Production quality checking system and production quality checking method

Abstract

Introduced is a production quality checking system comprising: an unmanned robot for extracting a two-dimensional image of a component in a state of assembly through an image sensor unit; a digital twin system in which a three-dimensional image of the component completely assembled is stored, and from which a two-dimensional image of the component in the state of assembly is extracted from the three-dimensional image; and a check unit for checking the state of assembly of the component by comparing the extracted two-dimensional images.

Description

Production quality assurance system and production quality assurance method

[0001] The present invention relates to a production quality verification system and a production quality verification method.

[0002] Most products are completed by fastening various parts together. If these parts are not properly fastened, not only will the product not achieve the desired effect, but it can also generate vibrations and noise, causing discomfort to the user and negatively impacting the product's durability.

[0003] In particular, the importance of part fastening increases for products with a large number of parts, and as production automation increases, the importance of part fastening verification work also increases.

[0004] Accordingly, the need for a production quality assurance system that can check the status of component fastening is increasing.

[0005] The present invention aims to provide a production quality confirmation system that reduces the man-hours and cost required for confirming the fastening status of parts in an automated production process by confirming the fastening status of parts using an unmanned robot, and, in some cases, determines whether correction of an image sensor unit performing a photographing function is necessary, thereby increasing the reliability of confirming the fastening status of parts.

[0006] The production quality confirmation system according to the present invention includes an unmanned robot that extracts a two-dimensional image of a part in an assembled state through an image sensor unit; a digital twin system that stores a three-dimensional image of a part in which assembly has been completed and extracts a two-dimensional image of the part in an assembled state from the three-dimensional image; and a check unit that compares the extracted two-dimensional images to confirm the assembly state of the part.

[0007] The image sensor section is composed of a multi-joint mechanism capable of rotation or extension, enabling the unmanned robot to extract two-dimensional images of assembled parts from multiple viewpoints.

[0008] The two-dimensional image extracted by the unmanned robot may include location information and movement information of the image sensor unit where the unmanned robot extracted the two-dimensional image.

[0009] The digital twin system extracts a two-dimensional image of an assembled part from a three-dimensional image, and can extract the two-dimensional image of an assembled part at the time when the unmanned robot extracts the assembled part.

[0010] The check section can determine the similarity between two-dimensional images through the histogram of the extracted two-dimensional images and can determine the differences between the two-dimensional images.

[0011] Based on the similarity of the 2D image, the assembly status of the part is judged as suitable or defective in the check section, and the defect of the image sensor section can be judged based on the similarity of the 2D image.

[0012] If the similarity is greater than or equal to the first value, it is judged as suitable. If the similarity is less than the first value, it is judged as defective. If the similarity is less than the first value and greater than or equal to the second value, it is judged as defective in the image sensor section. If the similarity is less than the second value, it can be judged as defective in the assembly state.

[0013] It further includes a judgment unit that compares the extracted two-dimensional images to calculate the degree of defect in the image sensor unit of the unmanned robot; and if the checking unit determines that the image sensor unit is defective, the judgment unit can calculate the degree of defect in the image sensor unit based on the visual difference between the unmanned robot and the digital twin system.

[0014] The judgment unit can extract the outline information of the extracted two-dimensional image and calculate the visual difference between the unmanned robot and the digital twin system.

[0015] An unmanned robot can extract a two-dimensional image of an assembled part through an image sensor at a location where a location information mark containing a QR code or barcode is recognized.

[0016] Unmanned robots can be equipped with cameras that recognize location information markers.

[0017]

[0018] A method for verifying production quality according to the present invention includes a step of extracting a two-dimensional image of a part in an assembled state by an unmanned robot; a step of extracting a two-dimensional image of a part in an assembled state from a three-dimensional image of a part in which assembly is completed and stored in a digital twin system; and a step of verifying the assembly state of the part by comparing the extracted two-dimensional images in a checking unit.

