Laser stripping device, information processing method, and recording medium capable of reading and recording programs.

Abstract

A laser stripping apparatus is used to irradiate a workpiece, including a substrate and a release layer formed on the substrate, with laser light and to peel the release layer from the substrate. The apparatus includes an imaging device to capture an image of the workpiece irradiated by the laser light; and a processing unit to process the image captured by the imaging device. The processing unit acquires the image of the workpiece captured by the imaging device; by inputting the acquired image of the workpiece into a learning model, it derives information about the state of the release layer; the learning model outputs information about the state of the release layer of the workpiece when the image of the workpiece is input; and outputs the derived information about the state of the release layer.

Description

[Technical Field]

[0001] This invention relates to a laser stripping device, an information processing method, and a recording medium capable of reading and recording programs. [Previous Technology]

[0002] A release layer (workpiece) formed on a substrate is known, and a laser lift-off apparatus (e.g., Patent Document 1) is used to peel the release layer from the substrate by irradiating it with a laser beam from the substrate side. Patent Document 1 discloses a laser lift-off apparatus (laser lift-off device) that continuously changes the irradiation area of ​​the workpiece by the laser beam, simultaneously irradiating adjacent irradiation areas to overlap. [Prior Art Documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2012-28740 [Summary of the Invention]

[0004] [The problem that the invention aims to solve]

[0005] However, the laser stripping apparatus of Patent Document 1 does not take into account the efficient extraction of information about the state of the stripping layer irradiated by laser light.

[0006] In view of this situation, the object of the present invention is to provide a laser stripping apparatus, etc., that can efficiently extract information about the state of the stripping layer irradiated by laser light. [Means for solving the problem]

[0007] Regarding the laser stripping apparatus of this type, for irradiating a workpiece including a substrate and a release layer formed on the substrate with laser light and stripping the release layer from the substrate, the laser stripping apparatus includes: an imaging device for capturing images of the workpiece irradiated by the laser light; and a processing unit for processing based on the image captured by the imaging device; wherein the processing unit acquires the image of the workpiece captured by the imaging device; by inputting the acquired image of the workpiece into a learning model, it derives information about the state of the release layer, and the learning model outputs information about the state of the release layer of the workpiece when the image of the workpiece is input; and outputs the derived information about the state of the release layer.

[0008] The information processing method for this sample performs the following processing in a computer: acquiring an image of the workpiece on which a release layer is formed on a substrate by irradiating it with laser light; by inputting the acquired image of the workpiece into a learning model, deriving information about the state of the release layer; the learning model outputs information about the state of the release layer of the workpiece when the image of the workpiece is input; and outputting the derived information about the state of the release layer.

[0009] The program for this embodiment performs the following processing in a computer: acquiring an image of the workpiece on which a release layer is formed on a substrate, by irradiating the workpiece with laser light; by inputting the acquired image of the workpiece into a learning model, deriving information about the state of the release layer; the learning model outputs information about the state of the release layer of the workpiece when the image of the workpiece is input; and outputting the derived information about the state of the release layer. [Effects of the Invention]

[0010] According to the present invention, a laser stripping apparatus or the like can be provided to efficiently extract information about the state of the stripping layer irradiated by laser light.

Implementation Method

[0012] (Example 1) Hereinafter, embodiments of the present invention will be described. Figure 1 is a diagram showing an example of the configuration of the laser lift-off apparatus 1 according to Embodiment 1. Figure 2 is a diagram showing an example of the configuration of the information processing apparatus 9 included in the laser lift-off apparatus 1. The laser lift-off apparatus 1 is an apparatus that irradiates a laser beam from the substrate 81 side and peels off the release layer 82 from the substrate 81, for example, and is used in the manufacture of organic electroluminescence (EL) displays. The substrate 81 is configured as a transparent substrate 81, such as a glass substrate 81, and the workpiece 8 is subjected to laser lift-off (LLO) processing by means of the substrate 81 and the release layer 82 formed on the underside of the substrate 81.

[0013] As illustrated in this embodiment, in the XYZ three-dimensional rectangular coordinate system, the Z direction is the vertical direction, which is perpendicular to the substrate 81. The XY plane is a plane parallel to the surface on which the release layer 82 of the substrate 81 is formed. For example, the X direction is the length direction of the rectangular substrate 81, and the Y direction is the width direction of the substrate 81. When using the Θ-axis stage 71 that can rotate from 0° to 90° around the Z-axis, the X direction can become the width direction of the substrate 81, and the Y direction can become the length direction of the substrate 81.

[0014] The laser stripping apparatus 1 includes a laser optical system 11, a laser irradiation chamber 7, and an information processing device 9. The laser irradiation chamber 7 houses a base 72 and a stage 71 disposed on the base 72. In the laser stripping apparatus 1, the stage 71 transports the substrate 81 in the +X direction while simultaneously irradiating the stripping layer 82 with laser light.

[0015] The laser optical system 11 includes a laser light source 2, an attenuator 3, a polarization ratio control unit 4, a beam shaping optical system 5, a projection mirror 61, and a projection lens 63, which emits linear laser light.

