System and method for pig carcass grading and yield quality analysis

Abstract

A system and method for determining pig carcass grades and analyzing yield quality using AI are disclosed. The system for determining pig carcass grades and analyzing yield quality using AI according to an embodiment of the present invention comprises: a carcass moving device including a blue board; a guide device for moving a carcass and guiding it to a point where the blue board is installed; an infrared light sensor installed adjacent to the blue board to recognize the carcass; a camera installed opposite the blue board that captures the carcass on the blue board and generates a carcass image when the carcass is recognized; and an image analysis device that determines the sex of the carcass using the generated carcass image and determines and labels the carcass part using an AI model.

Description

Technology Field

[0001] The present invention relates to a system and method for determining pig carcass grades and analyzing yield quality, and more specifically, to a system and method for determining pig carcasses grades and analyzing yield quality that can accurately analyze pig carcasses without using expensive equipment. Background Technology

[0002] Part-by-part analysis is crucial for analyzing pig carcasses. While this is not an issue for small-scale butchering, an automated system is required for large-scale operations. Accordingly, the VCS2000 system developed by Germany's E+V is primarily used for pig carcass analysis.

[0003] The VCS2000 system consists of one black-and-white camera and two color cameras, and photographs the pig carcass by photographing the pig's hind legs with the black-and-white camera and photographing the pig's upper and lower bodies with the color cameras respectively, and obtaining a single color image through synthesis.

[0004] However, the VCS2000 system is subject to various obstacles due to its highly complex structure in which the related equipment for photographing pig carcasses is intricately interconnected. For example, vibrations can cause the limit switch connections to become misaligned, resulting in a mismatch in the camera's shooting timing. This leads to a degradation in image quality and a loss of identification accuracy. Additionally, since the system photographs the upper and lower halves of the pig carcass separately and then composites them, image distortion may occur, which also affects identification accuracy.

[0005] As the VCS2000 system was developed in Germany, it was designed to suit the slaughter and butchery conditions in Europe. While ham production in Europe primarily relies on pork hind legs, the domestic market requires analysis centered on pork belly. Therefore, the software provided by the VCS2000 system has a problem in that it is not compatible with the slaughter and butchery situation in Korea.

[0006] Furthermore, domestic companies are experiencing significant inconvenience due to the lack of domestic solutions for product defects in the VCS2000 system, which makes maintenance procedures very complicated, and the gap in mutual understanding with the German local market. Prior art literature

[0007] Korean Patent Publication No. 10-2023-0115371 (Published Aug. 3, 2023) The problem to be solved

[0008] Therefore, the technical problem and objective of the present invention is to provide a pig carcass grading and yield quality analysis system and method that enables accurate and rapid analysis by training and labeling images of pig carcasses.

[0009] The problems solved by the present invention are not limited to those mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0010] To achieve the aforementioned technical challenges and objectives, a pig carcass grading and yield quality analysis system according to an embodiment of the present invention comprises: a carcass moving device including a blue board; a guide device that moves the carcass and guides it to a point where the blue board is installed; an infrared light sensor installed adjacent to the blue board to recognize the carcass; a camera installed opposite the blue board that captures the carcass on the blue board to generate a carcass image when the carcass is recognized; and an image analysis device that uses the generated carcass image to determine the sex of the carcass and to determine and label the parts of the carcass.

[0011] Preferably, the image analysis device may include a gender determination area tracking unit that tracks a gender determination area from a conductor image, a cropping unit that crops an area corresponding to the tracked gender determination area, and a gender learning unit that learns a gender reading model through an image of the cropped area.

[0012] Additionally, preferably, the image analysis device may include a vertebra tracking unit that tracks vertebrae from a conductor image, a labeling unit that individually labels the tracked vertebrae, and a storage unit that stores the coordinates of the labeled vertebrae in a database.

[0013] Additionally, preferably, the image analysis device may further include a conductor width calculation unit that calculates coordinates corresponding to the midpoint between the lumbar and thoracic vertebrae using the stored coordinates of the vertebral segments, and calculates the conductor width for each position using the calculated coordinates.