[0019] The step of extracting a 2D image in a digital twin system may be a step of capturing a 3D image at the same time as the time at which the unmanned robot extracted the 2D image.

[0020] The step of checking the assembly status of the parts may be a step of judging the similarity between the two-dimensional images through the histogram of the extracted two-dimensional images and judging the differences between the two-dimensional images.

[0021] The step of checking the assembly status of the parts may be a step of judging the assembly status of the parts as suitable or defective based on the similarity of the two-dimensional images, and a step of judging the defect of the image sensor part based on the similarity of the two-dimensional images.

[0022] If the similarity is greater than the first value, it is considered suitable.

[0023] If the similarity is less than the first value, it is judged as defective.

[0024] If the similarity is less than the first value or greater than the second value, it is judged as a defect in the image sensor.

[0025] If the similarity is less than the second value, it can be judged as a defective assembly.

[0026] In the step of checking the assembly status of the parts, if the image sensor part is judged to be defective, a step of comparing the extracted two-dimensional images and calculating the degree of defect in the image sensor part of the unmanned robot can be further performed.

[0027]

[0028] The step of calculating the degree of defect in the image sensor part of the unmanned robot by the judgment unit may be a step of calculating the degree of defect in the image sensor part based on the visual difference between the unmanned robot and the digital twin system.

[0029] The step of extracting a two-dimensional image by an unmanned robot may include a step of recognizing a location information mark including a QR code or barcode by the unmanned robot; and a step of extracting a two-dimensional image of an assembled part through an image sensor unit at a point where the location information mark is recognized.

[0030] According to the production quality confirmation system and production quality confirmation method of the present invention, I) an unmanned robot is used, and by comparing a two-dimensional image extracted by the unmanned robot and a master image of a digital twin system, it is possible to confirm whether or not a part is poorly fastened, so there is no need to collect images of poorly assembled cases.

[0031] II) In some cases, it is possible to determine whether the image sensor unit performing the shooting function requires correction, replace parts of the image sensor unit, or increase the reliability of checking the state of part fastening by exchanging the image sensor unit.

[0032] Figure 1 is a configuration diagram of a production quality verification system according to the present invention.

[0033] Figure 2 illustrates an unmanned robot extracting a two-dimensional image of the assembled state of a part.

[0034] Figure 3 is an enlarged view of the image sensor section of an unmanned robot.

[0035] Fig. 4 is a two-dimensional image extracted by moving the image sensor unit of the unmanned robot in the situation of Fig. 2 to capture a picture of Part A at a specific point in time. Fig. 5 shows an example of a three-dimensional image of Part A stored in the digital twin system, and Fig. 6 is a two-dimensional image extracted through a three-dimensional image of Part A stored in the digital twin system.

[0036] Figure 7 illustrates an example of using outline information to determine the degree of defect in an image sensor unit.

[0037] Figures 8 to 9 are flowcharts for one embodiment of the present invention.

[0038] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing numbers, identical or similar components are given the same reference numbers and redundant descriptions thereof will be omitted.

[0039] In describing the embodiments disclosed in this specification, detailed descriptions of related known technologies will be omitted if it is determined that such detailed descriptions may obscure the gist of the embodiments disclosed in this specification. In addition, the attached drawings are provided solely to facilitate understanding of the embodiments disclosed in this specification, and the technical concepts disclosed in this specification are not limited by the attached drawings, and should be understood to include all modifications, equivalents, and substitutes included within the spirit and technical scope of the present invention.

[0040] Terms that include ordinal numbers, such as first, second, etc., may be used to describe various components, but the components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0041] Singular expressions include plural expressions unless the context clearly indicates otherwise.

[0042] In this specification, terms such as “include” or “have” are intended to specify the presence of a feature, number, step, operation, component, part or combination thereof described in the specification, but should be understood not to exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

[0043] The suffixes "module" and "part" used for components in the following description are given or used interchangeably only for the convenience of writing specifications, and do not have distinct meanings or roles in themselves.

