Machine vision positioning method, system and storage medium in COB die bonding process

Abstract

The present invention provides a machine vision positioning method, system and storage medium in the COB die bonding process, the method comprising acquiring an image to be detected and a process data packet, analyzing the image to be detected and a preset image processing model to obtain the fullness of metal paste dots, determining the pressure control parameters in the COB die bonding process according to the process data packet and the fullness of the metal paste dots, performing detection according to the image to be detected and a preset edge detection algorithm to obtain the outer contour data of the image to be detected, judging the position deviation of the bare crystal and the area deviation of the metal paste dots according to the outer contour data to obtain the deviation result, and generating the movement compensation data of the control motor based on the deviation result, and generating the control instruction to control the operation of the motor in the COB die bonding process according to the pressure control parameter and the movement compensation data. The method improves the target positioning accuracy in the COB die bonding process and improves the overall production quality.

Description

Technical Field

[0001] The present invention relates to the technical field of COB die bonding, and in particular to a machine vision positioning method, system and storage medium in a COB die bonding process. Background Art

[0002] In the integrated circuit processing technology, die bonding is a key step to accurately adhere the bare die film or die to the electrically insulating carrier to achieve mechanical and electrical connections between circuit units. The die bonder is the core equipment in this process, performing precise die bonding operations through complex mechanical and visual systems.

[0003] At present, in one existing technology, image color features are used to analyze the actual target and background features as well as the appearance and color diversity of the target, so as to determine the characteristics of the entire target image. These technologies have made certain progress in image recognition, but they still face the problem of insufficient precision in the die bonding process. This lack of precision is mainly reflected in the inaccuracy of target recognition and positioning, especially when dealing with complex backgrounds, similar target colors and background colors, which can easily lead to errors and instability in the die bonding process, thereby affecting the overall production quality.

[0004] In summary, the existing technology has insufficient recognition accuracy in the die bonding process, which easily leads to errors and instability in the die bonding process, thereby affecting the overall production quality. Summary of the invention

[0005] The present invention provides a machine vision positioning method, system and storage medium in a COB die bonding process to solve the technical problem that the prior art has insufficient recognition accuracy in the die bonding process, which easily leads to errors and instability in the die bonding process, thereby affecting the overall production quality.

[0006] In the first aspect, in order to solve the above technical problems, the present invention provides a machine vision positioning method in a COB die bonding process, comprising:

[0007] Obtain the image to be inspected and the process data packet;

[0008] Analyze the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots;

[0009] Determine the pressure control parameters in the COB die bonding process according to the process data package and the fullness of the metal paste dots;

[0010] Perform detection based on the image to be detected and a preset edge detection algorithm to obtain outer contour data of the image to be detected;

[0011] Determine the position deviation of the bare die and the area deviation of the metal paste dots according to the outer contour data, obtain a deviation result, and generate movement compensation data for controlling the motor based on the deviation result;

[0012] According to the pressure control parameter and the movement compensation data, a control instruction is generated to control the operation of the motor in the COB die bonding process.

[0013] In an optional implementation, the analyzing the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots includes:

[0014] Identify the position data and size data of each metal paste dot according to the image to be detected;

[0015] Based on the image processing model, the actual fullness of all the cream dots is evaluated;

[0016] The actual fullness is compared with the standard fullness for similarity to obtain the fullness of the metal paste dots.

[0017] In an optional implementation manner, determining the position deviation of the bare die and the area deviation of the metal paste dots according to the outer contour data to obtain the deviation result includes:

[0018] Determine the position data of the bare die and the area data of the metal paste dots according to the outer contour data;

[0019] Subtracting the position data from the standard position parameters in the process data packet to obtain a position deviation result;

[0020] Subtracting the metal paste dot area data from the standard metal paste dot area in the process data packet to obtain an area deviation result;

[0021] The deviation results include position deviation results and area deviation results.

