KR20220139622A - Compensation system for laser processing apparatus and method for compensate laser processing apparatus - Google Patents

Abstract

A compensation system for a laser processing device according to the present invention comprises: a profile module including a first light source and an optical sensor for measuring the surface of a specimen using the reflected light of the first light source reflected from the specimen; a vision module including a second light source, and an image collecting part for collecting the image of the specimen using the image light of the second light source reflected from the specimen; a laser processing module including a guide laser output part for displaying the central axis of the scanner module and a processing laser output part for processing; and a scanner module including a lens part and a mirror part, which guides and reflects the paths of the first light source, the reflected light, the image light, guide laser, and processing laser, respectively, a stage on which the specimen is located, a stage driving part for moving the stage in the x-axis and y-axis directions, and an axis compensation part for rotating the axis of the scanner module, and substantially coaxially guiding the first light source, the reflected light, the image light, the guide laser, and the processing laser. Therefore, provided are a compensation system for a laser processing device and a compensation method for a laser processing device, wherein vision coordinates can be matched accurately with respect to scanner coordinates.

Description

Laser processing equipment compensation system and laser processing equipment compensation method

The present invention relates to a laser processing apparatus correction system and a laser processing apparatus correction method.

When it is necessary to precisely process a pattern with a complex shape at a high speed in a laser processing apparatus, it is common to apply a scanner-based laser processing apparatus. Here, machining refers to any process that can be performed using a laser, such as drilling, cutting, grinding, character marking, and welding. What is essential for processing with a scanner-based laser processing device is to use an F-theta lens so that the laser beam can be delivered not only to the center but also to the periphery in a focused state. By adopting an F-theta lens in the scanner, it is possible to secure a working area from several tens of mm to several hundred mm.

However, as a large-aperture lens, an F-theta lens has optical aberration and distortion aberration is quite large. Specifically, there is a problem in that barrel distortion or pincushion distortion cannot be avoided.

On the other hand, as recent technology trends in the fields of displays, semiconductors, printed circuit boards (PCBs) and electric vehicles have become lighter, thinner and smaller, the demand for processing microscopic areas with high precision and accuracy is increasing. To solve this problem, a 2D camera-based vision device is being used by grafting it to a scanner-based laser processing system. The vision device and the laser processing system share an F-theta lens, but some optical systems are different, which causes coordinate mismatch.

In addition, since the vision device mainly uses 650 nm or less of the visible light region, and the laser processing system uses an IR laser of 1064 nm and a green laser of 500 nm or less, chromatic aberration essentially exists between the two systems.

As a result, the shape of the actual workpiece processed by the laser processing system is different from the image of the workpiece captured by the CCD camera of the vision device. In other words, even if the scanner can properly draw a 200mm square, the image taken with this pattern is not a 200mm square. To solve this problem, the vision coordinates should be corrected based on the scanner coordinates.

The characteristics or disadvantages of the conventional correction method are roughly as follows. First, since a method of drawing a grid pattern for calibration directly on a specimen with a laser is generally used, the inconvenience of having to produce a calibration specimen every time it is calibrated and the operating cost of equipment increase. Second, in the case of a specimen that has been used once, it cannot be reused because a pattern is drawn on the surface and some deformation occurs due to heat caused by laser processing. Third, although there are deviations depending on the specifications and performance of the laser and optical system and the F-theta lens, the edge of the grid pattern drawn on the specimen is not sharp, the line width (usually 300um) is not constant, and the line width is Since it is difficult to process less than 100um, it acts as an error in the recognition of the edge of the pattern and the detection of the coordinates of the grid points during the correction process, affecting the correction accuracy. Fourth, non-coaxial illumination is often used as the illumination for the vision device used in the scanner-based laser processing system, and the quality difference between the central image and the peripheral image is large. Therefore, it is difficult to automate the correction mechanism because it is often necessary to manually adjust the image processing parameters when edge detection, intersection detection, and coordinate detection by applying a rule-based image processing algorithm.

Accordingly, there is a need for a laser processing apparatus correction system that can accurately match the vision coordinates based on the scanner coordinates, can be reused as a single specimen, and can easily correct the vision coordinates.

An object of the present invention is to provide a laser processing apparatus correction system and a laser processing apparatus correction method that can accurately match the vision coordinates based on the scanner coordinates, can be reused as a single specimen, and can easily correct the vision coordinates it is to do

All of the above and other objects of the present invention can be achieved by the present invention described below.

One aspect of the present invention relates to a laser processing machine calibration system.

