Transfer method, device and equipment of light emitting diode chip, medium and product

Abstract

The invention relates to the technical field of light-emitting diode chip transfer, and discloses a light-emitting diode chip transfer method, device, equipment, medium and product, the light-emitting diode chip transfer method comprises the following steps: obtaining a plurality of block data in a target group, the block data comprising parameter information of a plurality of light-emitting diode chips on corresponding blocks, the parameter information comprises wavelength and / or brightness; the multiple pieces of block data are converted into multiple pieces of first image information, and the multiple pieces of first image information and the multiple pieces of block data are in one-to-one correspondence; determining the type of block data corresponding to the plurality of pieces of first image information according to a first learning model; the light emitting diode chips on the blocks corresponding to the multiple pieces of target block data are transferred to a target area of the target driving substrate, and the multiple pieces of target block data are block data of complementary types in the multiple pieces of block data. According to the invention, the chromaticity difference of the display screen is reduced while the transfer of the light-emitting diode chip is completed in a short time.

Description

Technical Field

[0001] The present invention relates to the technical field of light emitting diode chip transfer, and in particular to a light emitting diode chip transfer method, device, equipment, medium and product. Background Art

[0002] In the manufacturing process of Micro Light-Emitting Diode (Micro LED) display screens, a key step is to accurately position a large number (e.g., millions or even hundreds of millions) of micron-sized Micro LED chips from the growth substrate (epitaxial wafer) on the driver substrate. However, due to the production process, the wavelength distribution of the Micro LED wafer is not uniform. If the Micro LED chips are transferred to the driver substrate in the order in which they are arranged on the epitaxial wafer, the display screen may have obvious color differences, making it difficult to meet the requirements of high-quality images and color accuracy.

[0003] In the manufacturing process of larger-sized light-emitting diode (LED) display screens, a binning method can be used to reduce color differences caused by extension. Specifically, before picking up the LED chips, multiple LED chips are classified based on the parameters (wavelength) of the LED chips. When picking up and placing the LED chips one by one, LED chips of the same type (similar wavelengths) are placed on the same driver substrate to reduce color differences. In the process of Micro-LED mass transfer, if the above binning method is used to reduce color differences, a lot of time costs will be consumed. Summary of the invention

[0004] In view of this, the present invention provides a method, device, equipment, medium and product for transferring a light emitting diode chip, so as to complete the transfer of the light emitting diode chip in a shorter time and reduce the chromaticity difference of the display screen.

[0005] In a first aspect, the present invention provides a method for transferring a light-emitting diode chip, the method comprising: acquiring multiple block data in a target group, wherein the block data includes parameter information of multiple light-emitting diode chips on corresponding blocks, the parameter information including wavelength and / or brightness; converting the multiple block data into multiple first image information, wherein the multiple first image information and the multiple block data correspond one-to-one; determining the type of block data corresponding to the multiple first image information according to a first learning model; transferring the light-emitting diode chips on the blocks corresponding to the multiple target block data to a target area of ​​a target driving substrate, wherein the multiple target block data are block data of complementary types among the multiple block data.

[0006] The method for transferring the light-emitting diode chip provided in this embodiment converts the multiple block data into multiple first image information after acquiring the multiple block data of the target group, and then determines the type of the block data corresponding to the multiple first image information according to the first learning model, and transfers the light-emitting diode chips on the multiple blocks of complementary types to the target area of ​​the target driving substrate. The present invention uses machine learning image recognition technology to determine the type of the block to be transferred, and transfers several types of complementary colors to the same area of ​​the same target substrate, thereby achieving scattered uniform distribution and reducing the chromaticity difference of the display screen.

[0007] In an optional implementation, converting the plurality of block data into a plurality of first image information includes: converting the plurality of block data into a plurality of first image information based on a correspondence between parameter information and color.

[0008] In an optional embodiment, the first learning model is a pre-trained convolutional neural network model, and the types of block data corresponding to multiple first image information are determined based on the first learning model, including: inputting the first image information into the first learning model, and determining the type of block data corresponding to the first image information based on the output of the first learning model.

[0009] In an optional embodiment, before determining the type of block data corresponding to multiple first image information according to the first learning model, the method also includes: determining a sample database, wherein the sample database includes multiple second image information and the types of block data corresponding one by one to the multiple second image information; based on the sample database, training an initial convolutional neural network model to determine the first learning model.

[0010] In an optional embodiment, determining a sample database includes: determining multiple feature vectors corresponding one-to-one to multiple second image information based on an autoencoder model; determining types of block data corresponding one-to-one to multiple second image information based on multiple feature vectors and a clustering algorithm; correcting types of block data corresponding one-to-one to multiple second image information in response to user operations; and determining the types of block data corresponding one-to-one to multiple second image information and the corrected multiple second image information as a sample database.

[0011] In this embodiment, a sample database is determined by an autoencoder model and a clustering algorithm, and then an initial convolutional neural network model is trained based on the sample database to obtain a first learning model, which can improve the accuracy of the determined first learning model.

[0012] In an optional embodiment, before transferring the LED chips on the blocks corresponding to the multiple target block data to the target area of ​​the target driving substrate, the method also includes: determining the multiple target block data based on the types corresponding to the multiple block data and the distances between the blocks corresponding to the multiple block data.

[0013] In this embodiment, multiple target block data are determined based on the types corresponding to the multiple block data and the distances between the blocks corresponding to the multiple block data, which can reduce the time of the transfer process and improve the transfer efficiency.

