Foreign matter tracking method, and terminal therefor

Abstract

The present disclosure relates to a foreign matter tracking method capable of accurately and promptly determining, through artificial intelligence, the type and cause of a foreign matter found in a product production line, and a terminal therefor, and provides a foreign matter tracking method comprising the steps of: capturing an image of a foreign matter through a camera; inferring the type and source of ingress of the foreign matter by inputting the captured image of the foreign matter into an artificial intelligence model; and outputting a measure for preventing additional ingress of the foreign matter on the basis of the inferred type and source of ingress of the foreign matter.

Description

Method for tracking foreign objects and terminal for the same

[0001] The present disclosure relates to a method and a terminal for analyzing foreign matter entering a product production line or factory using artificial intelligence to track the cause of the foreign matter's entry.

[0002] For example, foreign substances such as air bubbles and dust entering products during the manufacturing process, such as semiconductors or display panels, cause product defects. Foreign substances generated on the production line (or factory) have a significantly negative impact on product yield. Therefore, if the entry of foreign substances into the production line is detected, it is crucial to promptly identify the type of foreign substance, find its cause, and prevent its further entry.

[0003] However, foreign matter found on a product production line is very fine in size and varies in shape and / or color, so a lot of time and effort is required to identify the type and cause of occurrence of the foreign matter. This will be explained with reference to FIG. 1. FIG. 1 is a flowchart of a method for tracking foreign matter introduced into a product production line according to the prior art.

[0004] First, defects in the product may occur due to the ingress of foreign matter into the product production line [S101].

[0005] However, a worker (or manager) on the product production line can collect foreign matter from a defective product and identify the appearance of the foreign matter (e.g., shape, color, size, etc.) using an optical device such as a microscope [S103, S105].

[0006] If a worker can identify the type and cause of foreign matter through its appearance, they can take measures to prevent foreign matter from entering the product production line.

[0007] However, if the worker cannot determine the type and cause of the foreign object based only on its appearance, the worker may inevitably have to send the foreign object to an external foreign object analysis center [S107].

[0008] The foreign matter analysis center can analyze the physical and chemical properties of foreign matter and determine the type and cause of occurrence of foreign matter based on the analyzed properties and the appearance of foreign matter [S109, S111].

[0009] The product production line operator can receive a report from the foreign matter analysis center regarding the type and cause of the foreign matter and perform improvement work to prevent such problems from recurring [S113].

[0010] As such, significant time, cost, and effort can be invested from the time foreign matter is sent to an external analysis center and a report is received until corrective measures are taken; in some cases, the production line may need to be shut down from the occurrence of the foreign matter until the remediation work is completed. If there are errors in the analysis by the analysis center, the shutdown of the production line may be even longer.

[0011] Therefore, there is a need to continue researching methods to accurately and promptly identify the types and causes of foreign substances found on product production lines.

[0012] The present disclosure is proposed to solve such problems and aims to provide a foreign object tracking method and a terminal therefor that can accurately and quickly identify the type and cause of occurrence of foreign objects found in a product production line through artificial intelligence.

[0013] To achieve the above objective, the present disclosure may provide a terminal comprising a camera, a display, a memory storing an artificial intelligence model, and a control unit that photographs a foreign object through the camera, inputs the image of the photographed foreign object into the artificial intelligence model to infer the type and source of the foreign object, and outputs a measure to prevent further inflow of the foreign object based on the inferred type and source of the foreign object.

[0014] The above control unit can further output the cause of occurrence of the foreign substance inferred from the above inflow source.

[0015] The above control unit can further output information regarding the frequency of occurrence of the foreign substance in the production line where the foreign substance was found.

[0016] Training data for training the above artificial intelligence model can be generated by performing at least one of brightness adjustment, enlargement, reduction, rotation, etc. on a plurality of previously captured foreign object images.

[0017] The above artificial intelligence model may be a CNN (Convolutional Neural Network) family.

[0018] In addition, to achieve the above objective, the present disclosure may provide a foreign object tracking method comprising the steps of: photographing a foreign object through a camera; inputting the image of the photographed foreign object into an artificial intelligence model to infer the type and source of the foreign object; and outputting a measure to prevent further inflow of the foreign object based on the inferred type and source of the foreign object.

[0019] The effects of the foreign object tracking method and the terminal for the same according to the present disclosure are described as follows.