[0044] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0045] The digital twin system, check unit, judgment unit, ERP system, MES system, and robot control system described herein include all types of recording devices that store data readable by a computer system. Examples of computer-readable media include hard disk drives (HDDs), solid-state disk drives (SSDs), silicon disk drives (SDDs), ROMs, RAMs, CD-ROMs, magnetic tapes, floppy disks, and optical data storage devices.

[0046]

[0047] Figure 1 is a schematic diagram of a production quality assurance system according to the present invention. Referring to Figure 1, the production quality assurance system according to the present invention may be a system required for operating a smart factory.

[0048] Smart factories can incorporate automated production equipment to improve product production efficiency. For example, automation equipment can include assembly robots that assemble products, transfer robots that transport assembled products, and delivery robots that deliver parts.

[0049] Additionally, smart factories can be equipped with or connected to an Enterprise Resource Planning (ERP) system, receiving necessary information from the ERP system. Through the ERP system, smart factories can access information related to product production, such as inventory and production schedules for all currently available resources.

[0050] Additionally, smart factories can be equipped with or connected to an MES (Manufacturing Execution System) and receive necessary information from the system. The MES system establishes a production plan based on customer orders, anticipated future orders, production speed, and available resources. This plan can then be provided to the smart factory to ensure smooth product production.

[0051] Meanwhile, the production quality confirmation system according to the present invention includes an unmanned robot, a digital twin system, and a check unit.

[0052] The unmanned robot (100) may include an image sensor unit (150) for extracting images of the assembly status of parts. The image sensor unit (150) may include a photographing device such as a camera. The movement of the unmanned robot (100) may be controlled by a robot control system, and the robot control system may receive production plans, etc. from an ERP system or an MES system and schedule a plan for the movement of the unmanned robot.

[0053] The unmanned robot (100) can extract a two-dimensional image of the assembled parts through the image sensor unit (150). The two-dimensional image extracted by the unmanned robot (100) becomes a target for comparison with the two-dimensional image extracted from the digital twin system (200) described later.

[0054] A digital twin system (200) refers to a system that digitally models real-world objects or processes, equivalent to analogue ones, and monitors or analyzes them. Based on the results, the system can modify and optimize the real-world objects or processes. In other words, it refers to a system that replicates the real world into a digital world and can reflect the results of processes performed in the digital world back into the real world.

[0055] Meanwhile, a smart factory can be implemented within the digital twin system (200) of the present invention, and a two-dimensional image of the same assembled part photographed by an unmanned robot (100) within the digital twin system (200) can be extracted. The two-dimensional image extracted from the digital twin system (200) becomes data for comparing the two-dimensional image extracted by the unmanned robot (100) and determining the assembly state of the part, and the two-dimensional images can be mutually compared in the check unit (300) described later to check the assembly state of the part.

[0056] More specifically, the digital twin system (200) can store 3D images of assembled parts. For example, assuming that part A is assembled to part B, the digital twin system (200) can store a 3D image of part A assembled to part B, and can also store 3D images of the individual assembly completion states of all parts. For example, the 3D image can be a CAD (Computer Aided Design) drawing.

[0057] The digital twin system (200) can extract a 2D image from a 3D image. For example, if an unmanned robot (100) extracts a 2D image of the assembly status of Part A, the digital twin system (200) can capture a 3D image of the assembly status of Part A and extract it as a 2D image.

[0058] In this way, the two-dimensional images extracted from the unmanned robot (100) and the digital twin system (200) are transmitted to the check unit (300), and the check unit (300) can compare the extracted two-dimensional images with each other to check the assembly status of parts of an unfinished product within the smart factory.

[0059] With reference to FIGS. 2 to 6, examples of 2D-image extraction and 2D-image comparison are described.

[0060] Referring to FIG. 2, an unfinished product (M) with assembled parts may be moved to a specific location, or an unmanned robot (100) may move to a location where the unfinished product with assembled parts is located. An unfinished product (M) may be moved to a specific location by a transport robot, or an unmanned robot (100) may receive a command from a robot control system and move to a location where the unfinished product (M) with assembled parts is located.