[0014] In an optional embodiment, the method also includes: obtaining parameter information of light-emitting diode chips on multiple epitaxial wafers; storing the parameter information in multiple blocks divided by the multiple epitaxial wafers to obtain multiple block data; and dividing the multiple block data into multiple groups based on the parameter average value and preset parameter range of the block data, wherein each group includes multiple block data, and the target group is a group in the multiple groups corresponding to the target driving substrate.

[0015] In this embodiment, based on the parameter average value of the block data and the preset parameter range, a plurality of block data are divided into a plurality of groups, and a plurality of different LED chips are mixed within the preset parameter range, which can further improve the color mixing effect.

[0016] In an optional embodiment, before storing parameter information in multiple blocks divided by multiple epitaxial wafers to obtain multiple block data, the method also includes: dividing multiple epitaxial wafers into blocks according to the maximum number of light-emitting diode chips transferred by the transfer device each time to obtain multiple blocks.

[0017] In this embodiment, the maximum number of light-emitting diode chips that can be transferred by the transfer device each time is divided into multiple blocks of epitaxial wafers, which can reduce the number of transfers and thus improve the transfer efficiency.

[0018] In a second aspect, the present invention provides a device for transferring a light-emitting diode chip, the device comprising: an acquisition module for acquiring multiple block data in a target group, wherein the block data includes parameter information of multiple light-emitting diode chips on a corresponding block, and the parameter information includes wavelength and / or brightness; a conversion module for converting the multiple block data into multiple first image information, wherein the multiple first image information and the multiple block data correspond one-to-one; a determination module for determining the type of block data corresponding to the multiple first image information according to a first learning model; and a transfer module for transferring the light-emitting diode chips on the blocks corresponding to the multiple target block data to a target area of ​​a target driving substrate, wherein the multiple target block data are block data of complementary types among the multiple block data.

[0019] In an optional implementation, the conversion module includes: a first conversion unit, configured to convert a plurality of block data into a plurality of first image information based on a correspondence between parameter information and color.

[0020] In an optional embodiment, the first learning model is a pre-trained convolutional neural network model, and the determination module includes: a first determination unit, used to input the first image information into the first learning model, and determine the type of block data corresponding to the first image information according to the output of the first learning model.

[0021] In an optional embodiment, the determination module also includes: a second determination unit, used to determine a sample database, wherein the sample database includes multiple second image information and types of block data corresponding one to the multiple second image information; a training unit, used to train an initial convolutional neural network model based on the sample database to determine the first learning model.

[0022] In an optional embodiment, the second determination unit includes: a first determination subunit, used to determine multiple feature vectors corresponding one-to-one to multiple second image information based on an autoencoder model; a second determination subunit, used to determine the type of block data corresponding one-to-one to multiple second image information based on multiple feature vectors and a clustering algorithm; a correction unit, used to correct the type of block data corresponding one-to-one to multiple second image information in response to user operations; and a third determination subunit, used to determine the type of block data corresponding one-to-one to multiple second image information and the corrected multiple second image information as a sample database.

[0023] In an optional implementation, the determination module further includes: a fourth determination subunit, configured to determine a plurality of target block data based on types corresponding to the plurality of block data and distances between blocks corresponding to the plurality of block data.

[0024] In an optional embodiment, the device also includes: an acquisition unit, used to acquire parameter information of light-emitting diode chips on multiple epitaxial wafers; a first division module, used to store parameter information according to multiple blocks divided by multiple epitaxial wafers to obtain multiple block data; a second division module, used to divide the multiple block data into multiple groups based on the parameter average value and preset parameter range of the block data, wherein each group includes multiple block data, and the target group is a group in the multiple groups corresponding to the target driving substrate.

[0025] In an optional embodiment, the device further includes: a third division module, which is used to divide the blocks of the plurality of epitaxial wafers into a plurality of blocks according to the maximum number of light-emitting diode chips transferred by the transfer device each time.

[0026] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0027] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to cause a computer to execute the method of the first aspect or any corresponding embodiment thereof.

[0028] In a fifth aspect, the present invention provides a computer program product, comprising computer instructions for causing a computer to execute the method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the related technologies, the drawings required for use in the specific embodiments or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 is a schematic structural diagram of an epitaxial wafer according to an embodiment of the present invention;

[0031] Figure 2 is a schematic diagram of a light emitting diode chip on a driving substrate according to an embodiment of the present invention;

[0032] Figure 3 is a schematic flow chart of a method for transferring a light emitting diode chip according to an embodiment of the present invention;

[0033] Figure 4 is a schematic diagram of block division of an epitaxial wafer according to an embodiment of the present invention;

[0034] Figure 5 is a schematic diagram of a Micro LED chip on a driving substrate according to an embodiment of the present invention;

[0035] Figure 6 is a schematic diagram of first image information according to an embodiment of the present invention;

[0036] Figure 7 is a schematic diagram of a type of block data according to an embodiment of the present invention;

[0037] Figure 8 is a schematic diagram of another type of block data according to an embodiment of the present invention;

[0038] Fig. 9 is a schematic diagram of another block division of an epitaxial wafer according to an embodiment of the present invention;

[0039] Fig.10 is a schematic diagram of block division of another epitaxial wafer according to an embodiment of the present invention;

[0040] Fig.11 is a schematic diagram of a transfer result according to an embodiment of the present invention;

[0041] Fig.12 is a schematic flow chart of another method for transferring a light emitting diode chip according to an embodiment of the present invention;