[0016] The laser source 2 is a laser generating device that generates pulsed laser light, which is used to irradiate the stripping layer 82. The generated laser light is, for example, gas laser light such as excimer laser light with a center wavelength of 308 nm. Alternatively, the laser light is not limited to excimer laser light, but can also be laser light from a solid-state laser. In the laser source 2, a gas such as xenon is sealed in the chamber, and two resonator mirrors are arranged opposite each other on both sides of the gas. One resonator mirror is a total reflection mirror that reflects all light, and the other resonator mirror is a partial reflection mirror that transmits a portion of the light. The gas light excited by the gas is repeatedly reflected between the resonator mirrors, and the amplified light is released from the resonator mirrors as laser light. The laser source 2 repeatedly releases pulsed laser light with a period of, for example, 500 Hz to 600 Hz. The laser source 2 emits the laser light towards the attenuator 3.

[0017] Attenuators 3 attenuate the incident laser light, adjusting it to a specific energy density (ED). These attenuators 3 have transmittance characteristics, showing the ratio of the emitted laser light to the incident laser light, which is configured to be changed based on signals from a programmable logic controller (PLC). Attenuators 3 are positioned in the optical path diagram from the laser source 2 to the beam shaping optical system 5. Attenuators 3 attenuate the laser light emitted from the laser source 2 according to their transmittance. The energy density (E) emitted from attenuators 3 is the product of the energy density (E0) of the laser light emitted from the laser source 2 and the transmittance (T) of attenuators 3 (E=E0×T). The PLC (control unit) identifies (derives) and changes the transmittance of attenuators 3 to make the energy density emitted from attenuators 3 the optimal energy density.

[0018] A polarization ratio control unit 4 is disposed on the output side of attenuator 3. The polarization ratio control unit 4 is, for example, composed of a 1 / 2 wavelength plate (λ / 2 plate) and a polarizing beam splitter, and changes the polarization ratio of the P-polarized wave and S-polarized wave of the incident laser light. That is, the polarization ratio of the laser light emitted from attenuator 3 is changed by the polarization ratio control unit 4. The polarization ratio control unit 4 is configured to change (variable) the polarization ratio based on a control signal output from a PLC (control unit).

[0019] A laser beam shaping optical system 5, emitted from the polarization ratio control unit 4, shapes the incident laser light to generate a laser beam shape suitable for irradiating the stripping layer 82. The beam shaping optical system 5 generates a linear beam along the Y direction.

[0020] For example, the beam shaping optical system 5 uses a homogenizer composed of a lens array to split a single beam into multiple beams (multiple line beams arranged in the Z direction). After being split into multiple beams, they can be combined by a condenser lens to form a line beam. The beam shaping optical system 5 then outputs the generated (shaped) line laser light to the incident mirror 61.

[0021] The reflection mirror 61 is a rectangular reflector extending in the Y direction, reflecting the laser light generated by the multiple line beams of the beam shaping optical system 5. The reflection mirror 61 is, for example, a dichroic mirror, a partial reflector that transmits a portion of the light. The reflection mirror 61 reflects the line laser light and generates reflected light, while simultaneously generating transmitted light that partially transmits the line laser light. The reflection mirror 61 causes the reflected laser light to irradiate the release layer 82 of the substrate 81, and outputs the transmitted laser light to a pulse measuring device such as a dual-plane phototube 62.

[0022] A projection lens 63 is disposed above the substrate 81. The projection lens 63 includes a plurality of lenses for projecting laser light onto the substrate 81, that is, onto the release layer 82. The projection lens 63 focuses the laser light onto the substrate 81. On the substrate 81, the laser light forms a linear irradiation area along the Y direction. That is, the laser light becomes a line beam on the substrate 81 with the Y direction as its length direction. In addition, the workpiece 8 (substrate 81) is transported in the +X direction while the laser light irradiates the release layer 82. Therefore, the laser light can irradiate a strip-shaped area with a width equal to the length of the irradiation area in the Y direction.

[0023] The linear laser beam illuminating the reflector 61 becomes a beam shape with a wider short axis. That is, after exiting the condenser lens, the short axis width is slightly wider, and the shape is disrupted. The laser beam reflected by the reflector 61 is shaped into a linear laser beam with a short axis width of about 1 / 5 by passing through the projection lens 63.

[0024] The dual-plane phototube 62 is adjacent to the beam shaping optical system 5 and is disposed at the end of the laser optical system 11. It detects the pulse waveform of the laser light emitted from the laser source 2 based on the transmitted light from the transmission mirror 61. The dual-plane phototube 62 outputs (transmits) the detected waveform to the PLC (control device). The laser pulse energy of the laser light is derived based on this pulse waveform.

[0025] The photographic device 64 captures images of the workpiece 8 (substrate 81 and release layer 82) illuminated by laser light, and transmits the captured images of the workpiece 8 to the information processing device 9. For example, the photographic device 64 includes a coaxial incident linear camera (coaxial incident linear camera). This coaxial incident linear camera (photographic device 64) includes a semi-reflective mirror, an illumination device such as an LED (coaxial incident illumination), and a linear camera (camera body). The illumination light reflected by the semi-reflective mirror is configured to be coaxial with the camera body. Therefore, the illuminated and unilluminated parts can be captured with high contrast using the illumination device such as an LED. Although the photographic device 64 is a coaxial incident linear camera in this embodiment, it is not limited to this. In addition to a coaxial incident linear camera, the photographic device 64 may include, for example, a transmissive linear camera, a diffuse reflection linear camera, or a one-way reflection linear camera, or may be composed of a group of multiple linear cameras of these different types.

[0026] The information processing device 9 is communicatively connected to the imaging device 64, acquires an image of the workpiece 8 output from the imaging device 64, outputs information about the state of the peeling layer 82 of the workpiece 8 based on the acquired image of the workpiece 8, and performs processing based on the information to determine whether the peeling of the peeling layer 82 is proceeding normally. That is, the information processing device 9 operates as the main device constituting the peeling observation and monitoring system, which automatically determines whether the peeling layer 82 on the substrate has been peeled by the laser peeling device 1.