[0014] Additionally, preferably, the carcass width by location may include a carcass width including fat of the 4th and 5th thoracic vertebrae, a carcass width excluding fat of the 4th and 5th thoracic vertebrae, and a carcass width including fat of the 4th and 5th lumbar vertebrae.

[0015] Additionally, preferably, the conductor width calculation unit can calculate the conductor width for each location using the following formula:

[0016]

[0017]

[0018]

[0019] Here, W adepth , W bdepth , W cdepth is the conductor width by location, Length a The carcass width including the fat of the 4th and 5th thoracic vertebrae, Length b is the carcass width excluding fat at the 4th and 5th thoracic vertebrae, Length c is the width of the body including fat of the 4th and 5th lumbar vertebrae, and rate is the ratio between the actual body and the body image.

[0020] Additionally, preferably, the image analysis device may further include a body length calculation unit that calculates the body length at each position using the stored coordinates of the vertebral segments and the coordinates of the body's semi-segments.

[0021] Additionally, preferably, the length of the carcass by position may include the length from the top of the lumbar spine to the first cervical vertebra, the length from the bottom of the lumbar spine to the first cervical vertebra, the length from the seventh lumbar vertebra to the top of the first thoracic vertebra, the length from the top of the lumbar spine to the first thoracic vertebra, and the length from the bottom of the lumbar spine to the first thoracic vertebra.

[0022] Additionally, preferably, the conductor length calculation unit can calculate the conductor length for each location using the following formula:

[0023]

[0024]

[0025]

[0026]

[0027]

[0028] Here, W alenth , W blenth , W clenth , W dlenth , W elenthConductor length by 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.

[0029] Meanwhile, a method for determining pig carcass grade and analyzing yield quality according to another embodiment of the present invention is a carcass analysis method applied to a system comprising a carcass moving device including a blue board and a camera that generates a carcass image for a carcass moved by said carcass moving device, wherein when the carcass is moved to a point where the blue board is installed, the method includes the steps of: recognizing the carcass; when the carcass is recognized, photographing the carcass on the blue board to generate a carcass image; determining the sex of the carcass using the generated carcass image; and determining and labeling a part of the carcass.

[0030] Preferably, the determination step may include the step of tracking a gender determination region from a conductor image, the step of cropping an area corresponding to the tracked gender determination region, and the step of training a gender determination model through an image of the cropped area.

[0031] Additionally, preferably, the labeling step may include the step of tracking vertebrae from a conductor image, the step of individually labeling the tracked vertebrae, and the step of storing the coordinates of the labeled vertebrae in a database.

[0032] Additionally, preferably, the labeling step may further include the step of calculating coordinates corresponding to the midpoint between the lumbar and thoracic vertebrae using the stored vertebral segment coordinates, and the step of calculating the conductor width at each location using the calculated coordinates.

[0033] Additionally, preferably, the carcass width by position may include a carcass width including fat of the 4th and 5th thoracic vertebrae, a carcass width excluding fat of the 4th and 5th thoracic vertebrae, and a carcass width including fat of the 4th and 5th lumbar vertebrae.

[0034] Additionally, preferably, the calculation step can calculate the conductor width for each location using the following formula:

[0035]

[0036]

[0037]

[0038] Here, W adepth , W bdepth , W cdepth is the conductor width by location, Length a The carcass width including the fat of the 4th and 5th thoracic vertebrae, Length b is the carcass width excluding fat at the 4th and 5th thoracic vertebrae, Length c is the width of the body including fat of the 4th and 5th lumbar vertebrae, and rate is the ratio between the actual body and the body image.

[0039] Additionally, preferably, the labeling step may further include a step of calculating the carcass length by position using the stored coordinates of the vertebral segments and the semi-vertebral coordinates of the carcass.

[0040] Additionally, preferably, the carcass length by position may include 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.

[0041] In addition, preferably, the step of calculating the conductor length by location can be calculated using the following formula.

