Method and device for extracting viscera of carcass based on ai

The robot system uses 3D vision and AI to automate the evisceration process, ensuring precise and hygienic internal organ extraction and sanitation, addressing the limitations of manual labor in slaughtering processes.

KR103045271B1Active Publication Date: 2026-11-10ROBOS CO LTD
View PDF 15 Cites 0 Cited by

Patent Information

Application Number
KR1020250136975
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Priority Date
2025-07-11
Filing Date
2025-09-23
Publication Date
2026-11-10
Estimated Expiration
2045-09-23

AI Technical Summary

Technical Problem

Conventional slaughtering processes for large livestock, such as pigs, rely heavily on manual labor, leading to inconsistencies in incision precision, potential carcass contamination, and challenges in meeting stringent hygiene regulations due to the lack of an integrated system capable of autonomously performing the entire evisceration process, recognizing the complex internal structure of carcasses in real time.

Method used

A robot system utilizing 3D vision information and AI to identify anatomical singularities, perform precise internal organ extraction, and implement automated hygiene processing, including weight verification and sanitation, ensuring high precision and consistency.

Benefits of technology

The system achieves significantly higher precision and consistency in internal organ extraction, prevents contamination, and meets stringent hygiene standards by using a closed-loop control system and automated hygiene processes, thereby improving product yield and safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 112025108662790-PAT00002_ABST
    Figure 112025108662790-PAT00002_ABST
Patent Text Reader

Abstract

According to one embodiment, a method and device for extracting internal organs of a conductor based on AI are disclosed, comprising the steps of: acquiring three-dimensional shape information of a conductor through a three-dimensional vision sensor; inferring location information of first and second anatomical singularities for internal organ extraction, respectively, based on the three-dimensional shape information; generating work path information of a robot to sequentially extract the first internal organ and the second internal organ based on the inferred location information; and operating the robot to extract the internal organs according to the generated work path information.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present invention relates to an AI-based slaughter automation system, and more specifically, to a robot system and a method of operation thereof that automatically performs the process of extracting internal organs and hygienic processing of a carcass using three-dimensional vision information and artificial intelligence (AI). Background Technology

[0003] With the recent application of artificial intelligence (AI) and robotics technologies across various industrial sectors, the automation of processes that traditionally relied on manual labor is accelerating. In particular, the food manufacturing and processing sector, and specifically the slaughtering industry, is seeing a continuously increasing need for automation in terms of hygiene, safety, and efficiency.

[0004] Conventional slaughtering processes, specifically the evisceration of large livestock such as pigs, have relied entirely on the physical labor of highly skilled workers. This manual method had inherent limitations, such as product quality degradation or carcass contamination caused by internal organ rupture, as the precision of the incision location varied depending on the worker's skill level or fatigue. Furthermore, the manual cleaning and disinfection of work tools made it difficult to completely eliminate the risk of cross-contamination, making it challenging to meet increasingly stringent food hygiene regulations (such as HACCP). Although some automation technologies have been proposed, they were mainly limited to simple, repetitive partial processes such as perianal incisions or the removal of fat from specific areas; consequently, there was a lack of an integrated system capable of autonomously performing the entire evisceration process by recognizing the complex internal structure of carcasses, which vary in shape and size from individual to individual, in real time. Prior art literature

[0006] U.S. Patent Publication No. 4,087,886 (May 2, 1978) Automatic evisceration process for animal carcasses The problem to be solved

[0007] The present disclosure aims to solve the aforementioned problems by combining 3D vision information and AI technology to recognize the complex internal anatomical structure of a conductor in real time, and to provide a method and device that automatically perform the entire process from internal organ extraction to hygienic processing in a precise and hygienic manner through a robotic system.

[0008] The problems to be solved in this disclosure are not limited to those mentioned above, and other unmentioned problems will be clearly understood by a person skilled in the art from the description below. means of solving the problem

[0010] As a technical means for achieving the technical problem described above, a method for extracting internal organs of a conductor based on AI according to the first aspect of the present disclosure may include: a step of acquiring three-dimensional shape information of the conductor through a three-dimensional vision sensor; a step of inferring, based on the three-dimensional shape information, location information of a first anatomical singularity for extracting internal organs of the conductor and a second anatomical singularity different from the first anatomical singularity, respectively, by an AI model; a step of generating work path information of a robot to extract a first internal organ based on the location information of the first anatomical singularity and sequentially extract a second internal organ based on the location information of the second anatomical singularity; and a step of the robot operating according to the work path information to extract internal organs including the first internal organ and the second internal organ from the conductor.