[0027] The information processing device 9 may be an edge computer, personal computer, or server device, or it may be a device (control device) that also functions as a control panel or programmable logic controller (PLC) for controlling the operation of the laser stripping device 1. Alternatively, the PLC or the like that controlling the operation of the laser stripping device 1 may be configured as a separate device from the information processing device 9, and the information processing device 9 may be communicatively connected to the PLC or the like (control device) via the communication unit 93. In this case, the information processing device 9 may also be a cloud server, for example, connected to an external network such as the Internet, and communicate with the laser stripping devices 1 (photography device 64, PLC) located at multiple production sites, operating as a central server, and exporting the state of the stripping layer 82 through LLO processing performed separately in the multiple laser stripping devices 1.

[0028] The information processing device 9 includes a processing unit 91, a memory unit 92, a communication unit 93, and an input / output I / F 94, and communicates with the camera device 64 and the PLC through the communication unit 93 or the input / output I / F 94.

[0029] The processing unit 91 includes one or more central processing units (CPU), micro-processors (MPU), graphics processing units (GPU), and other computing processing devices with timing functions. It operates as a control unit and performs processing or control based on the image (image of workpiece 8) captured by the camera device 64 by reading and executing the program P (program product) stored in the memory unit 92.

[0030] The memory unit 92 includes volatile memory areas such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), and flash memory, as well as non-volatile memory areas such as Electrically Removable Programmable Read-Only Memory (EEPROM) or hard disks. The program P (program product) and data referenced during processing are pre-stored in the memory unit 92. The program P stored in the memory unit 92 can also store the program P (program product) read from a readable recording medium by the processing unit 91. Alternatively, the program P (program product) can be downloaded from an external computer (not shown) connected to a communication network (not shown) and stored in the memory unit 92. The physical file of the learning model 913, described later, can also be stored in the memory unit 92, for example, as a module of the program P.

[0031] The communication unit 93 is, for example, a communication module or communication interface conforming to the Ethernet (registered trademark) specification, and the communication unit 93 is connected to an Ethernet cable. The communication unit 93 is not limited to wired cases such as the Ethernet cable, and may also be, for example, a communication interface corresponding to short-range wireless communication networks such as Wi-Fi (registered trademark) and Bluetooth (registered trademark), or long-range wireless communication modules such as 4G and 5G.

[0032] The input / output I / F 94 is, for example, a communication interface conforming to communication standards such as RS232C or USB. The input / output I / F 94 connects to input devices such as a keyboard, and peripheral devices such as a display device 941 such as an LCD screen. The processing unit 91 of the information processing device 9 can be communicatively connected to the camera device 64 through the input / output I / F 94 or the communication unit 93.

[0033] Figure 3 is an explanatory diagram showing an example of a learning model 913 (stripping state determination module). The learning model 913 is, for example, composed of a convolutional neural network (CNN), which is a neural network (NN) that outputs information about the state of the stripping layer 82 of the workpiece 8 contained in the input image, and operates as a stripping state determination model.

[0034] The image of workpiece 8 input into the learning model 913 (peeling state determination module) can also be a block image segmented into specific sizes. That is, as a preprocessing step when using the learning model 913 (peeling state determination module), the image of workpiece 8 (peeling layer 82) captured by the camera device 64 can be segmented into specific sizes, and the segmented block image can be input into the learning model 913. Alternatively, the image of workpiece 8 (peeling layer 82) captured by the camera device 64 can be input into the learning model 913 without performing preprocessing such as segmentation.

[0035] The learning model 913 (peeling state determination module) can prepare training data for the image (problem data) of the workpiece 8 (peeling layer 82) and information (response data) about the state (peeled state or unpeeled state) of the peeling layer 82. Using this training data, a neural network is generated by machine learning. While using the laser peeling device 1, the image of the workpiece 8 can be stored sequentially as operation history data by, for example, a data logger. Information about the state of the peeling layer 82 can be obtained by collecting quality confirmation results from products processed by LLO. Training data for the learning model 913 (peeling state determination model) can be generated using this data.

[0036] When generating the learning model 913 (stripping state determination model), the already learned model can also be used as a classifier. The model can be transferred or fine-tuned using training data adapted to the stripping layer 82 of the aforementioned workpiece 8. The dataset containing the question data and answer data included in the training data used to learn the learning model 913, etc., is synonymous with the dataset containing the input data and output data when using the learning model 913, etc. If it is defined in one dataset, it will naturally apply to other datasets as well.

[0037] Assuming the neural network (stripping state determination model) learned using training data is used as a program module, this program module is part of artificial intelligence software. As described above, the learning model 913 (stripping state determination model) is used in an information processing device 9 including a processing unit 91 and a memory unit 92, and by executing in the information processing device 9 with such computing power, a neural network system is formed. That is, the processing unit 91 of the information processing device 9 performs the calculation of the feature quantities of the image of the workpiece 8 (stripping layer 82) input to the input layer according to the instructions from the learning model 913 (stripping state determination model) stored in the memory unit 92, and outputs information about the state of the stripping layer 82 from the output layer.