[0042] Fig.13 is a schematic diagram of the structure of an autoencoder model according to an embodiment of the present invention;

[0043] Fig.14 is a schematic diagram of second image information extraction according to an embodiment of the present invention;

[0044] Fig.15 is a schematic diagram of the structure of an initial convolutional neural network model according to an embodiment of the present invention;

[0045] Fig.16 is a schematic diagram of converting second image information into types according to an embodiment of the present invention;

[0046] Fig.17 is a schematic flow chart of another method for transferring a light emitting diode chip according to an embodiment of the present invention;

[0047] Fig.18 is a schematic diagram of block division of another epitaxial wafer according to an embodiment of the present invention;

[0048] Fig.19 is a schematic diagram of a specific process of a light emitting diode chip transfer method according to an embodiment of the present invention;

[0049] Fig. 20 is a structural block diagram of a light emitting diode chip transfer device according to an embodiment of the present invention;

[0050] Fig.21 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention.

[0051] Figure numerals: 110, LED chip; 210, first driving substrate; 220, second driving substrate; 230, third driving substrate; 510, Micro LED chip; 520, fourth driving substrate; 1110, target driving substrate; 1310, encoding module; 1320, decoding module; 2001, acquisition module; 2002, conversion module; 2003, determination module; 2004, transfer module; 2110, processor; 2120, memory; 2130, input device; 2140, output device. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0053] Micro LED display refers to a new display technology that miniaturizes and matrices the LED structure design. It is composed of a micron-level LED chip array. For example, the size of the LED chip is between 5 microns and 10 microns. Micro LED display has the advantages of high resolution, low power consumption and high response speed. It can be applied to scenarios such as smartphones, tablets, smart watches, car displays and digital billboards.

[0054] Accurately transferring a large number of Micro LED chips on the epitaxial wafer to the driving substrate is one of the key technologies in the manufacturing process of Micro LED display screens. Due to the production process, the wavelength distribution of the Micro LED wafer is not uniform. If the Micro LED chips are transferred to the driving substrate in sequence according to the arrangement order of the Micro LED chips on the epitaxial wafer, it may cause obvious color difference in the display screen, which is difficult to meet the requirements of high-quality images and color accuracy.

[0055] Among them, color difference manifests as some pixels being brighter or darker than others, or it may manifest as color shift. The driver substrate is the final carrier platform of the Micro-LED display. It not only has good physical stability, but also integrates drive circuits and signal lines to control the light-emitting state of each Micro-LED chip. The driver substrate can be rigid (such as glass, silicon or metal, etc.) or flexible (such as plastic, etc.).

[0056] In the manufacturing process of larger-sized LED displays, binning is used to reduce color differences caused by extension. Figure 1 As shown, the LED chips 110 on the epitaxial wafer are tested to determine the process parameters (e.g., wavelength (WLD) and light output power (LOP) etc.) of each LED chip 110, and the LED chips 110 within a specified range are classified into one category. Figure 1 As shown, the LED chips 110 are divided into three categories, and the three categories are identified by numbers 1, 2 and 3. Then, when picking up and placing the LED chips 110 one by one, the LED chips 110 of the same category are placed on the same driving substrate to reduce the color difference of the display screen. Figure 2 As shown, the LED chips 110 of category 1 are placed on the first driving substrate 210 , the LED chips of category 2 are placed on the second driving substrate 220 , and the LED chips of category 3 are placed on the third driving substrate 230 .

[0057] However, Micro LED chips are relatively small, usually only tens to hundreds of microns in size, and the transfer may involve millions or even hundreds of millions of Micro LED chips. If the Micro LED chips are transferred according to the above binning method, it will consume a lot of time.

[0058] In view of this, the present invention provides a method for transferring light-emitting diode chips, which uses machine learning image recognition technology to determine the type of blocks to be massively transferred, and mixes and transfers the Micro-LED chips on several types of complementary blocks to the target area of ​​the same target driving substrate, thereby achieving dispersed and uniform distribution, completing the transfer in a shorter time while reducing the chromaticity difference of the display screen.

[0059] According to an embodiment of the present invention, an embodiment of a method for transferring a light-emitting diode chip is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0060] In this embodiment, a method for transferring a light emitting diode chip is provided, which can be used in devices such as notebook computers, tablet computers or computers. Figure 3 FIG. 1 is a flow chart of a method for transferring a light emitting diode chip according to an embodiment of the present invention. Figure 3 As shown, the method comprises the following steps:

[0061] Step S301, obtaining a plurality of block data in a target group.

[0062] The block data includes parameter information of multiple light-emitting diode chips on the corresponding block, and the parameter information includes wavelength and / or brightness. The light-emitting diode chip may be a Micro LED chip.

[0063] Specifically, in order to improve the transfer efficiency, the epitaxial wafer is divided into multiple blocks, and a transfer machine can transfer multiple Micro LED chips on a block to the driving substrate at the same time. Figure 4 As shown, the epitaxial wafer can be divided into 9 blocks, such as Figure 5 As shown, a transfer machine can simultaneously transfer multiple Micro LED chips 510 on the A1 block to the A1 region of the fourth driving substrate 520 , and simultaneously transfer multiple Micro LED chips 510 on the A2 block to the A2 region of the fourth driving substrate 520 .

[0064] Among them, the size of the block can be set by a technician, and the size of the block can also be determined according to the maximum number of single transfers by the transfer machine. The transfer machine can be a micro-nano stamp device or other Micro LED chip transfer equipment commonly used in the field, which is not specifically described in the present invention.