According to one embodiment, the laser processing apparatus calibration system is a profile module comprising a first light source, and an optical sensor for measuring the surface of the specimen with the reflected light reflected from the first light source from the specimen; a vision module including a second light source and an image collecting unit configured to collect an image of the specimen as the image light reflected from the specimen from the second light source; a laser processing module including a guide laser output unit for displaying the x-axis and y-axis of the scanner module, and a processing laser output unit for processing; and a lens unit and a mirror unit for guiding and reflecting paths of the first light source, reflected light, image light, guide laser, and processing laser, respectively, a stage on which the specimen is located, and a stage for moving the stage in the x-axis and y-axis directions and a scanner module including a driving unit and an axis compensating unit for rotating the axis of the scanner module, and guiding the first light source, reflected light, image light, guide laser, and processing laser substantially coaxially.

A grid pattern is formed on the specimen, and the grid pattern includes a first pattern arranged at a first interval (α _x ) and a second pattern arranged at a second interval (α _y ) perpendicular to the first pattern. can do.

The specimen includes a specimen intersection point at which the first pattern and the second pattern intersect, the specimen center specimen intersection coordinate is (0, 0), and other specimen intersection coordinates are (x × α _x , y × α _y ) can be defined as

(where x is the x-th (x is positive) specimen intersection in the positive x-axis direction from the central specimen intersection or the -x-th (x is negative) specimen intersection in the negative direction, and y is y at the central specimen intersection) The y-th (y is positive) specimen intersection in the positive direction of the axis or -y-th (y is negative) specimen intersection in the negative direction).

The specimen may have a metal pattern formed on a substrate.

The substrate may be quartz or glass.

The pattern may include forming a metal film on the substrate; and forming a pattern on the metal film by exposure.

The pattern may have a width of 100 μm or less.

The specimen image may be obtained by collecting image light in the horizontal β _x and vertical β _y regions on the specimen.

The specimen image may have M×N image intersection points formed by the first pattern and the second pattern on the specimen image (wherein M and N are odd natural numbers).

In the specimen image, the unit coordinates of the central image intersection point may be defined as g(0,0), and the unit coordinates of other image intersection points may be defined as g(i,j).

(the i means the i-th (i is positive) image intersection in the positive x-axis direction at the central image intersection or the -i-th (i is negative) image intersection in the negative direction, and j is the y in the central image intersection. The jth (j is positive) image intersection in the positive direction or -jth (j is negative) image intersection in the negative direction).

The first interval α _x may satisfy Equation 1-1 below, and the second interval α _y may satisfy Equation 1-2 below.

[Equation 1-1]

1st interval (α _x ) = β _x /(M+1)

(The above M is an odd natural number)

[Equation 1-2]

2nd interval (α _y ) =β _y /(N+1)

(The above N is an odd natural number).

The profile module fixes one of the x-axis and y-axis and repeats the process of measuring the surface in the other axial direction to collect height information of the surface of the specimen, and from the height information, the first pattern direction on the specimen, the second 2 The pattern direction and origin can be calculated.

The stage driving unit moves the specimen so that the origin of the specimen calculated by the profile module coincides with the origin of the scanner module, and the axis correction unit has a first pattern direction and a second pattern direction of the specimen calculated by the profile module. The first correction may be performed by rotating the axis of the scanner module to coincide with the y-axis and the x-axis of the scanner module, respectively.

The vision module includes a central image that is a specimen image including the origin of the axis of the scanner module in the center, a plurality of adjacent images that are specimen images immediately adjacent to the central image, and a plurality of adjacent images that are another specimen image immediately adjacent to the adjacent image. A specimen image of the entire specimen is collected by acquiring an adjacent image, and the central image and the adjacent image may be horizontal β _x and vertical β _y images.

The unit coordinates of the specimen image of the central image may be defined as I(0,0), and the unit coordinates of the specimen image of the adjacent image may be defined as I(r,s).

(wherein r means an r-th (r is positive) specimen image in the positive x-axis direction in the central image or an -r-th (r is negative) specimen image in the negative direction, and s is the y-axis positive in the central image. means the sth (s is positive) specimen image in the direction of or -sth (s is negative) specimen image in the negative direction).

The laser processing apparatus correction system may further include a data processing module, and the data processing module may extract an image intersection point from the specimen image or the image of the entire specimen through a deep learning model.

The data processing module calculates the image intersection coordinates (i × α _x + r × β _x , j) based on the specimen image unit coordinates (I(r,s)) and the image intersection unit coordinates (g(i,j)). It is possible to obtain a transformation function that transforms into ×α _x + s × β _x ).

The data processing module may generate a correction table for the entire coordinates of the scan area from the conversion function, and correct the image collected by the image collection unit of the vision module in real time with the correction table.

Another aspect of the present invention relates to a method for calibrating a laser processing apparatus.