[0042]

[0043]

[0044]

[0045]

[0046]

[0047] Here, W alenth , W blenth , W clenth , W dlenth , W elenth Conductor length by 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. Effects of the invention

[0048] According to the present invention, by providing an algorithm for analyzing carcasses suitable for domestic demand conditions, it has the effect of providing a low-cost, high-efficiency pig carcass grading and yield quality analysis system and method.

[0049] In addition, since the software can be easily modified according to the conditions of slaughterhouses and farms, it provides convenience from the user's perspective.

[0050] 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 below. Brief explanation of the drawing

[0051] FIG. 1 is a block diagram schematically illustrating a pig carcass grading and yield quality analysis system according to a preferred embodiment of the present invention, FIG. 2 is a block diagram of an image analysis device shown in FIG. 1, and FIG. 3a to 3i are diagrams for explaining the process of analyzing a pig carcass using the image analysis device shown in FIG. 2. FIG. 4 is a flowchart for explaining a pig carcass grading and yield quality analysis method according to a preferred embodiment of the present invention. Specific details for implementing the invention

[0052] The above objects, other objects, features, and advantages of the present invention will be easily understood through the following preferred embodiments associated with the accompanying drawings. However, the present invention is not limited to the embodiments described herein and may be embodied in other forms. Rather, the embodiments introduced herein are provided to ensure that the disclosed content is thorough and complete, and to ensure that the spirit of the present invention is sufficiently conveyed to those skilled in the art.

[0053] In this specification, when a component is described as being on another component, it means that it may be formed directly on the other component or that a third component may be interposed between them. Additionally, in the drawings, the thicknesses of the components are exaggerated for the effective description of the technical content.

[0054] Where terms such as "first," "second," etc. are used in this specification to describe components, these components shall not be limited by such terms. These terms are used merely to distinguish one component from another. The embodiments described and illustrated herein also include complementary embodiments.

[0055] Furthermore, when it is stated that the first element (or component) operates or is executed on (ON) the second element (or component), it should be understood that the first element (or component) operates or is executed in the environment where the second element (or component) operates or is executed, or operates or is executed through direct or indirect interaction with the second element (or component).

[0056] Where any element, component, device, or system is described as including a component consisting of a program or software, it should be understood that, even without explicit mention, that element, component, device, or system includes hardware (e.g., memory, CPU, etc.) or other programs or software (e.g., an operating system or drivers required to run the hardware) necessary for the execution or operation of that program or software.

[0057] Furthermore, unless otherwise specified regarding the implementation of any element (or component), it should be understood that the element (or component) may be implemented in software, hardware, or in any form that is both software and hardware.

[0058] Furthermore, the terms used herein are for the purpose of describing embodiments and are not intended to limit the invention. In this specification, the singular form includes the plural form unless specifically stated otherwise in the text. As used in this specification, 'comprises' and / or 'comprising' do not exclude the presence or addition of one or more other components to the mentioned components.

[0060] FIG. 1 is a block diagram schematically illustrating a pig carcass grading and yield quality analysis system according to a preferred embodiment of the present invention.

[0061] Referring to FIG. 1, a pig carcass grading and yield quality analysis system (10) according to a preferred embodiment of the present invention includes a carcass moving device (100), a guide device (200), an infrared light sensor (300), a camera (400), and an image analysis device (500).

[0062] The conductor moving device (100) includes a blue board and is used to move the conductor, including a Type A gamble (not shown), to each device required for analyzing the pig conductor.

[0063] The guide device (200) guides the conductor, which is moved by the conductor moving device (100), to move to an accurate point, and can guide the conductor to a point where a blue board is installed.

[0064] The infrared light sensor (300) is installed adjacent to the blue board and is a reflective sensor that utilizes the property of light reflection using infrared light to recognize a conductor that has moved to the blue board.

[0065] The camera (400) is installed facing the blue board, and when a conductor is detected by the infrared light sensor (300), it photographs the conductor on the blue board to generate a conductor image. At this time, the camera (400) is a camera that generates a color image.