[0020] According to at least one of the various aspects of the present disclosure, there is an advantage in that the type and cause of occurrence of foreign matter found in a product production line can be accurately and quickly identified through artificial intelligence.

[0021] FIG. 1 is a flowchart of a method for tracking foreign matter introduced into a product production line according to the prior art.

[0022] FIG. 2 is a block diagram of a terminal that can be used in foreign object tracking according to the present disclosure.

[0023] FIG. 3 is an example of determining defects in a product produced by taking photos with a camera on a product production line according to one aspect of the present disclosure.

[0024] FIG. 4 illustrates an example of data construction for foreign matter analysis according to one aspect of the present disclosure.

[0025] FIGS. 5 and 6 are examples of a foreign matter management database according to one aspect of the present disclosure.

[0026] FIG. 7 is an example of training an artificial intelligence model for tracking foreign objects according to one aspect of the present disclosure.

[0027] Figure 8 is an example of the preprocessing of training data for the artificial intelligence model of Figure 7.

[0028] FIG. 9 illustrates an example of foreign object tracking using an artificial intelligence model according to one aspect of the present disclosure.

[0029] FIG. 10 illustrates an example of a user interface for displaying the results of foreign object tracking of FIG. 9.

[0030] Hereinafter, embodiments disclosed in this disclosure will be described in detail with reference to the attached drawings. Identical or similar components, regardless of drawing symbols, are assigned the same reference number, and redundant descriptions thereof will be omitted. The suffixes "module" and "part" used for components in the following description are assigned or used interchangeably solely for the ease of drafting the disclosure and do not inherently possess distinct meanings or roles. Furthermore, in describing embodiments disclosed in this disclosure, if it is determined that a detailed description of related prior art could obscure the essence of the embodiments disclosed in this disclosure, such detailed description will be omitted. Additionally, the attached drawings are intended only to facilitate understanding of the embodiments disclosed in this disclosure; the technical concept disclosed in this disclosure is not limited by the attached drawings, and it should be understood that they include all modifications, equivalents, and substitutions that fall within the spirit and technical scope of the present invention.

[0031] These components may each be composed of separate individual hardware modules or implemented as two or more hardware modules, or two or more components may be implemented as a single hardware module, and, of course, in some cases, may also be implemented as software.

[0032] Terms including ordinal numbers, such as first, second, etc., may be used to describe various components, but said components are not limited by said terms. These terms are used solely for the purpose of distinguishing one component from another.

[0033] When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. On the other hand, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between.

[0034] The singular expression includes the plural expression unless the context clearly indicates otherwise. In this disclosure, terms such as “comprising” or “having” are intended to specify the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the disclosure, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0035] In the present disclosure, the expression “at least one of A and B” may mean “A,” may mean “B,” or may mean both “A” and “B.”

[0036] The terminals described in the present disclosure may include mobile phones, smartphones, laptop computers, digital broadcasting terminals, PDAs (personal digital assistants), PMPs (portable multimedia players), navigation systems, slate PCs, tablet PCs, ultrabooks, wearable devices (e.g., smartwatches, smart glasses, head-mounted displays), etc.

[0037] However, those skilled in the art will readily understand that the configuration according to the embodiments described in this disclosure may also be applied to fixed terminals such as digital TVs, desktop computers, and digital signage, except in cases where it is applicable only to mobile terminals.

[0038] Hereinafter, with reference to FIG. 2, a terminal that can be used in foreign object tracking according to the present disclosure will be described. FIG. 2 is a block diagram of a terminal that can be used in foreign object tracking according to the present disclosure.

[0039] The above terminal (100) may include a communication unit (110), an input unit (120), a display (151), a memory (170), and a control unit (180), etc. The terminal described in the present disclosure may have more or fewer components than the components listed above.

[0040] The communication unit (110) may include one or more modules that enable wired or wireless communication between the terminal (100) and a wired or wireless communication system, between the terminal (100) and another terminal (100), or between the terminal (100) and an external server. Additionally, the communication unit (110) may include one or more modules that connect the terminal (100) to one or more networks.

[0041] The input unit (120) may include a camera (121) for inputting a video signal and a user input unit (123, e.g., a touch key, a mechanical key, a keyboard, a mouse, a touchpad, a touchscreen, etc.) for receiving information or commands from a user.