In one embodiment, the laser processing apparatus calibration method includes a first pattern arranged at a first interval (α _x ) on a substrate including quartz or glass and a second pattern perpendicular to the first pattern Preparing a specimen on which a metal grid pattern including a second pattern arranged at intervals (α _y ) is formed; Positioning the specimen on the stage of the scanner module so that the origin, the first pattern direction, and the second pattern direction of the specimen substantially coincide with the origin, y-axis, and x-axis of the scanner module using the guide laser of the laser processing module step; collecting, by a profile module, height information of the surface of the specimen, and calculating a first pattern direction, a second pattern direction, and an origin on the specimen; moving the specimen so that an origin of the specimen calculated by the profile module coincides with the origin of the scanner module by a stage driving unit of the scanner module; rotating, by the axis correction unit of the scanner module, the axis of the scanner module such that the first pattern direction and the second pattern direction of the specimen calculated by the profile module coincide with the y-axis and the x-axis of the scanner module; The vision module includes a central image that is a specimen image including the origin of the axis of the scanner module in the center, a plurality of adjacent images that are specimen images immediately adjacent to the central image, and a plurality of adjacent images that are another specimen image immediately adjacent to the adjacent image. acquiring an image to collect a specimen image of the entire specimen; extracting, by a data processing module, an image intersection point from the specimen image or the image of the entire specimen through a deep learning model; The data processing module is a regression analysis model that converts the image intersection coordinates into (i × α _x + r × β _x , j × α _x + s × β _x ) based on the specimen image unit coordinates and the image intersection point unit coordinates. obtaining; and correcting, by the data processing module, the image collected by the vision module in real time with the regression analysis model.

The present invention can accurately match the vision coordinates based on the scanner coordinates, and can be reused as a single specimen, and provides a laser processing apparatus correction system and a laser processing apparatus correction method for easy correction of vision coordinates has

1 is a schematic block diagram of a laser processing apparatus calibration system according to an embodiment of the present invention.
2 is a block diagram showing a path along which light travels between the components of a laser processing apparatus according to an embodiment of the present invention.
3 is a schematic diagram of a laser processing apparatus according to an embodiment of the present invention.
4 and 5 are diagrams illustrating a process in which the profile module measures the surface of the specimen in one embodiment of the present invention.
6 illustrates a process of matching the origin through the x-axis movement and y-axis movement of the specimen and the axial rotation of the scanner module based on the surface information of the specimen measured by the profile module in one embodiment of the present invention.
7 is a view for explaining the unit coordinates of the image intersection point in the specimen image in one embodiment of the present invention.
8 and 9 are diagrams for explaining a specimen image including a central image and an adjacent image in one embodiment of the present invention.
10 is a view for explaining the step of extracting the image intersection point from the specimen image in one embodiment of the present invention.
11 is a diagram for explaining a difference between an image intersection point of an adjacent image and a specimen intersection point of an actual specimen in one embodiment of the present invention.
12 is a view for explaining a regression analysis model for converting the corresponding image intersection coordinates based on the specimen image unit coordinates and the image intersection point unit coordinates.
13 is a flow chart of a laser processing apparatus correction method according to an embodiment of the present invention.

Hereinafter, specific embodiments of the present application will be described in more detail with reference to the accompanying drawings. However, the technology disclosed in the present application is not limited to the embodiments described herein and may be embodied in other forms.

However, the embodiments introduced herein are provided so that the disclosed content may be thorough and complete, and the spirit of the present application may be sufficiently conveyed to those skilled in the art. In the drawings, in order to clearly express the components of each device, the sizes of the components such as width and thickness are somewhat enlarged. In addition, although only some of the components are illustrated for convenience of description, those skilled in the art will be able to easily grasp the remaining parts of the components.

In the description of the drawings as a whole, it has been described from an observer's point of view, and when an element is referred to as being located above or below another element, it means that the element is located directly above or below another element, or an additional element between them. includes all meanings that may be interposed. In addition, those of ordinary skill in the relevant field will be able to implement the idea of the present application in various other forms without departing from the technical spirit of the present application. And, the same reference numerals on the plurality of drawings refer to elements that are substantially the same as each other.

On the other hand, the expression in the singular described in the present application should be understood to include a plural expression unless the context clearly indicates otherwise, and terms such as 'comprise' or 'have' are used to describe features, numbers, steps, It is intended to designate the presence of an action, component, part, or combination thereof, but precludes the possibility of the presence or addition of one or more other features or numbers, steps, actions, components, parts, or combinations thereof. should be understood as not

In addition, in this specification, 'X to Y' representing a range means 'X or more and Y or less'.

Throughout the specification, when a part is 'connected' to another part, this includes not only a case in which it is directly connected, but also a case in which it is 'electrically connected' with another element interposed therebetween. In addition, when a part 'includes' a certain component, this means that other components may be further included, rather than excluding other components, unless otherwise stated.