[0066] The image analysis device (500) determines the gender of the conductor using the conductor image generated by the camera (400) and labels the parts of the conductor by determining them using an AI (Artificial Intelligence) model. The results analyzed by the image analysis device (500) are stored in a database and used to retrain the AI ​​model to improve the performance of the AI ​​model.

[0067] Figure 2 is a block diagram of the image analysis device illustrated in Figure 1.

[0068] Referring to FIG. 2, an image analysis device (500) according to a preferred embodiment of the present invention includes a gender determination part tracking unit (510), a cropping unit (520), a gender learning unit (530), a vertebral segment tracking unit (540), a labeling unit (550), a storage unit (560), a conductor width calculation unit (570), a conductor length calculation unit (580), and a control unit (590).

[0069] The gender determination part tracking unit (510) tracks the gender determination part from the conductor image. If the conductor is accurately suspended from the A-type gamble, the gender determination part is located without deviating significantly from the preset area, and the gender determination part tracking unit (510) tracks that location.

[0070] The cropping unit (520) crops the tracked area when the gender determination area is tracked by the gender determination area tracking unit (510).

[0071] The gender learning unit (530) learns a gender reading model through an image of an area cropped by the cropping unit (520). As the learning operation of the gender learning unit (530) is repeated, the accuracy of gender reading improves.

[0072] The vertebral segment tracking unit (540) tracks the vertebral segments from the conductor image.

[0073] The labeling unit (550) individually labels the vertebrae tracked by the vertebra tracking unit (540).

[0074] The storage unit (560) stores all information necessary for the operation of the image analysis device (500). For example, it stores the coordinates of the vertebral segments labeled by the labeling unit (550) in a database. The storage unit (560) includes a database.

[0075] The carcass width calculation unit (570) calculates coordinates corresponding to the midpoint between the lumbar and thoracic vertebrae using the coordinates of the vertebral segments, and calculates the carcass width by position using the calculated coordinates. Here, the carcass width by position includes the carcass width including the fat of the 4th and 5th thoracic vertebrae, 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.

[0076] In addition, the conductor width calculation unit (570) can calculate the conductor width for each position using the following mathematical formula 1.

[0077]

[0078] delete

[0079] delete

[0080] W in mathematical formula 1 adepth , W bdepth , W cdepth is the conductor width by location, Length a The carcass width including the fat of the 4th and 5th thoracic vertebrae, Length b is the carcass width excluding fat at the 4th and 5th thoracic vertebrae, Length c is the width of the body including fat of the 4th and 5th lumbar vertebrae, and rate is the ratio between the actual body and the body image.

[0081] The carcass length calculation unit (580) calculates the carcass length by position using the coordinates of the vertebral segments and the coordinates of the carcass's lumbar vertebrae. Here, the carcass length by position includes the length from the upper lumbar vertebra to the 1st cervical vertebra, the length from the lower lumbar vertebra to the 1st cervical vertebra, the length from the 7th lumbar vertebra to the upper thoracic vertebra, the length from the upper lumbar vertebra to the 1st thoracic vertebra, and the length from the lower lumbar vertebra to the 1st thoracic vertebra.

[0082] The conductor length calculation unit (580) can calculate the conductor length for each location using the following mathematical formula 2.

[0083]

[0084] delete

[0085] delete

[0086] delete

[0087] delete

[0088] W in mathematical equation 2 alenth , W blenth , W clenth , W dlenth , W elenth Conductor length by 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.

[0089] The control unit (590) controls all functions of the image analysis device (500). That is, the control unit (590) controls signal input and output between the gender determination part tracking unit (510), the cropping unit (520), the gender learning unit (530), the vertebral segment tracking unit (540), the labeling unit (550), the storage unit (560), the conductor width calculation unit (570), and the conductor length calculation unit (580).

[0090] FIGS. 3a to 3i are drawings for explaining the process of analyzing a pig carcass using the image analysis device shown in FIG. 2.