[0061] In order to check the assembly quality of an unfinished product (M) with assembled parts, the unmanned robot (100) photographs the unfinished product (M) with assembled parts using an image sensor unit (150). The unmanned robot (100) can extract a two-dimensional image using the image sensor unit (150). Referring to FIG. 3, the image sensor unit (150) provided in the unmanned robot (100) is configured as a multi-joint mechanism and is capable of rotation or length extension.

[0062] That is, in some cases, the image sensor unit (150) can be inserted into the inside of an unfinished product to extract a two-dimensional image, and a number of joints can rotate, and some joints can be extended in length by a telescopic method.

[0063] In particular, the terminal joint (151) where a photographing means such as a camera is located can rotate within a range of 360 degrees based on one axis, so that two-dimensional images can be extracted from multiple angles, and the other joints can rotate based on two axes, so that the unmanned robot (100) can extract two-dimensional images from multiple viewpoints.

[0064] Fig. 4 is a two-dimensional image extracted by moving the image sensor unit of the unmanned robot in the situation of Fig. 2 to capture a picture of Part A at a specific point in time. Fig. 5 shows an example of a three-dimensional image of Part A stored in the digital twin system, and Fig. 6 is a two-dimensional image extracted through a three-dimensional image of Part A stored in the digital twin system.

[0065] Hereinafter, with reference to FIGS. 4 to 6, the assembly status of part A(a) will be specifically described.

[0066] First, the unmanned robot (100) can extract a two-dimensional image of a part in an assembled state through the image sensor unit (150) at a location where a location information mark (500) including a QR code or barcode is recognized.

[0067] Referring to FIG. 2, a location information mark (500) may be engraved on the floor or the side of a transport robot within a smart factory surrounding an unfinished product. When an unmanned robot (100) scans the location information mark (500), information (coordinates) about the location may be included in a 2D image at a later time, and a 3D image may be captured within a digital twin system (200) based on the information.

[0068] That is, in order to compare the 2D images extracted by the unmanned robot (100) and the digital twin system (200), the 2D images must be extracted at similar points in time to ensure the reliability of the image similarity, so that the location information extracted by the unmanned robot (100) can be included in the 2D image. Meanwhile, the unmanned robot (100) may be provided with a separate camera that recognizes the location information mark (500) and is responsible for recognizing the location information mark (500). The 2D image extracted by the unmanned robot for the part A (a) is as shown in Fig. 4.

[0069] Meanwhile, when the unmanned robot (100) extracts a 2D image for the assembled part, the digital twin system (200) also extracts a 2D image for the same part A(a). At this time, the 2D image can be extracted by capturing a 3D image for the part assembly status stored in the digital twin system (200). That is, the digital twin system (200) can retrieve and capture a 3D image showing the assembly status of the part A(a) as in FIG. 5 and extract a 2D image as in FIG. 6.

[0070] As described above, the image sensor unit (150) of the unmanned robot (100) is capable of extracting two-dimensional images from multiple viewpoints, and the two-dimensional images extracted by the unmanned robot (100) may include location information from which the unmanned robot (100) extracted the two-dimensional image and movement information of the image sensor unit (150).

[0071] In other words, the location (coordinates) within the smart factory where the unmanned robot (100) photographed the part, the range of rotation of a plurality of joints constituting the image sensor unit (150) to extract a two-dimensional image by photographing the part by the unmanned robot (100) or the length of extension of some joints may be included in the two-dimensional image.

[0072] Since the location information and movement information are included in the 2D image extracted by the unmanned robot (100), the digital twin system (200) can capture a 3D image at the same time as the image sensor unit (150) based on the location information and movement information extracted by the unmanned robot (100) and extract a 2D image.