In this specification, a "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. In addition, one unit may be implemented using two or more hardware, and two or more units may be implemented by one hardware. Meanwhile, '~ unit' is not limited to software or hardware, and '~ unit' may be configured to be in an addressable storage medium or may be configured to reproduce one or more processors. Thus, as an example, '~' denotes components such as software components, object-oriented software components, class components, and task components, and processes, functions, properties, and procedures. , subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. The functions provided in the components and '~ units' may be combined into a smaller number of components and '~ units' or further separated into additional components and '~ units'. In addition, components and '~ units' may be implemented to play one or more CPUs in a device or secure multimedia card.

In addition, in the present specification, the 'specimen intersection' means an intersection point at which the first pattern and the second pattern on the actual specimen intersect.

In addition, in the present specification, an 'image intersection' means an intersection point at which the first pattern and the second pattern of the specimen on the image in the specimen image intersect.

In addition, in this specification, the 'image intersection unit coordinate' represents the number of the image intersection in each specimen image in the x-axis or y-axis direction relative to the central image intersection (central image intersection among image intersection points). It is the unit coordinate for

In addition, in the present specification, the 'origin of the specimen' may mean a specimen intersection point closest to the origin of the scanner module (the intersection of the x-axis and y-axis of the scanner module) based on the height information of the specimen surface of the profile module.

In addition, in the present specification, the 'unit coordinates of the specimen image' are unit coordinates for expressing the position in the x-axis or y-axis direction relative to the central image.

Laser Processing Equipment Calibration System

A laser processing apparatus correction system according to an embodiment of the present invention will be described with reference to FIGS. 1 to 3 . 1 is a block diagram of a laser processing apparatus correction system according to an embodiment of the present invention is simply displayed, Figure 2 is a block between the configuration of the laser processing apparatus according to an embodiment of the present invention, the path that light moves It is indicated with a diagram, and Figure 3 is a simplified view of a laser processing apparatus according to an embodiment of the present invention.

1 to 3, the present invention laser processing apparatus correction system is a first light source 110, and the first light source 110 is an optical sensor 130 for measuring the surface of the specimen with the reflected light reflected from the specimen. Profile module 100 including; a vision module 200 including a second light source 210 and an image collecting unit 230 for collecting an image of the specimen as the image light reflected from the specimen from the second light source; A laser processing module 300 including a guide laser output unit 310 for displaying the x-axis and y-axis of the scanner module 400, and a processing laser output unit 330 for processing; and a lens unit 410 and a mirror unit 420 for guiding and reflecting paths of the first light source, reflected light, image light, guide laser, and processing laser, respectively, a stage 430 where the specimen is positioned, and the stage 430 ) including a stage driving unit 440 for moving the x-axis and y-axis direction and an axis correction unit 450 for rotating the axis of the scanner module, and substantially coaxial the first light source, reflected light, image light, guide laser, and processing laser. It includes a; scanner module 400 for guiding to.

A grid pattern is formed on the specimen, and the grid pattern includes a first pattern arranged at a first interval (α _x ) and a second pattern arranged at a second interval (α _y ) perpendicular to the first pattern. can do. The specimen serves as a reference for correcting the distorted image of the vision module 200 in the laser processing apparatus correction system of the present invention.

The specimen includes a specimen intersection point at which the first pattern and the second pattern intersect, the specimen center specimen intersection coordinate is (0, 0), and other specimen intersection coordinates are (x × α _x , y × α _y ) can be defined as (where x is the x-th (x is positive) specimen intersection in the positive x-axis direction from the central specimen intersection or the -x-th (x is negative) specimen intersection in the negative direction, and y is y at the central specimen intersection) The y-th (y is positive) specimen intersection in the positive direction of the axis or -y-th (y is negative) specimen intersection in the negative direction).

The specimen may have a metal pattern formed on a substrate, and the substrate may be made of quartz or glass. Specifically, the pattern may include: forming a metal film on the substrate; and forming a pattern on the metal film by exposure. A person skilled in the art may select a method suitable for the purpose of the present invention for the step of forming the metal film and the step of exposing, and may be appropriately applied according to the environment.

The pattern may have a width of 100 μm or less. In the above width range, there is an effect that the correction reliability is improved. Existing laser processing equipment calibration involves a process of directly processing a specimen, but these specimens have a line width of 300 μm, so calibration reliability is not secured. In addition, since the specimen in the laser processing apparatus correction system of the present invention can be reused, the process is simplified and the cost is reduced compared to the conventional preparation of the correction specimen for each correction.