[0011] Additionally, the first anatomical peculiarity may include at least one of the rectum or perianal region of the conductor.

[0012] In addition, the second anatomical peculiarity may include a region connecting the liver and the heart of the conductor.

[0013] Additionally, the first arm of the robot may grasp the first anatomical singularity, and the second arm of the robot may incise the first fascia exposed by the grasping motion of the first arm to extract the first internal organ, and subsequently, the first arm of the robot may grasp the second anatomical singularity, and the second arm of the robot may incise the exposed second fascia.

[0014] In addition, the first internal organ may include a cataract, and the second internal organ may include a red internal organ.

[0015] In addition, weight information can be obtained from an internal organ receiving module that accommodates the extracted internal organ, and the extraction of the internal organ can be determined based on the weight information.

[0016] In addition, the above weight information can determine whether the extraction of the internal organs is completed based on a preset threshold weight value.

[0017] In addition, while the first arm of the robot grasps and pulls the first anatomical singularity, the internal support module can support the first internal organ being pulled to reduce the load on the robot.

[0018] In addition, automatic hygiene treatment can be performed on at least one of the working tool of the robot and the lens of the vision sensor used for the internal organ extraction.

[0019] In addition, the above automatic hygiene treatment step may include a step of spraying hot water above a preset temperature, a step of cleaning using ultrasound, and a step of drying through air blowing.

[0020] In addition, if it is determined that the extraction of the internal organs is not completed, the robot can be controlled to perform the work path generation and extraction steps again.

[0021] Additionally, a device according to a second aspect of the present disclosure may include: a receiver that receives three-dimensional shape information of the conductor through a three-dimensional vision sensor; and a processor that, based on the three-dimensional shape information, infers location information of first and second anatomical singularities for internal organ extraction of the conductor, respectively, generates work path information of a robot to sequentially extract the first internal organ and the second internal organ based on the inferred location information, and controls the robot to operate according to the generated work path information.

[0022] Additionally, a third aspect of the present disclosure may provide a computer-readable recording medium having a program for executing the method of the first aspect on a computer. Effects of the invention

[0024] According to one embodiment of the present disclosure, by using a 3D vision sensor and an AI model to precisely infer the anatomical peculiarities of the body that differ from one body to another, and by having a robot perform the internal organ extraction operation based on this, significantly higher precision and consistency compared to manual work can be secured. In particular, through a unique process of sequentially extracting cataracts and red internal organs using different reference points, contamination of the body caused by internal organ rupture is fundamentally prevented, thereby having the effect of dramatically improving the yield and quality of the final product.

[0025] In addition, by implementing a closed-loop control system that physically verifies the success of the extraction operation through a load cell equipped in the built-in receiving module, process reliability is maximized and fully unmanned automated operation is enabled.

[0026] Furthermore, by performing an automated hygienic processing process including high-temperature water, ultrasound, and air blowing that meets HACCP standards in every work cycle, it provides the effect of blocking the risk of cross-contamination and ensuring the highest level of food safety. Brief explanation of the drawing

[0028] FIG. 1 is a block diagram schematically illustrating the configuration of a device for extracting the internal structure of a conductor based on AI according to one embodiment. FIG. 2 is a flowchart illustrating each step of a method for extracting internal organs of a conductor based on AI according to one embodiment. FIG. 3 is a drawing specifically illustrating the configuration of a dual-arm robot system according to one embodiment. FIG. 4 is a drawing for explaining the operation of an internal receiver module according to one embodiment. FIG. 5 is a drawing for explaining the configuration of an automatic hygiene treatment system according to one embodiment. Specific details for implementing the invention

[0029] The advantages and features of the present disclosure and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present disclosure is not limited to the embodiments disclosed below but can be implemented in various different forms, and the embodiments provided are merely to make the disclosure complete and to fully inform those skilled in the art of the scope of the present disclosure.