[0073] The digital twin system (200) extracts a two-dimensional image at the same time as the unmanned robot (100), and the two-dimensional images of FIGS. 4 and 6 are transmitted to the check unit (300) for comparison. Although the example is expressed as an image at one time point, two or more two-dimensional images may be used to check the assembly status of one component, and comparison of three or more two-dimensional images may be performed to ensure high reliability.

[0074]

[0075] Meanwhile, the check unit (300) can determine the similarity between two-dimensional images through the histogram of the extracted two-dimensional images and can determine the differences between the two-dimensional images. That is, the check unit can determine the similarity using the histogram of the two two-dimensional images to be compared. As a histogram-based method, a correlation coefficient, a chi-square distance, an intersection, or a Bhattacharyya technique can be used, and an appropriate technique can be selected and used depending on the characteristics of the part, the characteristics of the extracted two-dimensional image, etc.

[0076] Based on the similarity of the two-dimensional images extracted as described above, the assembly status of the component can be judged as suitable or defective in the check unit (300), and further, the defect of the image sensor unit (150) can be judged based on the similarity of the two-dimensional images. The similarity can have a value in the range of 0 to 100, and the closer the similarity is to 100, the more similar the two two-dimensional images are, and thus, the component assembly status is judged to be suitable.

[0077] For example, if the similarity is greater than or equal to the first value, it is judged as suitable, if the similarity is less than the first value, it is judged as defective, if the similarity is less than the first value and greater than or equal to the second value, it is judged as defective in the image sensor unit (150), and if the similarity is less than the second value, it can be judged as defective in the assembly state.

[0078] For example, the first value may be set to 95 and the second value may be set to 70. If the similarity is less than 70, the assembly state of the part may be judged to be poor. If the similarity is less than 95 and greater than 70, the assembly state of the part may be poor, but there may be an error in the movement (rotation or extension) of one of the joints of the image sensor unit.

[0079] That is, since the movement information of the image sensor unit (150) included in the 2D image captured by the image sensor unit (150) includes information that the terminal joint (151) of the image sensor unit (150) has rotated by 30 degrees, the digital twin system (200) also captures a 3D image at the same time, but in reality, if the terminal joint (151) of the image sensor unit (150) has rotated by 25 degrees due to insufficient operation of the motor included in the image sensor unit (150), the similarity may be measured as low.

[0080] In this way, the check unit (300) can determine the status of component assembly based on similarity or determine whether the image sensor unit (150) is defective.

[0081] Meanwhile, the production quality assurance system may further include a judgment unit (400) that compares the extracted two-dimensional images to calculate the degree of defect in the image sensor unit (150) of the unmanned robot (100). The judgment unit (400) is configured to calculate the degree of defect in the image sensor unit (150) when it is determined that the image sensor unit (150) is defective.

[0082] Specifically, the judgment unit (400) can calculate the degree of defect in the image sensor unit (150) based on the visual difference between the unmanned robot (100) and the digital twin system (200). The judgment unit (400) extracts the outline information of the extracted two-dimensional image to calculate the visual difference between the unmanned robot (100) and the digital twin system (200), and compares the outline information of the two-dimensional image extracted by the unmanned robot (100) based on the outline information acquired from the digital twin system (200). By comparing the outline information of the two two-dimensional images, the visual difference between the unmanned robot (100) and the digital twin system (200) can be calculated.

[0083] For example, the judgment unit (400) compares the outline information for similar points within two two-dimensional images, and calculates the visual difference between the unmanned robot and the digital twin system by determining how much the outline information of the two-dimensional image extracted by the unmanned robot (100) should be moved and reduced or enlarged to match the outline information of the digital twin system.

[0084] For example, referring to FIG. 7, for part A(a), if the outline information (P) for part A(a) obtained from the unmanned robot (100) is rotated by 30 degrees and enlarged by 1.25 times, the data that matches the outline information (Q) for part A(a) obtained from the digital twin system (200) can be used to calculate the visual difference between the unmanned robot (100) and the digital twin system (200).