[0038] The learning model 913 (stripping state determination model) has an input layer that receives image input, an output layer that outputs information about the state of the stripping layer 82 of the workpiece 8 contained in the image, and an intermediate layer that extracts image features. The input layer has multiple neurons that receive the pixel values ​​of each pixel contained in the image as input, and the input pixel values ​​are passed to the intermediate layer. The intermediate layer has multiple neurons that extract the image features of the image and pass the extracted image features to the output layer. For example, when the learning model 913 (stripping state determination model) is a CNN, the intermediate layer has a structure that interconnects convolutional layers that convolve the pixel values ​​of each pixel input from the input layer with pooling layers that map the pixel values ​​convolved in the convolutional layers, compressing the pixel information of the image while extracting the final image features. The output layer has one or multiple neurons that output information about the state of the stripping layer 82 of the workpiece 8, and outputs information about the state of the stripping layer 82 based on the image features output from the intermediate layer.

[0039] In this embodiment, as an example, the output layer is composed of a soft-mask layer and has two neurons (nodes) corresponding to the state of the peel layer 82 of the workpiece 8. The state of the peel layer 82 includes, for example, a peeled state and an unpeeled state. The neuron corresponding to the peeled state outputs a probability (score) indicating that the peel layer 82 is in the peeled state (peeling is normal), and the neuron corresponding to the unpeeled state outputs a probability (score) indicating that the peel layer 82 is in the unpeeled state (peeling is poor). This probability can be a value from 0 to 1 displaying a specific number of significant digits (e.g., 0.00 to 1.00 when the number of significant digits is 2 decimal places). Alternatively, the learning model 913 may only output the probability (score) indicating that the state of the peel layer 82 contained in the input block image is peeled (peeling is normal). Alternatively, the learning model 913 may also output binary information indicating that the state of the peel layer 82 contained in the input block image is either peeled or unpeeled (peeled state: 1, unpeeled state: 0).

[0040] Although it is assumed that the learning model 913 (peeling state determination model) outputs values ​​(scores) showing the probability of the peeled state and the unpeeled state as information about the state of the peeled layer 82, it is not limited to this. When the image of the workpiece 8 is input, the learning model 913 (peeling state determination model) can also output values ​​(percentages) showing the probability of the peeled state, the unpeeled state, and the soot-generating state (ash) as information about the state of the peeled layer 82 of the workpiece 8. At this time, the training data used to learn the learning model 913 includes images of the peeled state or the unpeeled state, as well as images of the soot-generating state. The output layer of the learning model 913 (peeling state determination model) consists of a soft masking layer containing three neurons corresponding to the unpeeled state and the soot-generating state (ash), and outputs the probability (score) of the state of each peeled layer 82 corresponding to each neuron. Therefore, in each segmented block image, it is possible to identify whether the peeling layer 82 of the workpiece 8 contained in that block image is in a peeled state, an unpeeled state, or a soot-generating state (ash). The processing unit 91 (control unit) of the information processing device 9 can identify the peeled, unpeeled, or soot-generating areas in all areas of the peeling layer 82 of the entire workpiece 8 (base plate) by identifying the state of the peeling layer 82 in these block images.

[0041] Alternatively, the learning model 913 (stripping state determination model) may be composed of, for example, Regions with Convolutional Neural Network (RCNN), Fast RCNN, Faster RCNN, Single Shot Multibook Detector (SSD), YOLO (You Only Look Once), etc., and may be a neural network (NN) for detecting objects, semantic segmentation, or instance segmentation. In this case, the learning model 913 (stripping state determination model) demonstrates the object extraction function of the image of the workpiece 8, and can identify the regions of the stripping layer 82 that are in a stripped state, an unstripped state, or a soot-generating state in the entire workpiece 8.

[0042] Although it is assumed in this embodiment that the learning model 913 (stripped state determination model) is CNN, it is not limited to this. It can also be a neural network other than CNN, a transformer, a bidirectional encoder representation from transformer (BERT), a generative pre-trained transformer (GPT), a recurrent neural network (RNN), a long short-term memory (LSTM), a CNN, a support vector machine (SVM), a Bayesian network, linear regression, regression tree, multiple linear regression, random forest, ensemble learning, etc., which are learning models 913 (stripped state determination models) constructed by other learning algorithms.

[0043] Figure 4 is a functional block diagram showing the functional units included in the processing unit 91 (control unit) of the information processing device 9. The processing unit 91 of the information processing device 9 operates as the acquisition unit 911, image segmentation unit 912, learning model 913, score calculation unit 914, heat map generation unit 915, judgment unit 916, and output unit 917 by executing the program P stored in the memory unit 92.

[0044] The acquisition unit 911 acquires an image of the workpiece 8 output from the imaging device 64. The image of the workpiece 8 includes a release layer 82 formed on the substrate 81 of the workpiece 8. The acquisition unit 911 outputs the acquired image to the image segmentation unit 912. For example, when the imaging device 64 sequentially outputs images of the workpiece 8 at a specific period, the acquisition unit 911 can continue to process the images captured at the specific period by sequentially outputting the sequentially acquired images to other functional units such as the image segmentation unit 912.

[0045] The image segmentation unit 912 generates block images by segmenting the image of the workpiece 8 from the acquisition unit 911 into blocks of a specific size. Therefore, multiple block images are generated from one image of the workpiece 8. The image segmentation unit 912 can also perform compression processing on the generated block images.

[0046] The image segmentation unit 912 can also assign a number (image number) to the position of each generated (segmented) block image. The image number is set, for example, by a combination of the vertical and horizontal addresses of the block images in the image of workpiece 8. By storing these addresses in array variables, each block image is assigned (associated). By using the image number, the segmented block images can be collected and the image of workpiece 8 can be reconstructed. The image segmentation unit 912 inputs the generated block images into the learning model 913 (stripping state determination model). When the processing unit 91 of the information processing device 9 includes a GPU with parallel processing capabilities, the image segmentation unit 912 can also perform input of the block images of the learning model 913 and inference of the learning model 913 in parallel processing.