[0091] FIG. 3a shows the installation state of the blue board (B). The conductor is moved through the conductor moving device (100) and positioned on the front of the blue board (B). Although not shown, a camera is installed on the front of the blue board (B), and multiple lights may be installed at specific locations to ensure accuracy in capturing images by the camera. Here, LED tubes may be used for the lights.

[0092] Figure 3b shows photographs of the vertebral segments and back fat regions of the carcass. Here, the average back fat thickness extends from between the 1st lumbar vertebra and the 14th thoracic vertebra to between the 11th and 12th thoracic vertebrae. Additionally, the fat thickness of the multifidus muscle refers to the fat thickness in the center of the multifidus muscle.

[0093] Figure 3c shows a photograph of the area for measuring the carcass width and carcass length. Here, carcass width a is the carcass width including the fat of the 4th and 5th thoracic vertebrae, carcass width b is the carcass width excluding the fat of the 4th and 5th thoracic vertebrae, and carcass width c is the carcass width including the fat of the 4th and 5th lumbar vertebrae. Additionally, 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.

[0094] FIG. 3d shows a photograph with each part marked to train an AI model. Each part is marked directly within the conductor image, and the AI ​​model undergoes continuous learning through the conductor image; as learning is repeated, the analysis performance of the conductor image improves. In this embodiment, a mechanical rod for fixing the conductor is included within the conductor image, and the AI ​​model can detect the position of the rod and delete it as it is an unnecessary image.

[0095] FIG. 3e shows a comparison of two types of conductor images for gender determination. To determine gender using conductor images, the gender determination area tracking unit (510) tracks the gender determination area, and the cropping unit (520) crops the area corresponding to the gender determination area. Subsequently, the gender learning unit (530) trains a gender determination model using the image of the cropped area. At this time, the gender determination model may apply the VGG-19 algorithm. Images classified by the gender determination model are stored in a database.

[0096] The area (A1) corresponding to the sex determination area in (a) shows female genitalia, and the area (A2) corresponding to the sex determination area in (b) shows a castrated state in which male genitalia were present.

[0097] FIG. 3f is a photograph showing the state in which the vertebrae of a conductor are extracted, and coordinates are assigned to each vertebra. The number of vertebrae varies depending on the conductor, ranging from 26 to 28. Therefore, the vertebra tracking unit (540) can track the vertebrae from the conductor image to determine the number of vertebrae, and using this result, the labeling unit (550) individually labels the vertebrae and assigns coordinates.

[0098] Figure 3g shows the result of tracking the isofat region of a carcass. 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.

[0099] Figure 3h shows the result of detecting the entire conductor area, and the coordinates corresponding to the entire conductor can be stored in a database. The conductor width can be calculated by detecting the entire conductor. The operation of detecting the entire conductor area can be performed by the Yolact algorithm.

[0100] Figure 3i shows the results of the segmentation processing of the back fat region by vertebral segment. The back fat is divided into 26 to 28 masking regions corresponding to the area of ​​each vertebral segment, and the number of all vertical pixels in the segmented back fat masking regions is measured to calculate the average number of pixels. 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 point of each vertebral segment is calculated to measure the carcass width.

[0101] As described above, by using the pig carcass grading and yield quality analysis system (10) according to the present invention, each part and size of the carcass can be accurately analyzed and tracked through images of the carcass.

[0102] To deliver these results to users, a dedicated application (or app) may be provided, and the following functions may be provided to users through the dedicated application. The functions that can be provided to users through the dedicated application are summarized in Table 1.