[0065] For example, the wavelength and / or brightness of multiple Micro LED chips 510 on the epitaxial wafer can be detected by a spot measuring machine, and the parameter information (i.e., wavelength and / or brightness) of the multiple Micro LED chips 510 corresponding to the block can be stored in blocks to form block data.

[0066] Specifically, the Micro LED display includes multiple driving substrates, each driving substrate is placed on a Micro LED chip 510 with a preset wavelength range (and / or a preset brightness range), and after the block data is determined, the multiple block data are divided into multiple groups based on parameter information, each group includes multiple block data, and the multiple groups correspond to the driving substrates one by one. The target group is a group in the multiple groups corresponding to the target driving substrate, and the multiple driving substrates can be set as the target driving substrate in sequence.

[0067] Exemplarily, the plurality of block data included in each of the plurality of groups may be stored in advance in a memory of the computer device, and after the target driving substrate is determined, the plurality of block data in the target group may be acquired from the memory.

[0068] Exemplarily, the wavelength and / or brightness of multiple Micro LED chips 510 on the epitaxial layer can be obtained from a spot measuring machine to form block data, and the block data can be grouped based on parameter information, and then multiple block data of the target group can be determined based on the target driving substrate.

[0069] Step S302, converting the plurality of block data into a plurality of first image information.

[0070] The plurality of first image information and the plurality of block data correspond one to one.

[0071] Specifically, after acquiring a plurality of block data of the target group, each block data may be converted into corresponding first image information based on the correspondence between the parameter information and the color.

[0072] For example, taking the parameter information as wavelength, the first image information can be as follows: Figure 6 As shown, the fixed wavelength color coordinates represent the corresponding relationship between parameter information and color. Figure 6 The vertical bar on the right may represent a fixed wavelength color coordinate, 460 nm represents the minimum wavelength in the multiple block data of the target group, and 466 nm represents the maximum wavelength in the block data of the target group.

[0073] Exemplarily, the fixed wavelength color coordinates may be configured in advance in a computer device by a designer, or may be automatically generated by the computer device.

[0074] Exemplarily, the first image information may be 128×128×3 image information, where 128×128×3 indicates that both the width and height of the first image information are 128 pixels, that is, the first image information is composed of 128 rows and 128 columns of pixels, 3 indicates the number of color channels, and each pixel contains three color components of red (Red), green (Green), and blue (Blue), and the first image information is an RGB image.

[0075] Step S303: determining the type of block data corresponding to the plurality of first image information according to the first learning model.

[0076] Exemplarily, the first learning model may be a classification model in a machine learning model.

[0077] Specifically, after determining the first image information, the first image information is input into the first learning model, and the type of block data corresponding to the first image information can be determined based on the output of the first learning model. For example, the upper left dark color is one type, and the upper left light color is another type.

[0078] Exemplarily, the type of the first image information includes 18 categories, and the 18 categories can be as follows: Figure 7 and Figure 8 shown.

[0079] Specifically, machine learning is a multidisciplinary discipline that covers probability theory, statistics, approximate theory and complex algorithm knowledge. It uses computers as tools and is committed to simulating human learning methods in real time, and divides existing content into knowledge structures to effectively improve learning efficiency. It has excellent performance in image recognition processing. Applying machine learning algorithms to Micro LED chip transfer can greatly improve the efficiency of transfer.

[0080] Step S304 , transferring the LED chips on the blocks corresponding to the target block data to the target area of ​​the target driving substrate.

[0081] The multiple target block data are block data of complementary types among the multiple block data.

[0082] Specifically, if the types corresponding to two block data have the same or similar colors in relative positions, the types of the two block data are complementary, and the relative positions may refer to left and right, top and bottom, upper left and lower right, and lower left and upper right, etc.

[0083] For example, Figure 7 Category 1 in the upper right corner and category 3 in the lower left corner are complementary. Figure 7 Category 2 in the upper left and orange in the lower right are complementary. Figure 8 Category 14 in the lower right center and Figure 8 Category 18 in blue in the upper left center is complementary. Figure 8 The blue category 15 and Figure 8 Category 17 in blue on the center left is complementary.

[0084] Specifically, based on the law of color mixing, designers have found through research that it is not necessarily monochromatic light that causes color stimulation, but often composite light, that is, the sum of radiation of various wavelengths. Moreover, the human visual organ has comprehensive performance, that is, it has the ability to combine light of different wavelengths emitted from objects of a certain color into one color. Therefore, adjusting the arrangement of Micro LEDs at a microscopic scale and mixing the Micro LED chips in blocks of complementary types can make the overall color change more evenly distributed, thereby reducing the visibility of chromaticity differences.

[0085] For example, after determining the multiple target block data, the computer device can control the transfer machine to sequentially transfer the LED chips on the blocks corresponding to the multiple target block data to the corresponding areas of the target drive substrate. Fig. 9 Block A and Fig.10 The types of block B on the epitaxial wafer shown are complementary, and the Micro LED chips on block A and block B can be mixed by a transfer machine and transferred to Fig.11 The corresponding regions (region A and region B) of the target driving substrate 1110 are shown.