[0022] In an optional implementation, determining the position data of the bare die and the area data of the metal paste dots according to the outer contour data includes:

[0023] Performing a grayscale extraction operation on the image to be detected to obtain grayscale information of the image to be detected;

[0024] Comparing the grayscale information of the image to be detected with the preset grayscale information, correcting the data at the grayscale difference to the standard value, and determining the accurate grayscale data of the image to be detected;

[0025] Performing binarization processing according to the accurate grayscale data to obtain binary image data of the image to be detected;

[0026] Obtaining metal paste dot data according to the binary image data;

[0027] Taking the metal paste dot data as a base point, generating a template image matching the image to be detected, and performing template image matching to identify the center point position of the paste dot of the metal paste dot data;

[0028] According to the center point position of the paste point and the actual bare die diameter contained in the process data packet, the area data of the metal paste point is obtained;

[0029] Acquire actual position data of the bare die, and determine the position of the bare die center point corresponding to the bare die from the actual position data;

[0030] Taking the center point of the bare crystal as the positioning point, the deviation between the center point of the bare crystal and the center point of the paste point is identified to obtain the relative position of the bare crystal to the metal paste point and determine the position data of the metal paste point.

[0031] In an optional implementation manner, obtaining actual position data of the bare die includes:

[0032] The image position data of the bare die is obtained by detecting the edge detection algorithm and the metal paste dot area data, wherein, i From 1 to n, it represents the detected n image position data;

[0033] According to the image position data and the actual die diameter contained in the process data packet, the image position data is converted by the following formulas (1) and (2) to obtain the actual die position data:

[0034] (1)

[0035] (2)

[0036] The actual position data includes the actual width distance and the actual height distance. is the actual width distance, is the actual height distance, is the actual die diameter width, is the actual bare die diameter height, is the pixel width, is the pixel height.

[0037] In an optional implementation, the subtraction of the metal paste dot area data from the standard metal paste dot area in the process data packet to obtain the area deviation result includes:

[0038] The area deviation result is calculated by the following formula:

[0039]

[0040] Among them, Diviation is the metal paste dot area deviation, Actualarea is the actual metal paste dot area data, and area is the metal paste dot area data.

[0041] In an optional implementation, the acquiring the image to be detected includes:

[0042] The original image data obtained by the image acquisition device is input into the pre-trained convolution model to obtain the image to be detected;

[0043] The training process of the convolutional model includes:

[0044] A training set is constructed based on historical image data and historical images to be detected, and the training set is input into the initial convolution model for iterative calculation. When the loss parameter value is less than the preset loss threshold, the training is determined to be completed, and a trained convolution model is obtained; the convolution model includes a convolution layer and a deconvolution layer.

[0045] In an optional embodiment, the method further includes:

[0046] According to the image to be detected, identifying outer contour edge points of a plurality of metal paste dots;

[0047] Determine the center point position of the metal paste dots according to the relative position relationship between the outer contour edge points of each two metal paste dots, and determine the degree of deviation of a single metal paste dot by calculating and comparing the center point position;

[0048] Generate overall error data based on the degree of deviation of all metal paste points;

[0049] Dividing the overall error data into units of graphic regions to generate regional error data;

[0050] The control instruction is updated according to the regional error data, so as to control the operation of the motor during the COB die bonding process according to the updated control instruction.

[0051] In a second aspect, the present invention provides a machine vision positioning system in a COB die bonding process, comprising:

[0052] A data acquisition module is used to acquire the image to be detected and the process data packet;

[0053] An analysis module, used for analyzing the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots;

[0054] A parameter determination module, used to determine the pressure control parameters in the COB die bonding process according to the process data package and the metal paste dot fullness;

[0055] A detection module, used to perform detection based on the image to be detected and a preset edge detection algorithm to obtain the outer contour data of the image to be detected;

[0056] A deviation module, used for determining the position deviation of the bare die and the area deviation of the metal paste dots according to the outer contour data, obtaining a deviation result, and generating movement compensation data for controlling the motor based on the deviation result;

[0057] The instruction module is used to generate a control instruction to control the operation of the motor during the COB die bonding process according to the pressure control parameter and the movement compensation data.