The profile module 100 may apply a laser as the first light source 110 , and the type of laser may be appropriately applied by those skilled in the art according to the pattern height of the specimen, the driving environment, and the like. As for the laser of the first light source 110 , the optical fiber end 113 becomes a virtual point light source of the laser light source 110 through the optical fiber 11 . The emitted laser becomes a parallel beam by the collimator 150 , and only the linearly polarized beam vibrating only in the polarization axis direction by the polarizer 170 passes. The linearly polarized beam passes through the optical splitter 270 of the vision module 200 and the optical splitter 370 of the laser processing module 300 and is reflected by the mirror 420 of the scanner module 400, which is a lens The scan area on the specimen is focused through the unit 410 . The reflected light reflected in the scan area passes through the lens unit 410, the mirror unit 420, the light splitters 370 and 270, the polarizer 170, and the collimator 150 again, and is focused on the point light source position 133, and is then focused on the optical fiber. It is input to the photosensor 130 through 131 and the intensity of light according to the wavelength is measured.

The profile module fixes one of the x-axis and y-axis and repeats the process of measuring the surface in the other axial direction to collect height information of the surface of the specimen, and from the height information, the first pattern direction on the specimen, the second 2 The pattern direction and origin can be calculated. (refer to FIGS. 4 to 6)

4 shows a process of collecting height information of the specimen surface by fixing one of the x-axis and y-axis of the profile module 100, and repeating the process of measuring the surface by scanning in the other axial direction, Figure 5 (A) is the y-coordinate is fixed and the surface height is measured in the x-axis direction, Figure 5 (B) is the x-coordinate is fixed and the surface height is measured in the y-axis direction, Figure 5 (C) is It shows that direction information of the first pattern 22-1 and the second pattern 23-1 is calculated based on the collected information, and FIG. 6(A) shows the calculated first pattern 22-1. ) and the direction information of the second pattern 23-1, the intersection of the extension lines of the first pattern and the second pattern is calculated as the origin.

The stage driving unit 440 moves the specimen so that the origin of the specimen calculated by the profile module 100 coincides with the origin of the scanner module 400 , and the axis correction unit 450 is the profile module 100 . The first correction may be performed by rotating the axis of the scanner module 400 so that the first pattern direction and the second pattern direction of the specimen calculated in , respectively, coincide with the y-axis and the x-axis of the scanner module.

The stage driver 440 may apply a method of moving the stage on which the specimen is located in the x-axis and/or the y-axis.

Referring to FIG. 6 , the process (A) of calculating the first pattern direction, the second pattern direction and the origin on the specimen, the process of matching the origin of the specimen (B), and the process of rotating the axis of the scanner module 400 (C) ), the first pattern, second pattern, and origin of the specimen and the y-axis, x-axis, and origin of the scanner module can be exactly matched.

In a specific embodiment, the first correction is performed more efficiently by minimizing the error between the origin of the specimen and the origin of the scanner module 400 by using the guide laser output unit 310 of the laser processing module 300 before the first correction. can do. Specifically, the guide laser output unit 310 transmits visible light to the specimen through the mirror unit 420 and the lens unit 410 of the scanner module 400 through the collimator 350 and the light splitter 370 . The x-axis and y-axis of the scanner module 400 can be displayed on the screen, and through this, the origin of the specimen and the origin of the scanner module can be roughly matched. Accordingly, the first correction can be performed more efficiently.

After the first correction, the vision module 200 may apply visible light as the light source 210 , and the light source 210 may be installed non-coaxially with the light source 110 . The light source 210 is reflected in the scan area of the surface of the specimen, and the reflected image light is the lens unit 410 , the mirror unit 420 , the light splitter 370 of the laser processing module 300 , and the vision module 200 . Image light may be collected by the image collecting unit 230 through the light splitter 270 and the imaging lens 250 of the . The image collection unit 230 may be, for example, a CCD camera.

The image collection unit 230 may generate a specimen image by collecting the image light. The specimen image may be a collection of image light in the horizontal β _x and vertical β _y regions on the specimen, and in a specific embodiment, the specimen image is the maximum region ( FOV, Field of View).

The specimen image may have M×N image intersection points formed by the first pattern and the second pattern on the specimen image (wherein M and N are odd natural numbers). M and N may each independently be 1 to 9, specifically 1 to 7, more specifically 3 to 5.

Referring to FIG. 7 , in the specimen image, the unit coordinates of the central image intersection point may be defined as g(0,0), and the unit coordinates of other image intersection points may be defined as g(i,j).

The i means the i-th (i is positive) image intersection in the positive x-axis direction at the central image intersection (g(0,0)) or the -i-th (i is negative) image intersection in the negative direction, and j may mean a j-th (j is positive) image intersection in a positive y-axis direction from the central image intersection, or a -j-th (j is negative) image intersection in a negative direction.

In an embodiment, the first interval (α _x ) may satisfy Equation 1-1 below, and the second interval (α _y ) may satisfy Equation 1-2 below. At this time, the calibration reliability of the laser processing apparatus calibration system is maximized.