[0086] The method for transferring the light-emitting diode chip provided in this embodiment converts the multiple block data into multiple first image information after acquiring the multiple block data of the target group, and then determines the type of the block data corresponding to the multiple first image information according to the first learning model, and transfers the light-emitting diode chips on the multiple blocks of complementary types to the target area of ​​the target driving substrate. The present invention uses machine learning image recognition technology to determine the type of the block to be transferred, and transfers several types of complementary colors to the same area of ​​the same target substrate, thereby achieving scattered uniform distribution and reducing the chromaticity difference of the display screen.

[0087] In this embodiment, a method for transferring a light emitting diode chip is provided, which can be used in devices such as notebook computers, tablet computers or computers. Fig.12 FIG. 1 is a flow chart of a method for transferring a light emitting diode chip according to an embodiment of the present invention. Fig.12 As shown, the method comprises the following steps:

[0088] Step S1201, obtaining multiple block data in the target group.

[0089] For details, please see Figure 3 Step S301 of the illustrated embodiment will not be described in detail here.

[0090] Step S1202: based on the correspondence between the parameter information and the color, convert the plurality of block data into a plurality of first image information.

[0091] For example, the first image information may be Figure 6 The corresponding relationship between parameter information and color can be pre-configured in the computer device by the designer.

[0092] Step S1203: determine the sample database.

[0093] The sample database includes a plurality of second image information and a plurality of types of block data corresponding to the second image information.

[0094] Specifically, the second image information is image information predetermined for training the initial convolutional neural network model. The second image information is converted based on block data corresponding to the pre-selected epitaxial wafer, and the first image information is converted based on block data corresponding to the epitaxial wafer to be transferred.

[0095] Exemplarily, the above step S1203 includes:

[0096] Step S12031, based on the autoencoder model, determine multiple feature vectors corresponding to multiple pieces of second image information.

[0097] Specifically, the autoencoder model is an unsupervised learning neural network model. The autoencoder model captures the intrinsic structure of the data by learning the compressed representation (encoding) and reconstruction (decoding) of the data, and then determines the feature vector corresponding to the data.

[0098] For example, Fig.13 As shown, the Autoencoder model includes an encoder module 1310 and a decoder module 1320. The encoder module 1310 includes three convolutional layers and three pooling layers (e.g., Fig.13 The decoding module 1320 includes 3 deconvolution layers and 3 de-pooling layers (for example, Fig.13 The first deconvolution layer, the first anti-pooling layer, the second deconvolution layer, the second anti-pooling layer, the third deconvolution layer and the third anti-pooling layer in ). Fig.14 As shown, the encoding module 1310 is used to convert the second image information of 128×128×3 into a feature vector of 16×16×8 through three convolution and three pooling operations, and the decoding module 1320 is used to convert the feature vector of 16×16×8 into the third image information of 128×128×3 through three deconvolution and three depooling operations.

[0099] The Autoencoder model may be a trained initial Autoencoder model. Specifically, after obtaining a plurality of second image information, the second image information may be input into the initial Autoencoder model, and the difference between the second image information and the third image information output by the initial Autoencoder model may be determined by a loss function, and then the parameters in the initial Autoencoder model may be adjusted based on the difference to obtain an adjusted initial Autoencoder model, and then the second image information may be input into the adjusted initial Autoencoder model, and the above process may be repeated until the difference is less than a preset difference. The adjusted initial Autoencoder model corresponding to a difference less than the preset difference is recorded as the trained initial Autoencoder model.

[0100] Exemplarily, the loss function may be a binary cross entropy (BCELoss) function.

[0101] Specifically, after obtaining the second image information, the second image information is input into the autoencoder model, and the feature vector corresponding to the second image information is determined based on the output result of the encoding module. For example, the second image information of 128×128×3 is input into the encoding module 1310, and a feature vector of 16×16×8 is obtained based on the output of the encoding module 1310 (i.e., the second image information is extracted by the encoding module 1310 to obtain 16×16×8 three-dimensional feature information), and then the 16×16×8 feature vector is fully connected to a one-dimensional feature vector of 1×2048, and the one-dimensional feature vector of 1×2048 is determined as the feature vector corresponding to the second image information.

[0102] Exemplarily, the block data of a plurality of different wafers to be epitaxially grown may be converted into the second image information through the above-mentioned steps S1201 and S1202, thereby obtaining the second image information.

[0103] Step S12032: determining the types of block data corresponding to the plurality of second image information based on the plurality of feature vectors and the clustering algorithm.

[0104] Exemplarily, the clustering algorithm may be a K-means clustering algorithm. Specifically, after determining the plurality of feature vectors, the plurality of feature vectors are classified by the K-means clustering algorithm to further determine the type of the second image information corresponding to the plurality of feature vectors.

[0105] For example, at least two feature vectors with close distances among multiple feature vectors can be classified into one category, where the distance refers to the Euclidean distance between the two feature vectors. If the distance from the first feature vector to the preset feature vector is less than the preset distance, and the distance from the second feature vector to the preset feature vector is less than the preset distance, then the first feature vector and the second feature vector are feature vectors with close distances. The first feature vector and the second feature vector are any two of the multiple feature vectors, and the preset feature vector is a preset cluster center.

[0106] Step S12033, in response to the user operation, modify the type of the block data corresponding to the plurality of second image information.

[0107] Specifically, after determining the types of block data corresponding to multiple second image information one by one, the designer can manually correct the types of block data corresponding to multiple second image information one by one based on the human-computer interaction device (mouse, keyboard or touch screen, etc.) to ensure the accuracy of the determined block data types.

[0108] Step S12034: determine the type of block data that corresponds one-to-one to the plurality of second image information and the plurality of corrected second image information as a sample database.