[0030] The terms used in this specification are for describing embodiments and are not intended to limit the disclosure. In this specification, the singular form includes the plural form unless specifically stated otherwise in the text. The terms “comprises” and / or “comprising” as used in this specification do not exclude the presence or addition of one or more other components in addition to the components mentioned. Throughout the specification, the same reference numerals refer to the same components, and “and / or” includes each of the mentioned components and all combinations of one or more. Although terms such as “first,” “second,” etc., are used to describe various components, these components are not limited by these terms. These terms are used merely to distinguish one component from another. Accordingly, the first component mentioned below may be the second component within the technical scope of this disclosure.

[0031] Unless otherwise defined, all terms used herein (including technical and scientific terms) may be used in a sense that is commonly understood by a person skilled in the art. Additionally, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise. Spatially relative terms such as "below," "beneath," "lower," "above," and "upper" may be used to easily describe the relationship between one component and other components, as illustrated in the drawings.

[0032] Spatially relative terms should be understood as terms that include different orientations of components during use or operation, in addition to the orientations depicted in the drawings. For example, if a component depicted in a drawing is inverted, a component described as "below" or "beneath" another component may be placed "above" of that other component. Therefore, the exemplary term "below" may include both the downward and upward directions. Components may also be oriented in other directions, and accordingly, spatially relative terms may be interpreted according to the orientation.

[0033] Embodiments of the present disclosure are described below with reference to the attached drawings so that those skilled in the art can easily implement them. However, the present disclosure may be embodied in various different forms and is not limited to the embodiments described herein.

[0035] FIG. 1 is a block diagram schematically illustrating the configuration of a device for extracting the internal structure of a conductor based on AI according to one embodiment.

[0036] Referring to the drawing, the device (100) may include a receiving unit (110), a processor (120), and a transmitting unit (130).

[0037] A person skilled in the art will understand that, in addition to the components shown in FIG. 1, other general-purpose components may be further included in the device (100). For example, the device (100) may further include a memory (not shown). Or, according to other embodiments, a person skilled in the art will understand that some of the components shown in FIG. 1 may be omitted.

[0038] A device (100) according to one embodiment may be a system that automates the process of extracting internal organs from a carcass based on AI. The device (100) is installed on a slaughter line and can continuously perform operations on carcasses transported via a conveyor system. The device (100) may be largely composed of parts that perform information acquisition, computation and control, and physical operations, and all components may be organically interconnected through real-time data communication.

[0039] The receiver (110) may be a component that receives various external information necessary for the operation of the invention. Specifically, the receiver (110) may receive point cloud data representing real-time three-dimensional shape information of a conductor from a three-dimensional vision sensor (e.g., LiDAR sensor, Time-of-Flight (ToF) camera, stereo camera, etc.). Additionally, the receiver (110) may receive weight information in analog or digital form from a load cell (440) provided in the built-in receiving module (400), or receive sensor signals including current state information from a plurality of sensors (e.g., temperature sensor, flow sensor) included in the automatic hygiene processing system (500).

[0040] The processor (120) is a central processing unit that controls the overall operation of the device (100) and can be implemented as one or more microprocessors or application-specific integrated circuits (ASICs). Based on three-dimensional shape information obtained from the receiver (110), the processor (120) can infer three-dimensional coordinates of anatomical singularities inside the conductor by running a pre-trained AI model (e.g., a deep learning model based on a Convolutional Neural Network (CNN)). Additionally, the processor (120) can generate an optimal work path for the robot based on the inferred coordinate information, determine whether internal organ extraction is successful by comparing weight information received from the load cell (440) with a preset threshold value, and generate signals to control valves, pumps, heaters, etc. of the hygiene processing system (500) according to each sensor value.

[0041] The transmitter (130) can perform the role of transmitting a control signal generated by the processor (120) to an external actuator. For example, the transmitter (130) can drive a two-arm robot by transmitting a control signal containing generated work path information (e.g., target angle, speed, torque of each joint) to a robot controller. Additionally, it can transmit an electrical signal to control the operation of the corresponding component, such as a drive unit (e.g., servo motor) of the built-in receiving module (400) or a solenoid valve, pump, etc. of the automatic hygiene processing system (500).