[0085] Through the generated visual difference, the judgment unit (400) calculates the degree of defect in the image sensor unit (150). For example, the result printed by the judgment unit (400) may be “the terminal joint motor of the image sensor unit needs to be replaced,” and accordingly, the unmanned robot (100) can be recovered to repair or replace the image sensor unit (150).

[0086] When the unmanned robot (100) is recovered and the image sensor unit (150) is repaired, the unmanned robot (100) can be put back into the smart factory to check the assembly quality of the parts, and then, through the extraction of 2D images of the parts and comparison with the 2D images extracted by the digital twin system (200), it can be determined whether a defect problem of the image sensor unit (150) is continuously detected, and if the defect problem of the image sensor unit (150) no longer occurs, the production quality check system can be continuously operated.

[0087]

[0088] Figures 8 and 9 are flowcharts for one embodiment of the present invention. Referring to Figures 8 and 9, a method for verifying production quality according to the present invention includes a step (S100) in which an unmanned robot extracts a two-dimensional image of a part in an assembled state; a step (S200) in which a two-dimensional image of a part in an assembled state is extracted from a three-dimensional image of a part in which assembly is completed and stored in a digital twin system; and a step (S300) in which a checking unit compares the extracted two-dimensional images to verify the assembly state of the part.

[0089] The step (S100) of extracting a 2D image in a digital twin system may be a step of capturing a 3D image at the same time as the time at which the unmanned robot extracted the 2D image.

[0090] The step (S200) of checking the assembly status of the parts may be a step of determining the similarity between the two-dimensional images through the histogram of the extracted two-dimensional images and determining the differences between the two-dimensional images.

[0091] The step (S300) of checking the assembly status of the parts may be a step of judging the assembly status of the parts as suitable or defective based on the similarity of the two-dimensional images, and judging the defect of the image sensor unit based on the similarity of the two-dimensional images.

[0092] If the similarity is greater than or equal to the first value, it is judged as suitable (S310). If the similarity is less than the first value, it is judged as defective. If the similarity is less than the first value and greater than or equal to the second value, it is judged as defective in the image sensor unit (S330). If the similarity is less than the second value, it can be judged as defective in the assembly state (S350).

[0093] In the step (S300) of checking the assembly status of the parts, if the image sensor part is judged to be defective, a step (S400) of comparing the extracted 2D images and calculating the degree of defect in the image sensor part of the unmanned robot may be further performed.

[0094] The step (S400) in which the judgment unit calculates the degree of defect in the image sensor unit of the unmanned robot may be a step in which the degree of defect in the image sensor unit is calculated based on the visual difference between the unmanned robot and the digital twin system.

[0095] The step (S100) of extracting a two-dimensional image by an unmanned robot may include a step (S110) of recognizing a location information mark including a QR code or barcode by the unmanned robot; and a step (S130) of extracting a two-dimensional image of a part in an assembled state through an image sensor unit at a point where the location information mark is recognized.

[0096]

[0097] Although the present invention has been illustrated and described with respect to specific embodiments thereof, it will be apparent to those skilled in the art that the present invention may be variously improved and modified without departing from the technical spirit of the present invention as defined by the following claims.

[0098]

[0099] [Explanation of symbols]

[0100] 100: Unmanned robot

[0101] 200: Digital Twin System

[0102] 300: Check section

[0103] 400: Judgment Department

[0104] 500: Location information marker

Claims

1. An unmanned robot that extracts a two-dimensional image of an assembled part through an image sensor; A digital twin system that stores a 3D image of a completed assembled part and extracts a 2D image of the assembled part from the 3D image; and A production quality verification system including a check section for checking the assembly status of parts by comparing extracted two-dimensional images.

2. In claim 1, A production quality assurance system characterized in that the image sensor section is configured as a multi-joint mechanism capable of rotation or extension, and an unmanned robot can extract two-dimensional images of assembled parts from multiple viewpoints.

3. In claim 1, A production quality assurance system characterized in that a two-dimensional image extracted by an unmanned robot includes location information and movement information of an image sensor unit where the two-dimensional image was extracted by the unmanned robot.