[0047] The learning model 913 (peeling state determination model) outputs information about the state of the peeling layer 82 contained in the input block image based on the input block image. As described above, the learning model 913 outputs a score indicating the probability (score) that the state of the peeling layer 82 contained in the input block image is peeled (normal peeling) and a score indicating the probability (score) that it is not peeled (poor peeling). Alternatively, the learning model 913 may only output a score indicating the probability (score) that the state of the peeling layer 82 contained in the input block image is peeled (normal peeling). Alternatively, the learning model 913 may output binary information indicating either the state of the peeling layer 82 contained in the input block image is peeled or not peeled. The learning model 913 outputs the inferred information about the state of the peeling layer 82 to the score calculation unit 914.

[0048] The score calculation unit 914 operates as a peeling rate calculation unit, calculating the peeling rate of the workpiece 8 based on information about the state of the peeling layer 82 in each block image output from the learning model 913. For example, when 100 block images are generated from an image of the workpiece 8, these block images are each assigned (associated) with a probability (score) representing the peeling state (normal peeling). For example, when this probability is represented by a value from 0 to 1 consisting of a specific number of significant digits, the probability of the peeling state (normal peeling) of the entire workpiece 8 (the entire peeling layer 82) can be calculated by collecting (adding) the probabilities of all these block images. The learning model 913 infers that when the peeling layer 82 contained in the block image is completely peeled, the probability (score) is 1, and as the degree of peeling decreases, the probability (score) becomes closer to 0. The score calculation unit 914 calculates the peeling rate of the workpiece 8 by dividing the sum of the probabilities (probabilities representing the peeling state (normal peeling)) of all block images by the number of block images. For example, when the probability of all block images is 1, the peeling rate of workpiece 8 is 1 (100%). The score calculation unit 914 outputs the calculated peeling rate to the determination unit 916.

[0049] The fraction calculation unit 914 calculates distribution data representing the distribution of the probability of the peeling state (normal peeling) of the entire workpiece 8 by mapping the positions of each block image together. The fraction calculation unit 914 outputs the calculated distribution data to the heat map generation unit 915. By dividing the image of the workpiece 8 into multiple block images, and deriving the probability of the peeling state (normal peeling) in each block image, areas of non-peeling state (poor peeling) that partially occur in the entire workpiece 8 can be efficiently identified.

[0050] The heat map generation unit 915 generates a heat map that visualizes the probability distribution of peeling states (normal peeling) throughout the entire workpiece 8 based on the distribution data output from the fraction calculation unit 914. In this heat map, the red chroma can be increased when the probability of a peeling state (normal peeling) is higher, and the blue chroma can be increased when the probability of a peeling state (normal peeling) is lower (indicating a higher probability of an unpeeled state (poor peeling)). The heat map generation unit 915 can also mark objects such as frames representing areas of unpeeled states to overlap with the heat map.

[0051] The determination unit 916 determines whether the peeling of the peeling layer 82 in the workpiece 8 is proceeding normally based on the peeling rate output from the fraction calculation unit 914. The memory unit 92 of the information processing device 9 stores in advance a reference value for determining whether the peeling of the peeling layer 82 in the workpiece 8 is proceeding normally. The determination unit 916 determines whether the peeling of the peeling layer 82 in the workpiece 8 is proceeding normally by comparing or contrasting such a specific reference value with the peeling rate in the workpiece 8. For example, the specific reference value is set to 0.99 (99%). When the peeling rate in the workpiece 8 is above the reference value, the determination unit 916 determines that the peeling of the peeling layer 82 in the LLO process is proceeding normally. When the peeling rate in the workpiece 8 does not reach the reference value, it determines that the peeling of the peeling layer 82 is not proceeding normally (poor peeling has occurred).

[0052] The output unit 917 uses the heat map from the heat map generation unit 915, the judgment result from the judgment unit 916, and the image of the workpiece 8 from the acquisition unit 911 to generate the management screen (screen data) described later. For example, the generated management screen (screen data) is output to the display device 941.

[0053] Figure 5 is a flowchart showing an example of the processing steps of the processing unit 91. For example, the processing unit 91 of the information processing device 9 receives the operator's operation via a keyboard or the like connected to input / output, and performs the following processing based on the received operation.

[0054] The processing unit 91 of the information processing apparatus 9 acquires an image of the workpiece 8 captured by the photographic device 64 (S101). The processing unit 91 of the information processing apparatus 9 divides the acquired image into a plurality of block images of a specific size (S102). The processing unit 91 of the information processing apparatus 9 generates a plurality of block images by dividing the acquired image of the workpiece 8 into block images of a specific size. The processing unit 91 of the information processing apparatus 9 may also assign image numbers to the divided block images.

[0055] The processing unit 91 of the information processing apparatus 9 inputs the segmented block images into the learning model 913 (peeling state determination model) (S103). The processing unit 91 of the information processing apparatus 9 obtains information about the state of the peeling layer 82 output by the learning model 913 (peeling state determination model) (S104). The processing unit 91 of the information processing apparatus 9 inputs each segmented block image sequentially or in parallel into the learning model 913 (peeling state determination model) and obtains information about the state of the peeling layer 82 output by the learning model 913. This information about the state of the peeling layer 82 includes a probability (score) indicating that the state of the peeling layer 82 contained in the block image is peeled (peeling is normal). It is inferred that when the peeling layer 82 contained in the block image is completely peeled, the probability (score) can be 1, and as the degree of peeling decreases, the probability (score) becomes closer to 0.