[0042] Memory (not shown) can store parameters of the AI ​​model, a bio-scan database (DB) used for training, a slaughter process sequence program, threshold setting data for various sensors, and work log data. The memory can be implemented as non-volatile or volatile memory, such as RAM, ROM, or flash memory.

[0044] FIG. 2 is a flowchart illustrating each step of a method for extracting the internal organs of a conductor based on AI using a device (100) according to one embodiment.

[0045] Referring to step S10, a device (100) according to one embodiment can acquire three-dimensional shape information of a carcass through a three-dimensional vision sensor. The three-dimensional vision sensor may be fixedly installed at a specific point on the slaughter line (e.g., before entering the robot work area) or attached to the robot arm itself. The device (100) performs a three-dimensional scan on a carcass moving along a conveyor at preset intervals (e.g., 100ms) or in response to a trigger signal, thereby acquiring high-density three-dimensional point cloud data including the size, shape, position, and orientation of the carcass. For the acquired raw data, a filtering algorithm (e.g., Statistical Outlier Removal) may be applied to remove noise occurring in the slaughter environment (e.g., mismeasurement due to steam, distortion due to blood reflection). This has the effect of improving the precision of the data so that the AI ​​model can perform more accurate inference in subsequent steps.

[0046] Referring to step S20, the device (100) according to one embodiment can, based on three-dimensional shape information, allow an AI model to infer location information of a first anatomical singularity for the extraction of internal organs of a conductor and a second anatomical singularity that is different from the first anatomical singularity. The processor (120) can provide refined three-dimensional point cloud data as input to a pre-trained AI model. The AI ​​model can be trained using a training dataset consisting of 3D scan data of tens of thousands of various objects and anatomical singularity coordinates directly labeled by an expert for each data. Based on a pattern learned of the relative positional relationships of internal organs that are not directly observed from the outside within the input 3D shape, the AI ​​model can infer the location of a first anatomical singularity (e.g., the three-dimensional center coordinate of the rectum) which serves as the starting point for cataract extraction and a second anatomical singularity (e.g., the three-dimensional center coordinate of the tissue connecting the liver and the heart) which serves as the standard for internal organ extraction, with high accuracy (e.g., within an average error of 5 mm). This has the effect of overcoming differences in body shape, size, and subtle positional variations among individuals, and always providing a consistent work reference point.

[0047] Referring to step S30, the device (100) according to one embodiment can generate robot work path information to extract a first internal organ based on location information of a first anatomical singularity and sequentially extract a second internal organ based on location information of a second anatomical singularity. The processor (120) can set the three-dimensional coordinates of the inferred first and second anatomical singularities as target points and calculate an optimal path to reach the target points from the robot's current position (start point). At this time, beyond simply generating the shortest path, a path planning algorithm (e.g., RRT, A*) to avoid collisions with other parts of the conductor or surrounding equipment may be applied. Additionally, complex and precise work path information can be generated, including the approach angle of the gripper arm for grasping the first singularity, the trajectory of the pulling motion to expose the fascia after grasping, and the movement path and speed profile of the knife arm for cutting the exposed fascia. This ensures that internal organ extraction operations requiring complex coordinated movements can be performed safely and efficiently without collisions.

[0048] Referring to step S40, the device (100) according to one embodiment can extract internal organs, including a first internal organ and a second internal organ, from a conductor by having the robot operate according to work path information. The processor (120) transmits a control command containing the generated work path information to a robot controller, and the robot can perform the internal organ extraction operation by precisely driving each joint according to the received command. First, an extraction sequence for the first internal organ (e.g., cataract) is performed, and then an extraction sequence for the second internal organ (e.g., red internal organ) can be performed sequentially. The entire work process is monitored in real time, and if an abnormal load on the robot (e.g., a situation where it gets caught on bone during incision) is detected through a torque sensor or the like, an interlock logic may be included to immediately stop the operation and switch to a safe state. This has the effect of protecting the system and the workpiece even in unexpected situations and ensuring the stability of the entire process.