[0042] The terminal (100) may be equipped with one or more cameras (121). The camera (121) processes image frames, such as still images or video, obtained by an image sensor in shooting mode. The processed image frames may be displayed on a display unit (151) or stored in a memory (170). Meanwhile, the multiple cameras (121) provided in the terminal (100) may be arranged to form a matrix structure, and through the cameras (121) forming such a matrix structure, multiple image information having various angles or focal points may be input to the terminal (100). Additionally, the multiple cameras (121) may be arranged in a stereo structure to acquire left and right images for implementing stereoscopic images. Furthermore, the multiple cameras (121) may include a depth camera and / or a Time of Flight (TOF) camera for stereoscopically sensing a subject.

[0043] The display (151) can be formed as a touch screen by forming a layered structure with the touch sensor or as an integral unit. This touch screen functions as a user input unit (123) that provides an input interface between the terminal (100) and the user, and at the same time can provide an output interface between the terminal (100) and the user.

[0044] The display (151) displays (outputs) information processed by the terminal (100). For example, the display (151) can display information on the execution screen of an application running on the terminal (100), or UI (User Interface) and GUI (Graphic User Interface) information based on such execution screen information.

[0045] The memory (170) stores data that supports various functions of the terminal (100). The memory (170) can store a number of application programs (or applications) running on the terminal (100), data for the operation of the terminal (100), and commands. At least some of these application programs may be downloaded from an external server via wireless communication. In addition, at least some of these application programs may exist on the mobile terminal (100) from the time of shipment for the basic functions of the terminal (100). Meanwhile, the application programs may be stored in the memory (170), installed on the terminal (100), and driven by the control unit (180) to perform the operation (or function) of the terminal.

[0046] The memory (170) can store a program for the operation of the control unit (180) and can also temporarily store input / output data (e.g., phonebook, message, still image, video, etc.).

[0047] The memory (170) may include at least one type of storage medium among flash memory type, hard disk type, SSD type (Solid State Disk type), SSD type (Silicon Disk Drive type), multimedia card micro type, card type memory (e.g., SD or XD memory, etc.), RAM (random access memory; RAM), SRAM (static random access memory), ROM (read-only memory; ROM), EEPROM (electrically erasable programmable read-only memory), PROM (programmable read-only memory), magnetic memory, magnetic disk, and optical disk. The terminal (100) may be operated in connection with web storage that performs the storage function of the memory (170) on the internet.

[0048] The memory (170) may include an artificial intelligence model (171). Artificial intelligence will be explained again later.

[0049] In addition to operations related to the application program, the control unit (180) typically controls the overall operation of the terminal (100). The control unit (180) can provide or process appropriate information or functions to the user by processing signals, data, information, etc. that are input or output through the components described above, or by running an application program stored in the memory (170).

[0050] Additionally, the control unit (180) can control at least some of the components examined together with FIG. 1 in order to run an application program stored in memory (170). Furthermore, the control unit (180) can operate at least two or more of the components included in the terminal (100) in combination with each other to run the application program.

[0051] The control unit (180) may further include an AI engine (182) for artificial intelligence processing.

[0052] Meanwhile, various embodiments described below may be implemented in a recording medium readable by a computer or a similar device, for example, using software, hardware, or a combination thereof.

[0053] Hereinafter, artificial intelligence (AI) that can be utilized in the present disclosure will be described.

[0054] Artificial intelligence refers to the field of researching artificial intelligence or the methodologies to create it, while machine learning refers to the field of researching methodologies to define and solve various problems addressed within the field of artificial intelligence. Machine learning is also defined as an algorithm that improves performance on a task through continuous experience.

[0055] An Artificial Neural Network (ANN) is a model used in machine learning that can refer to any model capable of problem-solving, composed of artificial neurons (nodes) that form a network through the connection of synapses. An artificial neural network can be defined by connection patterns between neurons in different layers, a learning process that updates model parameters, and an activation function that generates output values.

[0056] An artificial neural network may include an input layer, an output layer, and optionally one or more hidden layers. Each layer may include one or more neurons, and the artificial neural network may include synapses connecting the neurons. In an artificial neural network, each neuron may output a function value of an activation function for input signals, weights, and biases input through the synapses.