4. In claim 1, A production quality assurance system characterized in that the digital twin system extracts a two-dimensional image of an assembled part from a three-dimensional image, and extracts the two-dimensional image of the assembled part at the time when the unmanned robot extracts the assembled part.

5. In claim 1, A production quality verification system characterized in that the check section determines the similarity between two-dimensional images through a histogram of the extracted two-dimensional images and determines the differences between the two-dimensional images.

6. In claim 1, A production quality verification system characterized in that the assembly status of a part is judged as suitable or defective in a check section based on the similarity of a two-dimensional image, and the defect of an image sensor section is judged based on the similarity of a two-dimensional image.

7. In claim 6, If the similarity is greater than the first value, it is considered suitable. If the similarity is less than the first value, it is judged as defective. If the similarity is less than the first value or greater than the second value, it is judged as a defect in the image sensor. A production quality verification system characterized in that if the similarity is less than the second value, it is judged as a defective assembly.

8. In claim 6, It further includes a judgment unit that compares the extracted two-dimensional images to calculate the degree of defect in the image sensor part of the unmanned robot; A production quality assurance system characterized in that, when the inspection unit determines that the image sensor unit is defective, the judgment unit calculates the degree of defect in the image sensor unit based on the visual difference between the unmanned robot and the digital twin system.

9. In claim 8, A production quality assurance system characterized in that the judgment unit extracts outline information of the extracted two-dimensional image and calculates the visual difference between the unmanned robot and the digital twin system.

10. In claim 1, A production quality assurance system characterized in that an unmanned robot extracts a two-dimensional image of an assembled part through an image sensor at a location where a location information mark containing a QR code or barcode is recognized.

11. In claim 10, A production quality assurance system characterized by an unmanned robot equipped with a camera that recognizes location information tags.

12. A step in which an unmanned robot extracts a two-dimensional image of an assembled part; A step of extracting a two-dimensional image of an assembled part from a three-dimensional image of an assembled part stored in a digital twin system; and A method for verifying production quality, comprising: a step of verifying the assembly status of a part by comparing the extracted two-dimensional image in a check section; 13. In claim 12, A method for verifying production quality, characterized in that the step of extracting a two-dimensional image in a digital twin system is a step of capturing a three-dimensional image at the same time as the time at which the unmanned robot extracted the two-dimensional image.

14. In claim 12, A method for verifying production quality, characterized in that the step of verifying the assembly status of a part is a step of judging the similarity between two-dimensional images through a histogram of the extracted two-dimensional images and judging the differences between the two-dimensional images.

15. In claim 12, A method for verifying production quality, characterized in that the step of verifying the assembly status of a part is a step of judging the assembly status of the part as suitable or defective based on the similarity of a two-dimensional image, and a step of verifying the defect of the image sensor part based on the similarity of a two-dimensional image.

16. In claim 15, If the similarity is greater than the first value, it is considered suitable. If the similarity is less than the first value, it is judged as defective. If the similarity is less than the first value or greater than the second value, it is judged as a defect in the image sensor. A production quality verification method characterized in that if the similarity is less than the second value, it is determined that the assembly condition is defective.

17. In claim 15, In the step of checking the assembly status of the parts, if it is determined that the image sensor part is defective, A production quality verification method characterized by further performing a step of comparing the extracted two-dimensional images and calculating the degree of defects in the image sensor part of the unmanned robot.

18. In claim 17, A production quality assurance system characterized in that the step of calculating the degree of defect in the image sensor part of the unmanned robot by the judgment unit is a step of calculating the degree of defect in the image sensor part based on the visual difference between the unmanned robot and the digital twin system.

19. In claim 12, The steps for an unmanned robot to extract a 2D image are: A step in which an unmanned robot recognizes a location information mark containing a QR code or barcode; A production quality verification method characterized by including a step of extracting a two-dimensional image of a part in an assembled state through an image sensor unit at a point where a location information mark is recognized.

Images