[0103] division designation Detailed description 1 Registration of actual measurement standards for back fat thickness Registered rules for constructing the location for measuring back fat thickness by total spine number and the formula for calculating the average back fat thickness. 2 Select applicable slaughterhouse Select a slaughterhouse that applies backfat thickness standards; however, prior registration must be completed on the customer slaughterhouse registration screen. 3 Actual measurement of back fat thickness by carcass number The corresponding list is displayed based on the search criteria of slaughterhouse and slaughter date, and the actual location and value of backfat thickness are displayed depending on the selection. 4 Width and length measurement by conductor number Verify the actual measured values ​​of conductor width and length by conductor specification along with conductor photos. 5 Real-time back fat thickness measurement screen by carcass number Using the function, photos of carcasses are taken by carcass number, and back fat thickness is measured at each spinal location; the measurement location and value are automatically displayed on the on-site PC along with gender. 6 Meat quantity measurement by carcass number Retrieves all predicted items and actual backfat thickness locations / values ​​based on slaughterhouse / farm names, converts the entire file to an Excel file, and prints it. 7 52 pieces of lean meat measured per carcass number View the predicted meat yield for the selected carcass number in the measurement item prediction list. 8 Batch adjustment of average back fat thickness values A function that uniformly adjusts the average back fat thickness based on the appraiser's physical measurement data to minimize the error between the appraiser's physical measurement data and the AI ​​measurement data. 9 Registration of 1st Grade Assessment Criteria Register the grading standards of the Korea Animal Production and Grading Service, and determine the initial grade of pig carcasses based on these standards 10 Data Verification and Distribution Status by Slaughterhouse Confirmation of data verification based on RSD and R2 criteria 11 Check slaughterhouse operation status If it is necessary to check the system operation and error status for each slaughterhouse, identify failures by checking the error log files of the relevant slaughterhouse. 12 Slaughterhouse Registration Management Registration of information on client slaughterhouses to apply the system 13 Analysis results of measurement items by shipping farm It is configured to allow checking slaughter status information by shipping farm and period, and to view 52 measurement items and analysis information for the corresponding slaughter number. 14 Dashboard As an initial screen feature, it provides the ability to display the current slaughter status information of all slaughterhouses in graphs and tables. 15 User Management User registration processing for each slaughterhouse 16 User Login Login screen, restrictions on slaughterhouses accessible per ID 17 Data Import Monitoring Retrieval of slaughterhouse carcass image shooting history and receiving interface information on the carcass on the back

[0104] FIG. 4 is a flowchart illustrating a method for determining pig carcass grade and analyzing yield quality according to a preferred embodiment of the invention.

[0105] delete

[0106] In order to analyze a carcass using the pig carcass grading and yield quality analysis system (10) according to the present invention, the pig carcass to be analyzed is fixed to the carcass moving device (100) and the guide device (200) and moved to a predetermined location (S600).

[0107] When the conductor is guided by the guide device (200) to the point where the blue board is installed, the presence of the conductor is detected by the infrared light sensor (300) (S610). When the conductor is detected by the infrared light sensor (300), the control unit (590) controls the camera (400) to photograph the conductor.

[0108] The camera (400) photographs the conductor on the blue board when the conductor is positioned opposite the blue board (S620). The camera (400) photographs the conductor in color, thereby generating an image of the conductor.

[0109] The gender determination area tracking unit (510) tracks the gender determination area using an image captured by the camera (400), the cropping unit (520) crops the area corresponding to the gender determination area, and the gender is determined by a gender reading model (S640).

[0110] Afterwards, each part is identified and coordinates are assigned using the conductor image in the same manner as described in FIGS. 3a to 3i. The identified and labeled information is stored in a database (S650).

[0111] Through this process, an algorithm is provided that can analyze a conductor by capturing a color image without the need to capture the conductor image multiple times.

