device for quality evaluation of a pig carcass and the grading method using the same according to an embodiment of the present invention are industrially applicable because they collect accurate image data for each part of a pig carcass and synthesize and analyze it into a single image to increase the accuracy and speed of pig carcass grading.

Claims

1. A device for acquiring multi-faceted image of a divided carcass during the process of transporting a divided carcass suspended from left and right sides using a rotary gamble (a hook-shaped hook for hanging pig carcasses) on a trolley that travels along a rail installed on the ceiling inside a slaughterhouse. A chroma key screen fixedly installed at a certain distance below both the left and right sides of the above rail; A first vision camera installed below the side of the above rail and transmitting a conductor image obtained by photographing the belly portion of the above-mentioned bipartite conductor located within the above-mentioned chroma key screen; A turn guide installed below the rail and guiding the bipart conductor, while suspended from the gamble, to move along the rail by the movement of the trolley and rotate 90 degrees together with the gamble; A spread guide installed below the rail and guiding the lower part of the bipartite conductor, rotated 90 degrees by the turn guide, to spread out in both left and right directions of the rail; A second vision camera installed on the lower right side of the above rail and transmitting a conductor image obtained by capturing the conductor number of the right conductor among the two conductors located within the chroma key screen; A third vision camera installed on the lower right side of the above rail and transmitting a conductor image obtained by photographing the side surface of the right conductor among the two conductors located within the chroma key screen; A pull rod installed below the rail and pulling the lower part of the right conductor of the bipartite conductor, which moves along the rail by the movement of the trolley while suspended from the gamble, in a direction opposite to the direction of movement; Fourth to sixth vision cameras, each installed on the lower right side of the above rail and positioned within the chroma key screen, which transmit conductor images obtained by dividing and photographing the cross-section of the left conductor—which is revealed as the right conductor of the bipartite conductor is pulled in the direction opposite to the direction of movement by the pull rod—into the upper, central, and lower parts; An image acquisition device for quality evaluation of a pig carcass, comprising 2. In Paragraph 1, It further includes a controller that performs overall control functions, The above controller is, The first vision camera is controlled by the signal of the first sensor, which detects the position of the bipartite conductor moving along the rail by the movement of the trolley while suspended from the above gamble, converts it into an electrical signal, and outputs it. The second vision camera is controlled by the signal of the second sensor, which detects the position of the bipartite conductor moving through the first sensor, converts it into an electrical signal, and outputs it. The pull load and the third vision camera are controlled by the signal of the third sensor, which detects the position of the right conductor among the bipartite conductors moving through the second sensor and converts it into an electrical signal to output it. An image acquisition device for quality evaluation of a pig carcass, which controls the fourth to sixth vision cameras respectively by means of a signal from a fourth sensor that detects the position of the left conductor among the bipart conductors moving through the third sensor and converts it into an electrical signal to output.

3. In Paragraph 2, The above full load is, A rotary actuator controlled by the above controller; A lever arm that rotates at a constant angle by the rotary actuator to temporarily pull and hold the lower part of the right conductor of the bipartite conductor in the direction opposite to the direction of movement; An image acquisition device for quality evaluation of a pig carcass, comprising 4. In Paragraph 1, A lighting module that brightly illuminates the above-mentioned bipartite conductor; A guide roller installed at the leading end of the spread guide and guiding the bipartite conductor, which moves along the rail by the movement of the trolley while suspended from the gamble, to spread smoothly without shaking; Guard rails installed on both the left and right sides of the spread guide and guiding the bipartite conductor, which moves along the rail by the movement of the trolley while suspended from the gamble, so as not to deviate outward; An image acquisition device for quality evaluation of a pig carcass, further comprising 5. In Paragraph 1, The above 4th to 6th vision cameras are, Each is mounted on a vertical bar so that it can move up and down, and A setting board having reference points for determining the angle of view adjustment criteria and upper / lower positions of the 4th to 6th vision cameras, for setting the overlapping section between the image captured by the 4th vision camera and the 5th vision camera, which captures the cross-section of the left conductor of the bipartite by dividing it into upper, central, and lower parts, and the overlapping section between the image captured by the 5th vision camera and the 6th vision camera; An image acquisition device for quality evaluation of a pig carcass, further comprising 6. In Paragraph 1, A reader that tags an RFID tag attached to the gamble on which the above-mentioned bipartite conductor is suspended, obtains unique identification information, and sends it to a computing device; Includes more, The above computing device is an image acquisition device for quality evaluation of pig carcasses, which receives and synthesizes carcass images of the biparted carcass to extract feature points, analyzes them through an artificial intelligence model to determine the sex and whether the biparted carcass is a sow, measures backfat thickness, body length, and body width, verifies the carcass number, and determines the grade based on the above.