[Equation 1-1]

1st interval (α _x ) = β _x /(M+1)

(The above M is an odd natural number)

[Equation 1-2]

2nd interval (α _y ) =β _y /(N+1)

(The above N is an odd natural number).

8 and 9, the vision module 200 includes a central image that is a specimen image including the origin of the axis of the scanner module 400 in the center, a plurality of adjacent images that are specimen images immediately adjacent to the central image, and a plurality of adjacent images that are another specimen image immediately adjacent to the adjacent image are acquired to collect specimen images of the entire specimen, and the central image and adjacent image may be horizontal β _x and vertical β _y images. . Referring to FIG. 9 , it can be seen that the image intersection points in the adjacent images 61-1, 62-1, 63-1, and 64-1 that are distant from the central image are different from the specimen intersection points of the actual specimen. Through the following process, the coordinates of the image intersection points can be matched with respect to the specimen intersection point.

Referring to FIG. 9 , the unit coordinates of the specimen image of the central image may be defined as I(0,0), and the unit coordinates of the specimen image of the adjacent image may be defined as I(r,s).

(wherein r means an r-th (r is positive) specimen image in the positive x-axis direction in the central image or an -r-th (r is negative) specimen image in the negative direction, and s is the y-axis positive in the central image. means the sth (s is positive) specimen image in the direction of or -sth (s is negative) specimen image in the negative direction).

The present invention laser processing apparatus correction system may further include a data processing module.

Referring to FIG. 10 , the data processing module may extract an image intersection point from the specimen image or the image of the entire specimen through a deep learning model. A deep learning model, CoordNet 70, can be used for the data processing module to add the image intersection points from the specimen image, and the learning process of Codenet is when the specimen image is put into the model, all pixel values except for the image intersection points are The model parameters are determined using a mathematically optimized mechanism (Optimization) so that a binary image of the same size as 0, the pixel value of the image intersection is 1 or 255. Specifically, the neural layer used here is based on a well-known 2d convolution layer that is powerful for image analysis. Once the learning is completed, the codenet 70 displays only the expected image intersection points for any grid pattern image and outputs the image. Since the model parameters have already been tuned in the pre-training phase, there are no parameters to be tuned by the user in the actual calibration phase, and the calibration process can be automated. Hereinafter, the conversion net (83, see FIG. 12) can be automatically used in the calibration process only after pre-learning in a similar process.

In FIG. 11 , the image intersection point on the specimen image shows an error with the specimen intersection point of the actual specimen. The farther away from the central image, the larger the error tends to be.

Accordingly, the data processing module calculates the corresponding image intersection coordinates (i × α _x + r × β _x , j × α _x + s × β _x ) can be obtained. In the present invention, the process of obtaining the transformation function is performed by constructing a non-linear model using a non-linear regression deep learning model, thereby increasing the accuracy without requiring deep knowledge of the optical model. We tried to minimize model tuning.

12 is a nonlinear transformation function, a deep learning model for finding T (image intersection -> specimen intersection), Transform Net 83 . The input data 80 and 81 represent the distorted intersection points recognized on the image, and the output data 85 and 86 are the predicted values by the transform net 83 and T (image intersection -> specimen intersection) in learning. Both the input node 82 and the output node 84 are composed of two nodes. The transformation net 83 is a regression model that can be configured in various forms based on a fully connected layer. For example, it may be composed of a combination of a fully connected layer and a one-dimensional convolutional layer. The error of the model is defined by Equation 2 below, which is the Mean Squared Error (MSE) between the predicted value of the deep neural network model for the input data 80 and 81 and the undistorted coordinates of the pattern intersection points on the calibration specimen.

[Equation 2]

A result of the entire calibration process may be a lookup table (LUT) for the x and y coordinates of each scan position. Once the LUT is calculated, the calibration process can be terminated, the real-time image of the image collection unit of the vision module can be corrected, and a corrected panoramic image can be acquired for easy processing of objects with a size of several tens of cm.

For specimen processing, the processing laser output unit 430 of the laser processing module 400 may output an IR laser of about 1030 nm or a green laser of 515 nm or less. The output laser can be processed by passing through the collimator 350 and the light splitter 370 to focus on the processing position through the mirror unit 420 and the lens unit 410 of the scanner module 400 .

Another aspect of the present invention relates to a method for calibrating a laser processing apparatus.