[0109] Specifically, after determining the type of block data corresponding one-to-one to the multiple corrected second image information, the data (i.e., the multiple second image information) and the classification results (i.e., the type of block data corresponding one-to-one to the multiple corrected second image information) can be stored together to form a sample database.

[0110] In some other optional implementations, the sample database may also be determined by manual annotation.

[0111] Step S1204: Based on the sample database, train an initial convolutional neural network model to determine a first learning model.

[0112] For example, Fig.15 and Fig.16 As shown in Figure 2, the initial convolutional neural network model includes 3 convolutional layers, 3 pooling layers, and 2 fully connected layers (such as Fig.15 The fourth convolutional layer, the fourth pooling layer, the fifth convolutional layer, the fifth pooling layer, the sixth convolutional layer, the sixth pooling layer, the first fully connected layer, and the second fully connected layer in the initial convolutional neural network). The input of the initial convolutional neural network is the second image information, and the output of the initial convolutional neural network is the type corresponding to the second image information.

[0113] Specifically, the second image information in the sample data set can be input into the initial convolutional neural network model, and then the difference between the type of the output of the initial convolutional neural network model and the actual classification label of the second image information can be determined through the loss function, and the parameters in the initial convolutional neural network model can be adjusted based on the difference to obtain a modified initial convolutional neural network model, and then the above process is repeated until the difference is less than the preset difference. The modified initial convolutional neural network model corresponding to the difference less than the preset difference is the first learning model. Among them, the loss function can be a binary cross entropy (BCELoss) function.

[0114] Step S1205 , inputting the first image information into the first learning model, and determining the type of block data corresponding to the first image information according to the output of the first learning model.

[0115] Among them, the first learning model is a pre-trained convolutional neural network model.

[0116] Step S1206 , transferring the LED chips on the blocks corresponding to the target block data to the target area of ​​the target driving substrate.

[0117] For details, please see Figure 3 Step S304 of the illustrated embodiment will not be described in detail here.

[0118] In this embodiment, a sample database is determined by an autoencoder model and a clustering algorithm, and then an initial convolutional neural network model is trained based on the sample database to obtain a first learning model, which can improve the accuracy of the determined first learning model.

[0119] In this embodiment, a method for transferring a light emitting diode chip is provided, which can be used in devices such as notebook computers, tablet computers or computers. Fig.17 FIG. 1 is a flow chart of a method for transferring a light emitting diode chip according to an embodiment of the present invention. Fig.17 As shown, the method comprises the following steps:

[0120] Step S1701, obtaining parameter information of light emitting diode chips on multiple epitaxial wafers.

[0121] The parameter information includes wavelength and / or brightness. The wavelength and / or brightness of the LED chip are detected by a spot tester, and the computer device can obtain the parameter information of the LED chips on multiple epitaxial wafers from the spot tester.

[0122] Step S1702 , dividing a plurality of epitaxial wafers into blocks according to the maximum number of light-emitting diode chips that the transfer device transfers each time, to obtain a plurality of blocks.

[0123] Exemplarily, the size of the block is determined based on the maximum number of LED chips transferred by the transfer device each time, that is, the number of LED chips contained in the block is less than or equal to the maximum number of LED chips transferred by the transfer device each time, for example Fig.18 As shown, the epitaxial wafer can be divided into 9 blocks, namely block 1 to block 9.

[0124] Step S1703, storing parameter information according to multiple blocks divided by multiple epitaxial wafers to obtain multiple block data.

[0125] Specifically, the block data also includes location information of the corresponding block. The location information may be the name of the epitaxial wafer where the block is located plus the identifier of the block. The name and identifier may be numbers or letters, etc.

[0126] Step S1704: dividing the plurality of block data into a plurality of groups based on the parameter average value of the block data and the preset parameter range.

[0127] Each group includes a plurality of block data, and the target group is a group corresponding to the target driving substrate among the plurality of groups.

[0128] Exemplarily, taking the parameter information as wavelength as an example, the parameter average value is the wavelength average value, and the preset parameter range is the preset wavelength range. After acquiring multiple block data, the wavelength average value of each block data is determined, and the multiple block data are grouped based on the wavelength average value and the preset wavelength range. The preset parameter range can be determined by the designer according to the wavelength / or brightness range of the light-emitting diode chip on the epitaxial wafer. For example, the wavelength / or brightness range is divided into units of 2nm to obtain the preset parameter range.

[0129] For example, the preset wavelength ranges are [460nm, 462nm], (462nm, 464nm], and (464nm, 466nm], respectively. At this time, the block data with the wavelength average value in [460nm, 462nm] among the multiple block data are divided into the first group, the block data with the wavelength average value in (462nm, 464nm] among the multiple block data are divided into the second group, and the block data with the wavelength average value in (464nm, 466nm] among the multiple block data are divided into the third group. Among them, the interval [460nm, 462nm] includes the endpoint 460nm and the endpoint 462nm, the interval (462nm, 464nm] does not include the endpoint 462nm and includes the endpoint 464nm, and the interval (464nm, 466nm] does not include the endpoint 464nm and includes the endpoint 466nm.

[0130] Step S1705, obtaining multiple block data in the target group.

[0131] For details, please see Figure 3 Step S301 of the illustrated embodiment will not be described in detail here.

[0132] Step S1706, converting the plurality of block data into a plurality of first image information.

[0133] For details, please see Figure 3 Step S302 of the illustrated embodiment will not be described in detail here.