[0049] For example, the device (100) can perform internal organ extraction using a dual-arm robot. Specifically, the robot's first arm (310) (e.g., a gripper arm) can precisely grasp and pull an anatomical singularity (e.g., a rectum or a liver / heart junction) inferred by AI. Through this pulling motion, the fascial tissue previously hidden inside the body can be exposed to the outside. Then, the robot's second arm (320) (e.g., a knife arm) moves along the exposed fascia and precisely cuts it, thereby completely separating the internal organ from the surrounding tissue. By dividing the roles of the two arms and coordinating them organically, it is possible to efficiently perform complex and delicate tasks, much like a skilled human worker. This allows for the automation of high-difficulty tasks that are difficult to implement with a single robot arm, while simultaneously improving the accuracy and speed of the work.

[0050] For example, the device (100) can double-check whether the internal organ extraction is successful. When the internal organ is successfully extracted and placed on the internal organ receiving module (400), a load cell (440) installed at the bottom of the module can detect the weight of the internal organ, generate weight information, and transmit it to the processor (120). The processor (120) can determine whether the received weight information is greater than or equal to a preset threshold weight value (e.g., 3 kg, which corresponds to the minimum value of normal internal organ weight). If the weight information is below the threshold, it means that the operation was performed abnormally, such as the internal organ not being completely separated or only partially extracted; therefore, the processor (120) can send a control signal to the robot to re-perform the previous grasping and cutting steps. Through this feedback-based rework logic, it is possible to raise the reliability of the process to over 99.9% by compensating for potential misrecognition by the AI ​​vision system or minute errors in robot movement with physical sensors. This has the effect of fundamentally preventing defects and maximizing the stability of unmanned operation.

[0051] For example, the device (100) can automatically manage the hygiene status of the robot's work tools and sensors. After the internal extraction operation for one conductor is completed, the robot can automatically move to a designated hygiene processing station. At the station, the device (100) first activates a heating device (520) to heat the water in the tank to a sterilization temperature (e.g., 85°C) or higher as specified in HACCP standards, and uses a high-pressure pump (540) to spray the heated water onto a gripper (315), knife (325), camera lens, etc., to remove primary contaminants such as blood or fat. Then, by activating an ultrasonic generator (530) to create and burst fine bubbles within the tank through a cavitation phenomenon, it is possible to precisely clean even foreign substances in fine crevices that are difficult to remove by physical brushing. Finally, compressed air is blown at high speed through an air nozzle to quickly remove remaining moisture and dry the tank, thereby inhibiting bacterial growth and completing preparations for the next operation. This effectively prevents the risk of cross-contamination and ensures food safety by maintaining the highest level of hygiene at all times.

[0052] For example, the device (100) can dynamically respond to a process environment that changes in real time when determining the work path information of the robot. The work path information can be dynamically determined based on various factors including [1] shape information per conductor object, [2] inference accuracy of internal organ location, [3] gripping pressure feedback of the gripper, [4] load cell weight of the internal support, and [5] contamination information of the work tool. Here, 'shape information per conductor object' may refer to 3D point cloud data containing unique geometric features (e.g., width of the thorax, length of the abdomen) of the conductor currently being worked on, obtained by scanning with a 3D vision sensor (e.g., LiDAR). 'Inference accuracy of internal organ location' may refer to a probabilistic indicator (e.g., a confidence score of 95.2%) indicating the reliability of the result after the AI ​​model predicts the location of internal anatomical peculiarities (e.g., rectum) from the shape information. 'Gripper gripping pressure feedback' refers to a value of physical pressure (e.g., 3.5 N) measured in real time by a multi-axis force / torque sensor embedded in the gripper (315) when gripping soft tissue, and 'load cell measured weight of internal organ holder' refers to a value of real-time weight (e.g., 5.2 kg) measured by a load cell (440) when the internal organ is extracted and placed on the holder, which may be key data for physically verifying the success of the actual operation. Finally, 'contamination information of the work tool' may refer to a value (e.g., less than 0.1% of the total surface area) obtained by quantifying the area of ​​residual foreign matter through image analysis after a vision camera captures the surface of the knife or gripper following the completion of the automatic sanitation process.