Referring to Figure 13, in one embodiment, the laser processing apparatus calibration method is a first pattern and the first arranged at a first interval (α _x ) on a substrate including quartz (quartz) or glass (glass) Preparing a specimen in which a metal grid pattern including a second pattern arranged at a second interval (α _y ) perpendicular to the pattern is formed; Positioning the specimen on the stage of the scanner module so that the origin, the first pattern direction, and the second pattern direction of the specimen substantially coincide with the origin, y-axis, and x-axis of the scanner module using the guide laser of the laser processing module step; collecting, by a profile module, height information of the surface of the specimen, and calculating a first pattern direction, a second pattern direction, and an origin on the specimen; moving the specimen so that an origin of the specimen calculated by the profile module coincides with the origin of the scanner module by a stage driving unit of the scanner module; rotating, by the axis correction unit of the scanner module, the axis of the scanner module such that the first pattern direction and the second pattern direction of the specimen calculated by the profile module coincide with the y-axis and the x-axis of the scanner module; The vision module includes a central image that is a specimen image including the origin of the axis of the scanner module in the center, a plurality of adjacent images that are specimen images immediately adjacent to the central image, and a plurality of adjacent images that are another specimen image immediately adjacent to the adjacent image. acquiring an image to collect a specimen image of the entire specimen; extracting, by a data processing module, an image intersection point from the specimen image or the image of the entire specimen through a deep learning model; The data processing module is a regression analysis model that converts the image intersection coordinates into (i × α _x + r × β _x , j × α _x + s × β _x ) based on the specimen image unit coordinates and the image intersection point unit coordinates. obtaining; and correcting, by the data processing module, the image collected by the vision module in real time with the regression analysis model.

The definitions and procedures of terms are substantially the same as the laser processing apparatus calibration system which is one aspect of the present invention.

Although the specific embodiments of the present invention have been described above, the present invention is not limited to the above specific embodiments, but may be manufactured in various different forms, and those of ordinary skill in the art to which the present invention pertains will appreciate the technical spirit of the present invention. However, it will be understood that the invention may be embodied in other specific forms without changing essential features. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive.

100: profile module 110: first light source
130: optical sensor 150: collimator
170: polarizer 200: vision module
210: second light source 230: image collecting unit
250: imaging lens 270: splitter
300: laser processing module 310: guide laser output unit
330: processing laser output unit 350: collimator
370: optical splitter 400: scanner module
410: lens unit 420: mirror unit
430: stage 440: stage driving unit
450: Chukbo government

Claims (19)