[0057] Model parameters refer to parameters determined through learning, including synaptic connection weights and neuron biases. Hyperparameters, on the other hand, refer to parameters that must be set prior to training in a machine learning algorithm, including the learning rate, number of iterations, mini-batch size, and initialization function.

[0058] The objective of training an artificial neural network can be viewed as determining model parameters that minimize the loss function. The loss function can be used as an indicator to determine optimal model parameters during the training process of an artificial neural network.

[0059] Machine learning can be classified into supervised learning, unsupervised learning, and reinforcement learning depending on the learning method.

[0060] Supervised learning refers to a method of training an artificial neural network with labels provided for the training data; a label can refer to the correct answer (or result) that the neural network must infer when the training data is input. Unsupervised learning refers to a method of training an artificial neural network without labels provided for the training data. Reinforcement learning refers to a learning method in which an agent defined within an environment is trained to select an action or sequence of actions that maximizes the cumulative reward in each state.

[0061] Machine learning implemented using a Deep Neural Network (DNN) that includes multiple hidden layers among artificial neural networks is also called Deep Learning, and Deep Learning is a part of Machine Learning. Hereinafter, Machine Learning is used in a sense that includes Deep Learning.

[0062] Object detection models using machine learning include the single-stage YOLO (You Only Look Once) model and the two-stage Faster R-CNN (Regions with Convolutional Neural Networks) model.

[0063] The YOLO (You Only Look Once) model is a model that can predict objects present in an image and their locations by looking at the image only once.

[0064] The YOLO (You Only Look Once) model divides the original image into grids of equal size. Then, for each grid, it predicts the number of bounding boxes specified in a predefined form centered on the grid center, and calculates confidence based on this.

[0065] Subsequently, whether the image contains an object or consists solely of a background is included, and a location with high object confidence is selected so that the object category can be identified.

[0066] The Faster R-CNN (Regions with Convolutional Neural Networks) model is a model that can detect objects faster than the RCNN model and the Fast RCNN model.

[0067] This explains the Faster R-CNN (Regions with Convolutional Neural Networks) model in detail.

[0068] First, feature maps are extracted from the image using a Convolutional Neural Network (CNN) model. Based on the extracted feature maps, multiple Regions of Interest (RoIs) are identified. RoI pooling is performed for each region of interest.

[0069] RoI pooling is a process of setting a grid to fit a predetermined size of H x W for a feature map onto which a region of interest is projected, and extracting the largest value for each cell contained in the grid to extract a feature map with a size of H x W.

[0070] A feature vector is extracted from a feature map with size H x W, and identification information of an object can be obtained from the feature vector.

[0071] Hereinafter, with reference to FIG. 3, we will describe how to determine defects in a product manufactured by taking photos with a camera on a product manufacturing line. FIG. 3 is an example of determining defects in a product manufactured by taking photos with a camera on a product manufacturing line according to one aspect of the present disclosure.

[0072] A camera (121) of the terminal (100) of FIG. 1 can be installed at a specific location on the product production line (200) so as to photograph the product (300) being produced.

[0073] A camera (121) can be installed at a specific location on the product production line (200) to photograph a specific area (or part) of the product (300).

[0074] The control unit (180) can control each of the multiple products (300) being produced to take a specific area of ​​each product (300) when each of them reaches a specific location on the product production line (200).

[0075] The control unit (180) can determine whether the product (300) is defective by comparing the standard image of a normal product with the captured image through pattern recognition.

[0076] For example, the control unit (180) can determine whether there is an assembly defect in the product. The control unit (180) can also determine whether there is a formation defect in a pattern such as a semiconductor pattern. In addition, the control unit (180) can determine whether there are scratches on the product and / or whether foreign substances such as threads, stains, etc. have entered the product.

[0077] If a product is determined to be defective, it may be discarded or reworked.

[0078] As such, identifying product defects on the production line through camera imaging is a measure that can only be taken after a defect has already occurred, making it practically difficult to prevent such defects. Furthermore, in cases where defects are caused by the ingress of foreign matter, it is practically difficult to identify the type of foreign matter and its cause based solely on the captured product images.

[0079] Below, I will explain measures to prevent product defects caused by the ingress of foreign substances by promptly and accurately identifying the types of foreign substances that may be found on the product production line and their sources of entry.