[0053] For example, during the slaughtering process, a high-risk situation may occur in which an AI model calculates the inference accuracy of internal organ locations to be lower than a preset threshold (e.g., 90%) when handling an irregular carcass. In such cases, the device (100) may be configured to sequentially assign weights such that they gradually decrease in the order of 1) gripping pressure feedback of the gripper, 2) load cell weight measurement of the internal organ holder, 3) inference accuracy of internal organ locations, 4) contamination information of the work tool, and 5) shape information of each carcass, in order to ensure the stability of the operation as a top priority. This is because, in situations of high uncertainty, the actual reaction of the object currently being physically interacted with by the robot (e.g., sudden change in gripping pressure) has the greatest impact as it is the most immediate and important variable in preventing internal organ rupture, rather than AI predictions or geometric information; next, it can be given the second highest weight as the final success of the operation is determined by whether the internal organs are physically separated (e.g., whether the threshold of the load cell measured weight is reached); and finally, AI prediction information or basic shape information has a relatively small impact as it plays a supporting role to the physical sensor values. Therefore, based on this priority, if the gripping pressure deviates from a preset stable range (e.g., 3N ±0.5N), the cutting speed is immediately reduced to a certain percentage (e.g., 50%) of the current speed, and if the rate of change in load cell weight does not reach a reference value (e.g., 0.5kg / s), the robot's work path information can be corrected in real time to finely readjust the gripping position (e.g., -5mm in the Z-axis direction).

[0054] As another example, in a situation where the inference accuracy of the AI ​​model is very high (e.g., 98% or higher) and process variables are stable, the device (100) can change the order of weights to maximize overall productivity. In this case, weights can be assigned sequentially so that they gradually decrease in the order of 1) shape information per conductor object, 2) inference accuracy of internal organ position, 3) load cell weight of the internal receiver, 4) gripping pressure feedback of the gripper, and 5) contamination information of the work tool. This is because, in situations where the AI's judgment is reliable, generating an optimized shortest-time path in advance based on the most basic geometric information (shape information) and the AI's prediction (inference accuracy) and performing the work at high speed has the greatest impact as it is most effective in increasing the overall throughput (e.g., more than 60 heads per hour), and subsequently, physical feedback such as load cell weight or gripping pressure is sufficient for verification and auxiliary roles to detect exceptional situations that may occur during high-speed work (e.g., work failure occurring with a probability of less than 1%), so its impact on the work path generation itself is relatively small. Therefore, based on these priorities, the robot's work path information can be determined so that the device (100) performs the work based on a standardized high-speed path, but pauses the process and regenerates the path only in exceptional situations, such as when the load cell weight does not reach the target value within a preset time (e.g., 3 seconds).

[0056] FIG. 3 shows a dual-arm robot system according to one embodiment that may include a first arm (310) and a second arm (320). An end-effector (315) in the form of a gripper may be mounted on the end of the first arm (310), and the gripper (315) may include a plurality of fingers, and micro-protrusions or an elastic material (e.g., silicone) may be attached to the surface of each finger to prevent slipping. In addition, a force / torque sensor may be embedded inside the gripper (315) to precisely control the gripping force (e.g., in units of 0.1 N) so that excessive pressure is not applied when gripping soft tissues such as the rectum. An end-effector (325) in the form of a knife may be mounted on the end of the second arm (320). The knife (325) may have a curved blade to efficiently cut fascial tissue, and may be made of medical-grade stainless steel (e.g., SUS316L) to ensure corrosion resistance and durability. The device (100) can perform highly coordinated operations by precisely synchronizing the movements of the two arms, such as the first arm (310) pulling the internal organ at a speed of 50 mm per second while the second arm (320) cuts along the exposed fascia with a positional precision of within 0.2 mm. This minimizes errors that may occur in manual work and has the effect of dramatically improving the consistency of the operation.