a profile module including a first light source and an optical sensor for measuring the surface of the specimen with reflected light reflected from the first light source;
a vision module including a second light source and an image collecting unit configured to collect an image of the specimen as the image light reflected from the specimen from the second light source;
a laser processing module including a guide laser output unit for displaying the x-axis and y-axis of the scanner module, and a processing laser output unit for processing; and
A lens unit and a mirror unit for guiding and reflecting paths of the first light source, reflected light, image light, guide laser, and processing laser, respectively, a stage on which the specimen is located, and a stage driving unit for moving the stage in the x-axis and y-axis directions and a scanner module including an axis correction unit for rotating the axis of the scanner module, and guiding the first light source, reflected light, image light, guide laser, and processing laser substantially coaxially;
A laser processing device calibration system comprising a.
According to claim 1,
The specimen is formed with a grid pattern,
The grid pattern is
A laser processing apparatus correction system comprising a first pattern arranged at a first interval (α _x ) and a second pattern arranged at a second interval (α _y ) perpendicular to the first pattern.
3. The method of claim 2,
The specimen includes a specimen intersection point at which the first pattern and the second pattern intersect,
The coordinates of the specimen intersection point at the center of the specimen are (0,0),
The other specimen intersection coordinates are defined by (x×α _x , y×α _y ) laser processing device calibration system:
(where x is the x-th (x is positive) specimen intersection in the positive x-axis direction from the central specimen intersection or the -x-th (x is negative) specimen intersection in the negative direction, and y is y at the central specimen intersection) The y-th (y is positive) specimen intersection in the positive direction of the axis or -y-th (y is negative) specimen intersection in the negative direction).
4. The method of claim 3,
The specimen is a laser processing apparatus calibration system in which a metal pattern is formed on a substrate.
5. The method of claim 4,
The substrate is a quartz (quartz) or glass (glass) laser processing apparatus calibration system.
6. The method of claim 5,
the pattern is
forming a metal film on the substrate; and
forming a pattern on the metal film by exposure;
A laser processing device calibration system formed by
7. The method of claim 6,
The pattern is a laser processing device calibration system having a width of 100 μm or less.
8. The method of claim 7,
The specimen image is a laser processing apparatus calibration system that collects image light in the horizontal β _x and vertical β _y regions on the specimen.
9. The method of claim 8,
The specimen image is a laser processing apparatus correction system in which the image intersection points formed by the first pattern and the second pattern on the specimen image are M×N (wherein M and N are odd natural numbers).
10. The method of claim 9,
In the specimen image, the unit coordinates of the central image intersection point are g(0,0), and the unit coordinates of other image intersection points are defined as g(i,j).
(the i means the i-th (i is positive) image intersection in the positive x-axis direction at the central image intersection or the -i-th (i is negative) image intersection in the negative direction, and j is the y in the central image intersection. The jth (j is positive) image intersection in the positive direction or -jth (j is negative) image intersection in the negative direction).
11. The method of claim 10,
The first interval (α _x ) satisfies the following Equation 1-1, and the second interval (α _y ) is a laser processing apparatus correction system that satisfies the following Equation 1-2:
[Equation 1-1]
1st interval (α _x ) = β _x /(M+1)
(The above M is an odd natural number)
[Equation 1-2]
2nd interval (α _y ) =β _y /(N+1)
(The above N is an odd natural number).
12. The method of claim 11,
The profile module collects height information of the surface of the specimen by fixing one of the x-axis and the y-axis and repeating the process of measuring the surface in the other axial direction,
A laser processing apparatus correction system for calculating a first pattern direction, a second pattern direction, and an origin on the specimen from the height information.
13. The method of claim 12,
The stage driving unit moves the specimen so that the origin of the specimen calculated by the profile module coincides with the origin of the scanner module,
The axis correction unit rotates the axis of the scanner module so that the first pattern direction and the second pattern direction of the specimen calculated by the profile module coincide with the y-axis and the x-axis of the scanner module, respectively,
A laser processing apparatus calibration system for performing the first calibration.
14. The method of claim 13,
The vision module includes a central image that is a specimen image including the origin of the axis of the scanner module in the center, a plurality of adjacent images that are specimen images immediately adjacent to the central image, and a plurality of adjacent images that are another specimen image immediately adjacent to the adjacent image. By acquiring an adjacent image, the specimen image of the entire specimen is collected,
The central image and the adjacent image are images of horizontal β _x and vertical β _y sizes.
15. The method of claim 14,
The unit coordinate of the specimen image of the central image is I(0,0),
The laser processing device calibration system, wherein the unit coordinates of the specimen image of the adjacent image are defined as I(r,s):
(wherein r means an r-th (r is positive) specimen image in the positive x-axis direction in the central image or an -r-th (r is negative) specimen image in the negative direction, and s is the y-axis positive in the central image. means the sth (s is positive) specimen image in the direction of or -sth (s is negative) specimen image in the negative direction).
16. The method of claim 15,
The laser processing device calibration system further comprises a data processing module,
The data processing module is a laser processing apparatus correction system for extracting the image intersection point through a deep learning model from the image of the specimen or the entire specimen.
17. The method of claim 16,
The data processing module calculates the image intersection coordinates (i × α _x + r × β _x , j) based on the specimen image unit coordinates (I(r,s)) and the image intersection unit coordinates (g(i,j)). ×α _x + s × β _x ) to obtain a conversion function that converts a laser processing equipment calibration system.
18. The method of claim 17,
The data processing module generates a correction table for the entire coordinates of the scan area from the conversion function, and corrects the image collected by the image collecting unit of the vision module with the correction table in real time.
A first pattern arranged at a first interval (α _x ) and a second pattern arranged at a second interval (α _y ) perpendicular to the first pattern on a substrate including quartz or glass Preparing a specimen including a metal grid pattern formed thereon;
Positioning the specimen on the stage of the scanner module so that the origin, the first pattern direction, and the second pattern direction of the specimen substantially coincide with the origin, y-axis, and x-axis of the scanner module using the guide laser of the laser processing module step;
collecting, by a profile module, height information of the surface of the specimen, and calculating a first pattern direction, a second pattern direction, and an origin on the specimen;
moving the specimen so that an origin of the specimen calculated by the profile module coincides with the origin of the scanner module by a stage driving unit of the scanner module;
rotating, by the axis correction unit of the scanner module, the axis of the scanner module such that the first pattern direction and the second pattern direction of the specimen calculated by the profile module coincide with the y-axis and the x-axis of the scanner module;
The vision module includes a central image that is a specimen image including the origin of the axis of the scanner module in the center, a plurality of adjacent images that are specimen images immediately adjacent to the central image, and a plurality of adjacent images that are another specimen image immediately adjacent to the adjacent image. acquiring an image to collect a specimen image of the entire specimen;
extracting, by a data processing module, an image intersection point from the specimen image or the image of the entire specimen through a deep learning model;
The data processing module is a regression analysis model that converts the image intersection coordinates into (i × α _x + r × β _x , j × α _x + s × β _x ) based on the specimen image unit coordinates and the image intersection point unit coordinates. obtaining; and
correcting, by the data processing module, the image collected by the vision module in real time with the regression analysis model;
A laser processing apparatus calibration method comprising a.

Images