[0080] First, with reference to FIG. 4, I will explain the construction of data for foreign matter analysis. FIG. 4 illustrates an example of the construction of data for foreign matter analysis according to one aspect of the present disclosure.

[0081] First, as shown in (4-1) of Fig. 4, foreign matter may be found on the product production line or around it.

[0082] Then, as shown in (4-2) of Fig. 4, the worker (or manager) of the product production line can collect the foreign object as soon as it is discovered.

[0083] And, as shown in (4-3) and (4-4) of FIG. 4, the captured foreign matter can be observed under a microscope to determine the external characteristics and size of the foreign matter, and the organic and inorganic composition of the foreign matter can be analyzed.

[0084] And, as shown in (4-5) of FIG. 4, the results of observation and analysis of the captured foreign matter can be updated in the foreign matter management database. In addition, source information for each foreign matter can be updated together. The data stored in the foreign matter management database can be used as training data for an artificial intelligence model.

[0085] Hereinafter, with reference to FIGS. 5 and FIGS. 6, we will examine an example of a foreign matter management database. FIGS. 5 and FIGS. 6 are examples of a foreign matter management database according to one aspect of the present disclosure.

[0086] As shown in Fig. 5, the foreign matter management database can record the product production line where each foreign matter occurred and the number of times, by type and composition of the foreign matter.

[0087] Additionally, as illustrated in FIG. 6, the foreign matter management database may record information regarding the location where each foreign matter occurred (610), information regarding the shape, size, and composition of each foreign matter (620), and information regarding the source of each foreign matter.

[0088] As mentioned above, data within the foreign matter management database can be used as training data for artificial intelligence algorithms.

[0089] Hereinafter, with reference to FIGS. 7 and FIGS. 8, the training of an artificial intelligence model for foreign object tracking will be described. FIG. 7 is an example of the training of an artificial intelligence model for foreign object tracking according to one aspect of the present disclosure. FIG. 8 is an example of the preprocessing of training data of the artificial intelligence model of FIG. 7.

[0090] As shown in FIG. 7, data within the foreign matter management database can be used as training data for an artificial intelligence model [S710].

[0091] An artificial intelligence learning system (or processor) (1000) can perform data preprocessing to utilize data within a foreign matter management database as training data. Data preprocessing may include data augmentation preprocessing for augmenting training data and image preprocessing for ease of learning.

[0092] Through data augmentation preprocessing, as illustrated in FIG. 8, a plurality of modified foreign object images (820) can be generated from a single original foreign object image (810) by performing at least one of brightness adjustment, enlargement, reduction, rotation, etc. for each foreign object image (810). Both the original foreign object image (810) and the plurality of replicated modified foreign object images (820) can be used as training data.

[0093] An artificial intelligence learning system (1000) can train an artificial intelligence model using a portion of the training data (e.g., 80% of the total training data) and use the trained artificial intelligence model for foreign object classification [S730, S740]. The artificial intelligence model being trained may be a model of the Convolutional Neural Network (CNN) family.

[0094] The artificial intelligence learning system (1000) can evaluate the learned artificial intelligence model using the remainder of the learning data (e.g., 20% of the total learning data) [S750].

[0095] The artificial intelligence model prepared in this way can be loaded into the memory (170) of the terminal of FIG. 2 as the artificial intelligence model (171) described above.

[0096] The use of the artificial intelligence model prepared in this manner will be explained with reference to FIGS. 9 and FIG. 10. FIG. 9 illustrates an example of foreign object tracking using an artificial intelligence model according to one aspect of the present disclosure, and FIG. 10 illustrates an example of a user interface for displaying the results of foreign object tracking of FIG. 9.

[0097] As shown in (9-1) of FIG. 9, if a foreign object is found anywhere on the product production line, the operator can photograph the foreign object using a terminal (100).

[0098] The control unit (180) of the terminal (100) can infer the type and cause of occurrence of foreign matter by inputting the captured image of foreign matter into the artificial intelligence model, as shown in (9-2) of FIG. 9.

[0099] The control unit (180) can output the type of foreign matter and the cause of occurrence of the inferred foreign matter and countermeasures for it to the display (151).

[0100] Therefore, as illustrated in (9-3) of FIG. 9, the worker can take measures against the foreign object immediately upon discovery. For example, the worker can discard the product manufactured at the time the foreign object is discovered or perform rework on the product.