[0058] According to one embodiment of FIG. 4, the internal support module (400) may include a Z-axis drive unit (410), an X-axis drive unit (420), a support plate (430), and a load cell (440). The Z-axis drive unit (410) and the X-axis drive unit (420) are each composed of a high-precision ball screw and an AC servo motor, respectively, and can move the support plate (430) quickly and accurately in vertical and horizontal directions (e.g., maximum speed 1,000 mm / s, repeatable positioning precision ±0.02 mm). Before the robot begins the motion of pulling the internal organ, the device (100) can drive the internal support module (400) to position the support plate (430) directly below the expected drop point of the internal organ to be extracted. Considering the force limit of the robot arm (e.g., maximum payload of 10 kg), the support plate (430) can stably support the weight of the internal organs (e.g., 5 to 8 kg), thereby reducing the load applied to the robot and allowing the robot to concentrate its force only on pulling motions. When the internal organs are extracted and placed on the support plate (430), a plurality of load cells (440) connected to the lower part of the plate detect the load, and the summed analog signal can be converted into digital weight data through an amplifier and transmitted to the processor (120). Through this, the reliability of the entire system is maximized by adding a physical verification step beyond simple vision reading.

[0060] FIG. 5 shows that an automatic sanitation treatment system (500) according to one embodiment may include a hot water tank (510), a heating device (520), an ultrasonic generator (530), a high-pressure pump (540), and a plurality of valves and nozzles (550). The hot water tank (510) may be controlled by a water level sensor and a water supply valve to always maintain a constant water level. At the start of the process, the device (100) operates a heating device (520, e.g., a 5kW industrial heater) to heat the water in the tank to a target sterilization temperature (e.g., 85°C), and when the temperature sensor confirms that the target temperature has been reached, the heating operation can be stopped or switched to a maintenance (warming) mode. When the robot moves the work tool to the cleaning position, the device (100) can perform sequential control by first driving a high-pressure pump (540) to spray heated hot water through a nozzle (550), then driving an ultrasonic generator (530) to perform cleaning for about 10 seconds, and finally opening the valve of the air supply line to complete drying by air blowing. The flow sensor and pressure sensor continuously monitor whether the fluid line is operating normally throughout the entire process, and can stop the process immediately and generate an alarm if an abnormality occurs. This has the effect of ensuring consistent and verified sanitation results at all times without human intervention.

[0062] Various embodiments of the present disclosure may be implemented as software comprising one or more instructions stored in a storage medium (e.g., memory) readable by a machine (e.g., a display device or a computer). For example, a processor of the machine (e.g., processor (120)) may call at least one of the one or more instructions stored from the storage medium and execute it. This enables the machine to be operated to perform at least one function according to the at least one called instruction. The one or more instructions may include code generated by a compiler or code that can be executed by an interpreter. The storage medium readable by the machine may be provided in the form of a non-transitory storage medium. Here, 'non-temporary' merely means that the storage medium is a tangible device and does not contain a signal (e.g., electromagnetic waves), and this term does not distinguish between cases where data is stored semi-permanently and cases where it is stored temporarily.

[0063] According to one embodiment, the method according to the various embodiments disclosed herein may be provided by being included in a computer program product. The computer program product may be traded between a seller and a buyer as a product. The computer program product may be distributed in the form of a device-readable storage medium (e.g., compact disc read-only memory (CD-ROM)), or distributed online (e.g., download or upload) through an application store (e.g., Play Store™) or directly between two user devices (e.g., smartphones). In the case of online distribution, at least a portion of the computer program product may be temporarily stored or temporarily created in a device-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.

[0064] Those skilled in the art related to the embodiments described above will understand that they may be implemented in modified forms without departing from the essential characteristics of the description. Therefore, the disclosed methods should be considered in an illustrative rather than a restrictive sense. The scope of the invention is defined by the claims, not by the foregoing description, and all variations within the scope of equivalence should be interpreted as being included in the invention. Explanation of the symbols

[0066] 100: Device 110: Receiver 120: Processor 130: Transmitter 310: 1st Arm 315: Gripper 320: 2nd Arm 325: Knife 400: Built-in receiving module 410: Z-axis drive unit 420: X-axis drive unit 430: Support plate 440: Load cell 500: Automatic Sanitation System 510: Hot water tank 520: Heating device 530: Ultrasonic generator 540: High-pressure pump 550: Valves and Nozzles