[0056] The processing unit 91 of the information processing apparatus 9 calculates the peeling rate (S105) based on the acquired information regarding the state of the peeling layer 82. The processing unit 91 of the information processing apparatus 9 calculates the peeling rate of the workpiece 8 by dividing the sum of the probabilities of all block images (representing the probability of peeling state (normal peeling)) by the number of block images. For example, when the probability of all block images is 1, the peeling rate of the workpiece 8 is 1 (100%). As the peeling rate decreases (approaches 0), it indicates that the area where peeling failure occurs increases.

[0057] The processing unit 91 of the information processing device 9 determines whether the calculated peeling rate is above a specific reference value (S106). The processing unit 91 of the information processing device 9 determines whether the peeling of the peeling layer 82 in the workpiece 8 is proceeding normally based on a specific reference value (e.g., 0.99) that is stored in the memory unit 92 in advance.

[0058] When the peeling rate is above a specific reference value (S106: Yes), the processing unit 91 of the information processing device 9 determines that the peeling of the peeling layer 82 in the workpiece 8 is proceeding normally (S107). When the peeling rate is not above the specific reference value (S106: No), that is, when the peeling rate does not reach the specific reference value, the processing unit 91 of the information processing device 9 determines that the peeling of the peeling layer 82 in the workpiece 8 is not proceeding normally, and peeling failure occurs (S1061).

[0059] The processing unit 91 of the information processing apparatus 9 generates a heat map based on the acquired information about the state of the peeling layer 82 (S108). The processing unit 91 of the information processing apparatus 9 calculates distribution data representing the distribution of the probability of the peeling state (normal peeling) of the entire workpiece 8 by mapping the positions of the respective blocks of the image together, and generates a heat map using the distribution data.

[0060] The processing unit 91 of the information processing device 9 outputs the determination result and the heat map (S109). The processing unit 91 of the information processing device 9 generates a management screen (management data) for displaying the heat map and the determination result, and outputs the generated management screen (management data) to a portable terminal device of the administrator, such as the display device 941 or the laser stripping device 1, and displays the management screen on the display device 941 or the portable terminal device.

[0061] Figure 6 is a diagram illustrating one example of the management screen of the laser stripping device 1. The management screen of the laser stripping device 1 includes, for example, a list display area that displays the stripping determination results in a list, a workpiece 8 image display area that displays an image of the workpiece 8, a thermal image display area that displays a thermal image, and a suggestion display area that displays suggestion information according to the stripping determination results.

[0062] In the list display area, a display bar is configured to display the worksheet number, stripping judgment result, stripping rate, block image size (block image size), and number of block images (block image number) of the LLO processing matters recorded when using the laser stripping device 1.

[0063] By arranging the image display area of ​​workpiece 8 and the thermal image display area side by side, the degree of peeling in each area of ​​the workpiece 8 captured by the camera device 64 can be compared with that of the workpiece 8, thereby improving the visibility for the manager of the laser peeling device 1. Areas with low peeling degree, that is, areas determined to be in an unpeeled state, can be displayed by overlaying objects such as red frames to attract the manager's attention.

[0064] The suggestion display area displays suggestions (opinions) derived based on the stripping determination results. The memory unit 92 of the information processing device 9 stores suggestion content corresponding to the stripping determination results in the form of a form, for example. The processing unit 91 of the information processing device 9 can derive suggestions based on the stripping determination results by referring to the form.

[0065] According to this embodiment, the information processing device 9 (processing unit 91) inputs the image of the workpiece 8 captured by the imaging device 64 into the learning model 913 (peeling state determination model) to derive information about the state of the peeling layer 82 of the workpiece 8. By using the learning model 913 (peeling state determination model) in this way, information about the state of the peeling layer 82 can be efficiently derived during the laser lift-off (LLO) process of irradiating the workpiece 8 with laser light and peeling off the peeling layer 82. The state of the peeling layer 82 can include the peeling state of the peeling layer 82 being peeled off normally, the unpeeled state produced when the energy density (ED) of the laser light (line beam) is low, and the soot state (ash) produced when the energy density (ED) of the laser light (line beam) is high. Since the learning model 913 outputs (infers) that the state of the peeling layer 82 of the workpiece 8 is peeled, unpeeled, or soot produced when the image of the workpiece 8 is input, more detailed information about the state of the peeling layer 82 can be efficiently derived. The information processing device 9 (processing unit 91) associates the exported information about the status of the stripping layer 82 with the acquired image (the image of the workpiece 8 captured by the photographic device 64), and outputs it, for example, to the display device 941 connected to the information processing device 9, so as to efficiently notify the operator of the laser stripping device 1 about the status of the stripping layer 82.

[0066] According to this embodiment, since the information processing device 9 (processing unit 91) determines whether peeling in the peeling layer 82 is proceeding normally based on the information about the state of the peeling layer 82 output by the learning model 913, the determination of peeling or non-peeling in the LLO process can be performed efficiently. The information processing device 9 (processing unit 91) identifies and displays information about the area (peeling region) of the peeled portion of the peeling layer 82 that is being peeled normally, based on the information about the state of the peeling layer 82. The information processing device 9 (processing unit 91) calculates the area ratio of the peeled portion relative to the area of ​​the entire substrate 81 (the entire peeling layer 82 formed on the substrate 81) as the peeling rate. If the peeling rate (area ratio) is, for example, 99% (0.99) or higher, the peeling state is determined to be normal; if it is less than 99%, the peeling state is determined to be abnormal (poor peeling). By using the peeling rate calculated based on the information about the state of the peeling layer 82 in this way, it is possible to efficiently determine whether peeling in the peeling layer 82 is proceeding normally.