[0112] Those skilled in the art to which the present invention pertains will understand that the present invention may be implemented in other specific forms without altering its technical concept or essential features. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. The scope of the present invention is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and equivalent concepts should be interpreted as being included within the scope of the present invention. Explanation of the symbols

[0113] 10: Pig carcass grading and yield quality analysis system 100: Carcass moving device 200: Guide device 300: Infrared light sensor 400: Camera 500: Image analysis device

Claims

Claim 1 A conductor moving device equipped with a blue board; a guide device for guiding the movement of the conductor to a point where the blue board is installed; an infrared light sensor installed adjacent to the blue board to recognize the conductor; a camera installed opposite the blue board and, when the conductor is recognized, photographs the conductor on the blue board to generate a conductor image; and an image analysis device that determines the gender and part of the conductor and labels it using the generated conductor image; wherein the image analysis device comprises: a spinal segment tracking unit that tracks spinal segments from the conductor image; a labeling unit that individually labels the tracked spinal segments; a storage unit that stores the coordinates of the labeled spinal 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 spinal segment coordinates and calculates the conductor width for each location using the calculated coordinates. A pig carcass grading and yield quality analysis system comprising: a carcass length calculation unit that calculates the carcass length at each position using the coordinates of the stored vertebrae and the coordinates of the carcass's semi-deformed bones; wherein the carcass width calculation unit calculates the carcass width at each position using the following mathematical formula 1, and the carcass length calculation unit calculates the carcass length at each position using the following mathematical formula 2. [Mathematical Formula 1] Here, W adepth , W bdepth , W cdepth is the conductor width by location, Length a The carcass width including the fat of the 4th and 5th thoracic vertebrae, Length b is the carcass width excluding fat at the 4th and 5th thoracic vertebrae, Length c ε is the width of the body including fat of the 4th and 5th lumbar vertebrae, and rate is the ratio between the actual body and the body image.[Equation 2] Here, W alenth , W blenth , W clenth , W dlenth , W elenth Conductor length by 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. Claim 2 A pig carcass grading and yield quality analysis system according to claim 1, wherein the image analysis device comprises: a gender determination area tracking unit that tracks a gender determination area from the carcass image; a cropping unit that crops an area corresponding to the tracked gender determination area; and a gender learning unit that learns a gender reading model through an image of the cropped area. Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 delete Claim 7 delete Claim 8 delete Claim 9 delete Claim 10 A method for determining conductor grade and analyzing yield quality applied to a system comprising a conductor moving device equipped with a blue board and a camera that generates a conductor image of a conductor moved by said conductor moving device, comprising: a step of recognizing a conductor moved to a point where said blue board is installed; a step of photographing the conductor on said blue board to generate a conductor image when said conductor is recognized; a step of determining the gender of said conductor using said conductor image; and a step of determining and labeling a part of said conductor; wherein the labeling step comprises: a step of tracking a vertebral segment from said conductor image; a step of individually labeling said tracked vertebral segments; a step of storing said coordinates of said labeled vertebral segments in a database; a step of calculating a coordinate corresponding to the midpoint between the lumbar and thoracic vertebrae using said stored vertebral segment coordinates; and a step of calculating a conductor width by position using said calculated coordinates. A method for determining pig carcass grade and yield quality analysis, comprising: a step of calculating the carcass length at each position using the coordinates of the stored vertebrae and the semi-bone coordinates of the carcass; wherein the step of calculating the carcass width at each position is calculated using the following mathematical formula 1, and the step of calculating the carcass length at each position is calculated using the following mathematical formula 2. [Mathematical Formula 1] Here, W adepth , W bdepth , W cdepth is the conductor width by location, Length a The carcass width including the fat of the 4th and 5th thoracic vertebrae, Length b is the carcass width excluding fat at the 4th and 5th thoracic vertebrae, Length c ε is the width of the body including fat of the 4th and 5th lumbar vertebrae, and rate is the ratio between the actual body and the body image.[Equation 2] Here, W alenth , W blenth , W clenth , W dlenth , W elenth Conductor length by 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. h The length rate from the lower end of the pelvis to the 1st thoracic vertebra is the ratio between the actual body and the body image. Claim 11 A method for determining pig carcass grade and yield quality analysis, wherein the determining step comprises: a step of tracking a sex determination area from the carcass image; a step of cropping an area corresponding to the tracked sex determination area; and a step of training a sex reading model through the image of the cropped area. Claim 12 delete Claim 13 delete Claim 14 delete Claim 15 delete Claim 16 delete Claim 17 delete Claim 18 delete

Images