[0101] Hereinafter, with reference to FIG. 10, I will describe an interface that outputs the type of foreign matter inferred above, the cause of occurrence, and countermeasures therefor.

[0102] In the first item (2100) of the foreign matter interface (2000), the operator can input information about the product production line and process where the foreign matter was found.

[0103] And, the operator can upload the captured image of the foreign object in the first area (2210) of the second item (2200) of the foreign object interface (2000).

[0104] Then, multiple types of foreign substances can be listed in order of accuracy in the second area (2220) of the second item (2200) of the foreign substance interface (2000). In FIG. 10, the first type of foreign substance inferred is exemplified as polyester.

[0105] Images of the same or similar foreign substances found in the same product production line in the past may be output in the third area (2230) of the second item (2200) of the foreign substance interface (2000).

[0106] And, the composition analysis result of the invented foreign substance can be output in the fourth area (2240) of the second item (2200) of the foreign substance interface (2000).

[0107] The result of the inferred foreign substance source can be output in the first area (2310) of the third item (2300) of the foreign substance interface (2000).

[0108] In the second area (2320) of the third item (2300) of the foreign matter interface (2000), the cause of foreign matter generation from the inferred source of foreign matter can be output.

[0109] A method to prevent the inferred foreign matter from further entering the third area (2330) of the third item (2300) of the foreign matter interface (2000) may be output.

[0110] In the fourth item (2400) of the foreign matter interface (2000), the frequency of occurrence of the corresponding foreign matter by process can be output.

[0111] In the 5th item (2500) of the foreign matter interface (2000), the actual defect history caused by the foreign matter can be output.

[0112] In the foregoing, it has been described that the artificial intelligence model is installed within the terminal (100). However, the present disclosure is not limited thereto. Even if the artificial intelligence model is not installed within the terminal (100), the terminal (100) may communicate with an external server (not shown) on which the artificial intelligence model is installed to perform artificial intelligence inference regarding the foreign object.

[0113] The above-described disclosure can be implemented as computer-readable code on a medium on which a program is recorded. A computer-readable medium includes all types of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable media include a Hard Disk Drive (HDD), a Solid State Disk (SSD), a Silicon Disk Drive (SSD), ROM, RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc. Additionally, the computer may include a processor of an artificial intelligence device.

Claims

1. Camera; display; Memory for storing artificial intelligence models; and A foreign object is photographed through the above camera, and The image of the foreign substance captured above is input into the artificial intelligence model to infer the type and source of the foreign substance, and A terminal comprising: a control unit that outputs measures to prevent further inflow of foreign matter based on the type and source of inflow of the foreign matter inferred above.

2. In claim 1, the control unit A terminal characterized by further outputting the cause of occurrence of the foreign substance inferred from the above-mentioned input source.

3. In claim 1, the control unit A terminal characterized by further outputting information regarding the frequency of occurrence of the said foreign substance in the production line where the said foreign substance was found.

4. In Paragraph 1, A terminal device characterized by generating training data for training the above artificial intelligence model by performing at least one of brightness adjustment, enlargement, reduction, rotation, etc. on a plurality of previously captured foreign object images.

5. In Paragraph 1, A terminal characterized by the above artificial intelligence model being of the CNN (Convolutional Neural Network) family.

6. Step of photographing foreign objects through a camera; A step of inputting the image of the foreign substance captured above into an artificial intelligence model to infer the type and source of the foreign substance; and A foreign substance tracking method comprising the step of outputting measures to prevent further inflow of the foreign substance based on the type and source of inflow of the foreign substance inferred above.

7. In Paragraph 6, A foreign substance tracking method further comprising the step of outputting the cause of occurrence of the foreign substance inferred from the above-mentioned source.

8. In Paragraph 6, A foreign matter tracking method further comprising the step of outputting information on the frequency of occurrence of the foreign matter in the production line where the foreign matter was found.

9. In Paragraph 6, A foreign object tracking method characterized by generating training data for training the above artificial intelligence model by performing at least one of brightness adjustment, enlargement, reduction, rotation, etc. on a plurality of previously captured foreign object images.

10. In Paragraph 6, A foreign object tracking method characterized by the above artificial intelligence model being of the CNN (Convolutional Neural Network) family.

Images