[0067] According to this embodiment, the information processing device 9 (processing unit 91) segments the image acquired from the imaging device 64 into multiple block images of a specific size and inputs them into the learning model 913, deriving information about the state of the peeling layer 82 in each block image. Since the learning model 913 outputs information about the state of the peeling layer 82 contained in each input multiple block image, compared to the case where the entire image acquired from the imaging device 64 is input into the learning model 913, the accuracy of the output information can be improved. Furthermore, by deriving the possibility of representing the peeling state (normal peeling) in each segmented block image, areas of non-peeling state (poor peeling) that partially occur in the entire workpiece 8 (the entire peeling layer 82) can be efficiently identified.

[0068] According to this embodiment, since the information processing device 9 (processing unit 91) generates and outputs a heat map representing the state distribution of the peeling layer 82 in the workpiece 8 based on information about the state of the peeling layer 82 in each block image, the visibility of the peeling state of the entire peeling layer 82 in the workpiece 8 can be improved when the manager of the laser peeling device 1 grasps it. Furthermore, since objects such as frames of the unpeeled state are superimposed on the heat map, the operator of the laser peeling device 1 can be effectively drawn to the unpeeled state.

[0069] (Example 2) Figure 7 is a flowchart showing one example of the processing steps of the processing unit 91 in Example 2 (control for poor peeling). For example, the processing unit 91 of the information processing device 9 receives operations from the operator through a keyboard or the like connected to input / output, and performs the following processing based on the received operations.

[0070] The processing unit 91 of the information processing device 9 acquires an image of the workpiece 8 captured by the photographic device 64 (S201). The processing unit 91 of the information processing device 9 divides the acquired image into multiple block images of a specific size (S202). The processing unit 91 of the information processing device 9 inputs the divided block images into the learning model 913 (peeling state determination model) (S203). The processing unit 91 of the information processing device 9 acquires information about the state of the peeling layer 82 output by the learning model 913 (peeling state determination model) (S204). The processing unit 91 of the information processing device 9 calculates the peeling rate based on the acquired information about the state of the peeling layer 82 (S205). The processing unit 91 of the information processing device 9 determines whether the calculated peeling rate is above a specific reference value (S206). When the peeling rate is above the specific reference value (S206: Yes), the processing unit 91 of the information processing device 9 determines that the peeling of the peeling layer 82 in the workpiece 8 is proceeding normally (S207). The processing unit 91 of the information processing device 9 performs the processing of S201 to S207 in the same way as the processing of S101 to S107 in Embodiment 1.

[0071] When the peeling rate is not above a specific reference value (S206: No), that is, when the peeling rate does not reach the specific reference value, the processing unit 91 of the information processing device 9 determines that the peeling of the peeling layer 82 in the workpiece 8 has not proceeded normally, and peeling failure has occurred (S2061). The processing unit 91 of the information processing device 9 performs the processing of S2061 in the same way as the processing of S1061 in Embodiment 1.

[0072] The processing unit 91 of the information processing apparatus 9 performs a process to stop irradiating the workpiece 8 with laser light (S2062). When the peeling rate does not reach a specific reference value and a peeling defect is determined to have occurred, the processing unit 91 of the information processing apparatus 9 stops irradiating the workpiece 8 with laser light and stops production, for example, by outputting a processing stop signal to the laser peeling apparatus 1 or the programmable logic controller (PLC) that controls the laser peeling apparatus 1. Therefore, it is possible to prevent the substrate 81 with peeling defects from being manufactured in subsequent production.

[0073] The corresponding measures for poor peeling are not limited to stopping the laser light irradiation of the workpiece 8, but may also change the energy density of the laser light irradiating the workpiece 8. When the peeling rate does not reach a specific benchmark value and the processing unit 91 of the information processing device 9 determines that a poor peeling has occurred, the type of the poor peeling, that is, based on whether the poor peeling is due to no peeling (unpeeled state) or the generation of soot (ash) (soot generation state), changes the energy density of the laser light irradiating the workpiece 8.

[0074] When the learning model 913 (peeling state determination model) outputs (infers) the probability (score) of each of the following states—peeling state, non-peeling state, and soot-generating state—based on the input image of the workpiece 8 (segmented block image), the processing unit 91 of the information processing device 9 identifies the peeling layer 82 as being in any state based on the output of the learning model 913. That is, when the peeling rate does not reach a specific benchmark value and a peeling defect is determined to have occurred, the peeling defect is caused by either the non-peeling state or the soot-generating state. When a peeling defect is determined to have occurred, the processing unit 91 of the information processing device 9 identifies the state of the peeling layer 82 as either the non-peeling state or the soot-generating state based on the output of the learning model 913.

[0075] When the state of the release layer 82 is identified as unreleased, the processing unit 91 of the information processing apparatus 9 increases the energy density of the laser light irradiating the workpiece 8. When the state of the release layer 82 is identified as producing soot, the processing unit 91 of the information processing apparatus 9 decreases the energy density of the laser light irradiating the workpiece 8. The processing unit 91 of the information processing apparatus 9 can increase or decrease the energy density by, for example, increasing or decreasing the transmittance of the attenuator 3 or increasing or decreasing the output of the laser light source 2. By increasing or decreasing the energy density of the laser light according to the state of the release layer 82, the yield of the substrate can be improved.