[0134] Step S1707: determining the type of block data corresponding to the plurality of first image information according to the first learning model.

[0135] For details, please see Figure 3 Step S303 of the illustrated embodiment will not be described in detail here.

[0136] Step S1708 , determining a plurality of target block data based on the types corresponding to the plurality of block data and the distances between the blocks corresponding to the plurality of block data.

[0137] Specifically, when the number of block data of complementary types corresponding to multiple block data is greater than the number required by the target area, one of the epitaxial wafers is used as a benchmark (the benchmark epitaxial wafer), and a preset number of block data of complementary types with the shortest distance to the benchmark epitaxial wafer are determined as target block data.

[0138] For example, if the number of blocks required in the target area is 2, two block data of complementary types on the same epitaxial wafer can be determined as target block data. If there are no two complementary block data on the same epitaxial wafer, the two complementary block data with the shortest physical distance can be determined as target block data.

[0139] Step S1709 , transferring the LED chips on the blocks corresponding to the target block data to the target area of ​​the target driving substrate.

[0140] For details, please see Figure 3 Step S304 of the illustrated embodiment will not be described in detail here.

[0141] In this embodiment, according to the maximum number of light-emitting diode chips transferred by the transfer device each time, the multiple epitaxial wafers are divided into blocks, which can reduce the number of transfers and thus improve the transfer efficiency. Based on the parameter average value and preset parameter range of the block data, the multiple block data are divided into multiple groups, and multiple different light-emitting diode chips are mixed within the preset parameter range, which can further improve the color mixing effect. Based on the types corresponding to the multiple block data and the distances between the blocks corresponding to the multiple block data, multiple target block data are determined, which can reduce the time of the transfer process and improve the transfer efficiency.

[0142] Combine the following Fig.19 The specific process of the light emitting diode chip transfer method provided by the present invention is described in detail.

[0143] like Fig.19 As shown, the first step is to process data. Specifically, after obtaining multiple block data, the multiple block data are grouped based on the parameter information in the block data, and the block data is converted into the first image information. The second step is to train the model. Specifically, the second image information is roughly classified using the Autoencoder model and the K-Means clustering algorithm, and then the type of block data corresponding to the second image information is manually corrected to obtain a sample database. After that, the initial convolutional neural network is trained based on the sample database to obtain a first learning model (i.e., the trained initial convolutional neural network model). The third step is to classify. Specifically, the block data in each group is classified using the first learning model. The fourth step is to perform block matching transfer. Specifically, blocks of complementary types are transferred to the same drive substrate to achieve a color mixing effect, thereby reducing chromaticity differences.

[0144] In this embodiment, a device for transferring a light emitting diode chip is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made are not repeated here. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0145] This embodiment provides a device for transferring a light emitting diode chip, such as Fig. 20 As shown, including:

[0146] An acquisition module 2001 is used to acquire a plurality of block data in a target group, wherein the block data includes parameter information of a plurality of light-emitting diode chips on a corresponding block, and the parameter information includes wavelength and / or brightness;

[0147] A conversion module 2002, configured to convert the plurality of block data into a plurality of first image information, wherein the plurality of first image information and the plurality of block data correspond one to one;

[0148] A determination module 2003, configured to determine the type of block data corresponding to the plurality of first image information according to the first learning model;

[0149] The transfer module 2004 is used to transfer the LED chips on the blocks corresponding to the multiple target block data to the target area of ​​the target driving substrate, wherein the multiple target block data are block data of complementary types among the multiple block data.

[0150] In some optional implementations, the conversion module 2002 includes:

[0151] The first conversion unit is used to convert the plurality of block data into a plurality of first image information based on the correspondence between the parameter information and the color.

[0152] In some optional implementations, the first learning model is a pre-trained convolutional neural network model, and the determination module 2003 includes:

[0153] The first determination unit is used to input the first image information into the first learning model, and determine the type of block data corresponding to the first image information according to the output of the first learning model.

[0154] In some optional implementations, the determining module 2003 further includes:

[0155] A second determining unit is used to determine a sample database, wherein the sample database includes a plurality of second image information and a type of block data corresponding to the plurality of second image information;

[0156] The training unit is used to train an initial convolutional neural network model based on a sample database to determine a first learning model.

[0157] In some optional implementations, the second determining unit includes:

[0158] A first determining subunit is used to determine a plurality of feature vectors corresponding one to one to the plurality of second image information based on an autoencoder model;

[0159] A second determination subunit, configured to determine the type of block data corresponding to the plurality of second image information one by one based on the plurality of feature vectors and a clustering algorithm;

[0160] A correction unit, configured to correct the type of block data corresponding to the plurality of second image information in response to a user operation;

[0161] The third determining subunit is used to determine the types of block data that correspond one-to-one to the plurality of second image information and the corrected plurality of second image information as a sample database.

[0162] In some optional implementations, the determining module 2003 further includes:

[0163] The fourth determining subunit is used to determine a plurality of target block data based on the types corresponding to the plurality of block data and the distances between the blocks corresponding to the plurality of block data.

[0164] In some optional embodiments, the device further comprises:

[0165] An acquisition unit, used for acquiring parameter information of light-emitting diode chips on a plurality of epitaxial wafers;

[0166] A first division module is used to store parameter information according to a plurality of blocks divided by a plurality of epitaxial wafers to obtain a plurality of block data;

[0167] The second division module is used to divide the plurality of block data into a plurality of groups based on the parameter average value and the preset parameter range of the block data, wherein each group includes a plurality of block data, and the target group is a group corresponding to the target drive substrate among the plurality of groups.