Claims

Claim 1 A method for extracting internal organs of a tube based on AI, comprising: (a) acquiring three-dimensional shape information of the tube through a three-dimensional vision sensor; (b) based on the three-dimensional shape information, a step in which an AI model infers location information of a first anatomical singularity for extracting internal organs of the tube and a second anatomical singularity different from the first anatomical singularity, respectively; (c) generating work path information of a robot to extract a first internal organ based on the location information of the first anatomical singularity and sequentially extract a second internal organ based on the location information of the second anatomical singularity; and (d) the robot operating according to the work path information to extract internal organs including the first internal organ and the second internal organ from the tube, wherein the first anatomical singularity includes at least one of the area around the rectum or anus of the tube. Claim 2 delete Claim 3 A method according to claim 1, wherein the second anatomical peculiarity comprises a region connecting the liver and the heart of the conductor. Claim 4 A method according to claim 1, wherein step (d) comprises: a step in which the first arm of the robot grasps the first anatomical singularity; a step in which the second arm of the robot incises the first fascia connected to the first internal organ, which is exposed by the grasping motion of the first arm; and a step in which the first internal organ is extracted from which the first fascia has been incised. Claim 5 A method according to claim 4, wherein step (d) further comprises: a step in which, after the first internal organ is extracted, the first arm of the robot grasps the second anatomical singularity; and a step in which the second arm of the robot incises the second fascia connected to the second internal organ, which is exposed by the grasping motion of the first arm. Claim 6 A method according to claim 5, characterized in that the first internal organ includes a cataract and the second internal organ includes a red internal organ. Claim 7 A method according to claim 1, further comprising: a step of obtaining weight information from an internal organ receiving module that accommodates the extracted internal organ; and a step of determining whether the extraction of the internal organ is complete based on the weight information. Claim 8 A method according to claim 7, wherein the determining step comprises: a step of determining that the extraction of the internal organs is not completed if the weight information is less than a preset threshold weight value; and a step of determining that the extraction of the internal organs is completed if the weight information is greater than or equal to the threshold weight value. Claim 9 A method according to claim 4, wherein step (d) further comprises the step of reducing the load on the robot by having the internal support module support the first internal organ being pulled while the first arm of the robot grasps and pulls the first anatomical singularity. Claim 10 A method according to claim 1, further comprising an automatic sanitation step of performing automatic sanitation on at least one of the working tool of the robot and the lens of the vision sensor used for internal organ extraction. Claim 11 A method according to claim 10, wherein the above automatic sanitation treatment step comprises: a step of spraying hot water at a preset temperature (e.g., 85℃) or higher; a step of cleaning using ultrasound; and a step of drying through air blowing. Claim 12 A method according to claim 7, further comprising the step of controlling the robot to perform steps (c) and (d) again if it is determined in the above-determining step that the extraction of the internal organs is not completed. Claim 13 A method according to claim 1, characterized by determining work path information of the robot based on different weights assigned to each of the conductor shape information, internal organ location inference accuracy, gripping pressure feedback of the gripper, load cell measured weight of the internal support, and contamination information of the work tool. Claim 14 A device for extracting internal organs of a conductor based on AI, comprising: a receiver that receives three-dimensional shape information of the conductor through a three-dimensional vision sensor; and a processor that, based on the three-dimensional shape information, infers location information of a first anatomical singularity and a second anatomical singularity different from the first anatomical singularity for extracting internal organs of the conductor, respectively, generates work path information of a robot to extract a first internal organ based on the location information of the first anatomical singularity and sequentially extract a second internal organ based on the location information of the second anatomical singularity, and controls the robot to operate according to the work path information to extract internal organs including the first internal organ and the second internal organ from the conductor, wherein the first anatomical singularity includes at least one of a region around the rectum or anus of the conductor. Claim 15 A computer-readable recording medium having a program for executing the method of any one of paragraphs 1 and 3 through 13 on a computer.

Citation Information

Patent Citations

  • Automatic eviscerating system of poultry robot

    CN111685157A

  • A robot intelligent slaughtering and cutting method, device and storage medium

    CN114494295B

  • Positioning detection method for domestic animal bore opening

    CN115994905A

  • Pig slaughtering and segmenting device

    CN210054447U

  • Viscera recycling and conveying device for beef cattle slaughtering

    CN212650334U