7. In Paragraph 6, The above computing device is, A gender determination region tracking unit that tracks a gender determination region from a conductor image of the above-mentioned bipartite conductor; A cropping unit that crops an area corresponding to the tracked gender determination part; and Gender learning unit that learns a gender reading model through conductor images of cropped areas; An image acquisition device for quality evaluation of a pig carcass, comprising 8. In Paragraph 6, The above computing device is, A vertebral segment tracking unit that tracks vertebral segments from the conductor image of the above-mentioned bipartite conductor; A labeling unit that individually labels tracked vertebrae; A storage unit that stores the coordinates of labeled vertebrae in a database; A conductor width calculation unit that calculates coordinates corresponding to the midpoint between the lumbar and thoracic vertebrae using the stored coordinates of the vertebrae, and calculates the conductor width for each position using the calculated coordinates; An image acquisition device for quality evaluation of a pig carcass, comprising 9. In Paragraph 8, The conductor width for each of the above positions is, An image acquisition device for quality evaluation of a pig carcass, comprising a carcass width including fat of the 4th and 5th thoracic vertebrae and lumbar vertebrae of the above-mentioned bipartite, and a carcass width excluding fat of the 4th and 5th thoracic vertebrae.

10. In Paragraph 9, The above conductor width calculation unit is, An image acquisition device for quality evaluation of pig carcasses, which calculates the carcass width for each position using the following formula. W adepth =Length a ×rate W bdepth =Length b ×rate W cdepth =Length c ×rate Here, W adepth , W bdepth , W cdepth is the conductor width and Length for each of the above positions a is the carcass width including the fat of the 4th and 5th thoracic vertebrae mentioned above, Length b is the carcass width excluding the fat of the 4th and 5th thoracic vertebrae mentioned above, Length c is the width of the body including the fat of the 4th and 5th lumbar vertebrae mentioned above, and rate is the ratio between the actual body and the body image.

11. In Paragraph 6, The above computing device is, A conductor length calculation unit that calculates the conductor length at each position using the coordinates of the stored vertebrae and the semi-bone coordinates of the bipartite conductor; An image acquisition device for quality evaluation of a pig carcass, further comprising 12. In Paragraph 11, The conductor lengths for each location mentioned above are, An image acquisition device for quality evaluation of a pig carcass, comprising the length from the upper end of the carcass to the first cervical vertebra, the length from the lower end of the carcass to the first cervical vertebra, the length from the seventh lumbar vertebra to the upper end of the first thoracic vertebra, the length from the upper end of the carcass to the first thoracic vertebra, and the length from the lower end of the carcass to the first thoracic vertebra.

13. In Paragraph 12, The above conductor length calculation unit is, An image acquisition device for quality evaluation of pig carcasses, which calculates the carcass length for each position using the following formula. W alength =Length d ×rate W blength =Length e ×rate W clength =Length f ×rate W dlength =Length g ×rate W elength =Length h ×rate Here, W alenth , W blenth , W clenth , W dlenth , W elenth is the conductor length by the above location, Length d is the length from the top of the helix to the 1st cervical vertebra, Length e is the length from the lower end of the helix to the 1st cervical vertebra, Length f is the length from the 7th lumbar vertebra to the top of the 1st thoracic vertebra. g is the length from the top of the helix to the 1st thoracic vertebra, Length h is the length from the lower end of the thorax to the 1st thoracic vertebra, and rate is the ratio between the actual body and the body image.

14. A grading method using an image acquisition device for quality evaluation of pig carcasses, comprising each of the following steps. (a) A step of detecting the position of a bifurcated carcass, which is transported by being suspended from a trolley using a gambler and traveling along a rail installed on the ceiling inside a pig carcass slaughterhouse; (b) A step of photographing the abdomen of the bipartite conductor located within a chroma key screen fixedly installed at a certain distance below both the left and right sides of the rail, and transmitting the conductor image to a computing device; (c) A step of rotating the bipartite conductor, which moves along the rail by the movement of the trolley while suspended from the gamble, 90 degrees together with the gamble; (d) A step of photographing the conductor number marked on the rear leg portion of the bipartite conductor rotated 90 degrees and transmitting the conductor image to the computing device; (e) A step of spreading the lower part of the bipartite conductor, rotated 90 degrees, in both left and right directions of the rail; (f) A step of photographing the side surface of the right conductor among the bipart conductors located within the chroma key screen and transmitting the conductor image to the computing device; (d) A step of pulling the lower part of the right conductor of the above-mentioned bipartition conductor toward the direction opposite to the direction of movement; (a) A step of photographing the cross-section of the left conductor, which is revealed as the right conductor among the bipartition conductors located within the chroma key screen is pulled in the direction opposite to the direction of movement, by dividing it into upper, central, and lower sections, and transmitting the conductor images to the computing device; (i) A step of receiving carcass images of the split carcass from the computing device, synthesizing them to extract feature points, analyzing them through an artificial intelligence model to determine the sex and whether the split carcass is a sow, measuring back fat thickness, body length, and body width, verifying the carcass number, and determining the grade based on the above; 15. In Paragraph 14, A method for determining grade using an image acquisition device for quality evaluation of pig carcasses, wherein the computing device of the above (ja) step receives information regarding the carcass weight of the biparted carcass from the outside, and determines whether it is a sow based on the carcass image obtained by photographing the belly portion of the biparted carcass and the carcass weight, body length, and body width.

16. In Paragraph 14, The above artificial intelligence model is, A method for determining grade using an image acquisition device for quality evaluation of pig carcasses, which detects sex characteristic points based on carcass images of standard pigs (Grade 1+; carcass weight 83~92kg, back fat thickness 17~24mm) and generates a sex determination artificial intelligence model through deep learning training by classification by sex determination target items (penile ring, ischiocavernosus muscle, semimembranosus muscle, penile eyelid muscle).

Images