[0168] In some optional embodiments, the device further comprises:

[0169] The third division module is used to divide the blocks of the multiple epitaxial wafers into multiple blocks according to the maximum number of light-emitting diode chips transferred by the transfer device each time.

[0170] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0171] The transfer device of the light-emitting diode chip in this embodiment is presented in the form of a functional unit, where the unit refers to an application specific integrated circuit (ASIC) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0172] The embodiment of the present invention also provides a computer device, see Fig.21 , Fig.21 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Fig.21 As shown, the computer device includes: one or more processors 2110, memory 2120, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig.21 A processor 2110 is taken as an example.

[0173] The processor 2110 may be a central processing unit, a network processor or a combination thereof. The processor 2110 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0174] The memory 2120 stores instructions executable by at least one processor 2110 so as to enable at least one processor 2110 to implement the method shown in the above embodiment.

[0175] The memory 2120 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 2120 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 2120 may optionally include a memory remotely arranged relative to the processor 2110, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0176] The memory 2120 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 2120 may also include a combination of the above types of memory.

[0177] The computer device also includes an input device 2130 and an output device 2140. The processor 2110, the memory 2120, the input device 2130 and the output device 2140 may be connected via a bus or other means. Fig.21 The example of connecting through bus is taken in the following.

[0178] The input device 2130 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 2140 may include a display device, an auxiliary lighting device (e.g., an LED), and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display, and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0179] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0180] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0181] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the present invention.

Claims

1. A method for transferring a light emitting diode chip, characterized in that: The method comprises: Acquire multiple block data in the target group, wherein the block data includes parameter information of multiple light-emitting diode chips on the corresponding block, and the parameter information includes wavelength and / or brightness; Converting the plurality of block data into a plurality of first image information, wherein the plurality of first image information corresponds to the plurality of block data in one-to-one correspondence; Determining, according to the first learning model, the type of block data corresponding to the plurality of first image information; The light emitting diode chips on the blocks corresponding to the plurality of target block data are transferred to the target area of ​​the target driving substrate, wherein the plurality of target block data are block data of complementary types among the plurality of block data.

2. The method according to claim 1, characterized in that The converting the plurality of block data into a plurality of first image information comprises: Based on the correspondence between the parameter information and the color, the plurality of block data are converted into a plurality of first image information.

3. The method according to claim 1, characterized in that: The first learning model is a pre-trained convolutional neural network model, and determining the type of block data corresponding to the plurality of first image information according to the first learning model includes: The first image information is input into the first learning model, and the type of block data corresponding to the first image information is determined according to the output of the first learning model.

4. The method according to claim 3, characterized in that Before determining the types of block data corresponding to the plurality of first image information according to the first learning model, the method further includes: Determine a sample database, wherein the sample database includes a plurality of second image information and types of block data corresponding to the plurality of second image information; Based on the sample database, an initial convolutional neural network model is trained to determine the first learning model.

5. The method according to claim 4, characterized in that The determining of the sample database comprises: Based on the autoencoder model, determining a plurality of feature vectors corresponding one to one to the plurality of second image information; Based on the multiple feature vectors and the clustering algorithm, determine the type of block data corresponding to the multiple pieces of second image information one by one; In response to a user operation, modifying the type of the block data corresponding to the plurality of second image information one by one; The types of block data corresponding one-to-one between the plurality of second image information and the plurality of corrected second image information are determined as the sample database.

6. The method according to any one of claims 1 to 5, characterized in that Before transferring the light emitting diode chips on the blocks corresponding to the plurality of target block data to the target area of ​​the target driving substrate, the method further comprises: The plurality of target block data are determined based on the types corresponding to the plurality of block data and the distances between the blocks corresponding to the plurality of block data.

7. The method according to any one of claims 1 to 5, characterized in that The method further comprises: Obtaining parameter information of light-emitting diode chips on a plurality of epitaxial wafers; Storing parameter information according to the multiple blocks divided by the multiple epitaxial wafers to obtain multiple block data; Based on the parameter average value and the preset parameter range of the block data, the plurality of block data are divided into a plurality of groups, wherein each group includes a plurality of block data, and the target group is a group corresponding to the target drive substrate among the plurality of groups.

8. The method according to claim 7, characterized in that Before storing parameter information in the multiple blocks divided by the multiple epitaxial wafers to obtain multiple block data, the method further includes: According to the maximum number of light-emitting diode chips transferred by the transfer device each time, the multiple epitaxial wafers are divided into blocks to obtain the multiple blocks.

9. A device for transferring a light emitting diode chip, characterized in that: The device comprises: An acquisition module, used for acquiring a plurality of block data in a target group, wherein the block data includes parameter information of a plurality of light-emitting diode chips on a corresponding block, and the parameter information includes wavelength and / or brightness; A conversion module, used for converting the plurality of block data into a plurality of first image information, wherein the plurality of first image information and the plurality of block data correspond one to one; A determination module, configured to determine the type of block data corresponding to the plurality of first image information according to a first learning model; The transfer module is used to transfer the light emitting diode chips on the blocks corresponding to the multiple target block data to the target area of ​​the target driving substrate, wherein the multiple target block data are block data of complementary types among the multiple block data.

10. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method according to any one of claims 1 to 7 by executing the computer instructions.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 7.

12. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the method according to any one of claims 1 to 7.

Images