[0076] The processing unit 91 of the information processing apparatus 9 generates a heat map based on the acquired information regarding the state of the peeling layer 82 (S208). The processing unit 91 of the information processing apparatus 9 outputs a determination result and a heat map (S209). When outputting the determination result and heat map, if a peeling defect is determined to have occurred, the processing unit 91 of the information processing apparatus 9 can output these determination results and heat map along with information indicating that laser irradiation has been stopped. The processing unit 91 of the information processing apparatus 9 performs the processing of S208 to S209 in the same manner as processing S108 to S109 in Embodiment 1.

[0077] According to this embodiment, when the information processing device 9 (processing unit 91) determines that a peeling defect has occurred because the peeling state is abnormal, i.e., the peeling in the peeling layer 82 is not proceeding normally, it outputs a processing stop signal to the laser peeling device 1 or the programmable logic controller (PLC) controlling the laser peeling device 1 to stop the laser light irradiation on the workpiece 8. Therefore, the manufacture of products with peeling defects can be prevented.

[0078] All features of the embodiments disclosed herein are illustrative and should not be considered as limitations. The technical features described in the various embodiments can be combined with each other, and the scope of this invention includes all modifications within the scope of the claims and scopes equivalent to the claims. [Simplified Explanation of the Diagram]

[0011] [Figure 1] is a diagram showing an example of the configuration of the laser stripping apparatus of Embodiment 1. [Figure 2] is a diagram showing an example of the configuration of the information processing apparatus included in the laser stripping apparatus. [Figure 3] is an explanatory diagram showing an example of a learning model (stripping state determination model). [Figure 4] is a functional block diagram showing the functional units included in the processing unit (control unit) of the information processing apparatus. [Figure 5] is a flowchart showing an example of the processing steps through the processing unit. [Figure 6] is a diagram explaining an example of the management screen of the laser stripping apparatus. [Figure 7] is a flowchart showing an example of the processing steps through the processing unit of Embodiment 2 (corresponding control for poor stripping).

Claims

1. A laser stripping apparatus for irradiating a workpiece comprising a substrate and a release layer formed on the substrate with laser light, and stripping the release layer from the substrate, comprising: A photographic device for photographing a workpiece irradiated by the aforementioned laser light; The processing unit processes the images captured by the aforementioned photographic device. Specifically, the processing unit acquires an image of the workpiece captured by the aforementioned photographic device; segments the image acquired by the aforementioned photographic device into multiple block images of a specific size; inputs the segmented block images into a learning model, and derives information about the state of the aforementioned peeling layer from each block image. The learning model outputs information about the state of the peeling layer contained in the block image when the aforementioned block image is input; and outputs the derived information about the state of the aforementioned peeling layer.

2. The laser stripping apparatus of claim 1, wherein the state of the aforementioned stripping layer includes at least one of a stripped state, an unstripped state, and a state that generates soot.

3. The laser stripping apparatus of claim 1, wherein the aforementioned processing unit associates and outputs information about the state of the aforementioned stripped layer with an image obtained from the aforementioned photographic apparatus.

4. The laser stripping apparatus of claim 1, wherein the aforementioned processing unit determines whether the stripping in the stripping layer is proceeding normally based on the derived information about the state of the aforementioned stripping layer; and outputs the aforementioned determination result.

5. The laser stripping apparatus as described in claim 4, wherein, The aforementioned information regarding the state of the release layer includes information about the release portion region, which shows the portion of the release layer that has been peeled off; the aforementioned processing unit calculates the area ratio of the aforementioned release portion relative to the aforementioned substrate; when the calculated area ratio is above a certain ratio, the peeling state is determined to be normal; and when the calculated area ratio does not reach the certain ratio, the peeling state is determined to be abnormal.

6. The laser stripping apparatus of claim 5, wherein when the aforementioned processing unit determines that the stripping state is abnormal, it performs a process to stop irradiating the aforementioned workpiece with laser light.

7. The laser stripping apparatus of claim 1, wherein the aforementioned processing unit generates a heat map based on information about the state of the stripping layer in each of the exported images of the aforementioned blocks, the heat map showing the state distribution of the stripping layer in the aforementioned workpiece; and outputs the generated heat map.

8. The laser stripping apparatus of claim 7, wherein the aforementioned processing unit, in the state distribution of the stripped layers displayed in the aforementioned heat map, displays objects in the unstripped state overlapping with the aforementioned heat map and outputs them.

9. The laser stripping apparatus of claim 1, wherein the aforementioned photographic apparatus includes a coaxial incident linear camera.

10. An information processing method, wherein the following processing is performed in a computer: acquiring an image of a workpiece on which a release layer is formed on a substrate is irradiated with laser light; dividing the acquired image into a plurality of block images of a specific size; by inputting the divided block images into a learning model, deriving information about the state of the release layer from each block image, wherein the learning model outputs information about the state of the release layer contained in the block image when the block image is input; and outputting the derived information about the state of the release layer.

11. A readable recording medium for recording a program, the program being used to perform the following processing in a computer: acquiring an image of a workpiece on which a release layer is formed on a substrate is irradiated with laser light; dividing the acquired image into a plurality of block images of a specific size; by inputting the divided block images into a learning model, deriving information about the state of the release layer from each of the block images, wherein the learning model outputs information about the state of the release layer contained in the block image when the block image is input; and outputting the derived information about the state of the release layer.

Images