[0058] In a third aspect, the present invention further provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the machine vision positioning method in the COB die bonding process described in any one of the above is implemented.

[0059] In a fourth aspect, the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the machine vision positioning methods in the COB bonding process described above.

[0060] Compared with the prior art, the present invention has the following beneficial effects: the present invention obtains the image to be detected and the process data packet, and uses the preset image processing model to analyze the image to determine the fullness of the metal paste dots. Based on this fullness and process data packet, the pressure control parameters in the COB die bonding process are set. At the same time, the preset edge detection algorithm is applied to obtain the outer contour data of the image to be detected, and the position deviation of the bare crystal and the area deviation of the metal paste dots are judged accordingly to generate corresponding motion compensation data. Finally, the pressure control parameters and the motion compensation data are combined to generate control instructions to adjust the working state of the motor in the COB die bonding process.

[0061] Through the above technical solution, the image processing model is first used to accurately analyze the fullness of the metal paste dots, and then determine the appropriate pressure control parameters. Then, the outer contour data is obtained through the edge detection algorithm, the position deviation and area deviation are identified, and the compensation data is calculated. These compensation data are combined with the pressure control parameters to generate effective control instructions, optimize the working state of the motor, and thus achieve accurate target positioning. The present invention significantly improves the target positioning accuracy in the COB solid crystal process through precise image processing and deviation compensation mechanism, thereby reducing production defects caused by positioning errors and improving production stability and overall production quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a flow chart of a machine vision positioning method in a COB die bonding process provided by a first embodiment of the present invention;

[0063] Figure 2 It is a schematic diagram of a process for obtaining the fullness of metal paste dots provided by the present invention;

[0064] Figure 3 It is a schematic diagram of a process for obtaining a deviation result provided by the present invention;

[0065] Figure 4 It is a structural schematic diagram of a machine vision positioning system in a COB die bonding process provided by a second embodiment of the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] Reference Figure 1 The first embodiment of the present invention provides a machine vision positioning method in a COB die bonding process, comprising the following steps:

[0068] S1, obtaining the image to be inspected and the process data packet;

[0069] S2, analyzing the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots;

[0070] S3, determining the pressure control parameters in the COB die bonding process according to the process data package and the fullness of the metal paste dots;

[0071] S4, performing detection according to the image to be detected and a preset edge detection algorithm to obtain outer contour data of the image to be detected;

[0072] S5, determining the position deviation of the bare die and the area deviation of the metal paste dots according to the outer contour data, obtaining a deviation result, and generating movement compensation data for controlling the motor based on the deviation result;

[0073] S6, generating a control instruction according to the pressure control parameter and the movement compensation data to control the operation of the motor during the COB die bonding process.

[0074] In step S11, the image to be inspected and the process data packet are obtained.

[0075] It should be noted that the image to be detected is the image captured by the die bonding machine through the visual processing device during the die bonding process, and the process data packet is a data set used to guide and adjust the die bonding process during the die bonding process. The user can input or obtain historical process data in advance.

[0076] In one embodiment, a light source and a camera are provided in a visual image processing device of a die bonding machine, and the light source is usually placed below the camera. The camera captures the original image data of the die bonding object through a digital sensor, performs image enhancement processing, and stores the collected original image data in a computer system or database. The original image data obtained by the image acquisition device is input into a pre-trained convolution model to obtain an image to be detected.

[0077] Among them, the image enhancement processing includes white balance adjustment, noise reduction and sharpening; the training process of the convolution model includes: constructing a training set based on historical image data and historical images to be detected, inputting the training set into the initial convolution model for iterative calculation, and when the loss parameter value is less than the preset loss threshold, the training is judged to be completed, and the trained convolution model is obtained; the convolution model includes a convolution layer and a deconvolution layer.

[0078] In the present invention, the convolution algorithm is a typical image processing algorithm, which is a method of extracting image features using digital image processing technology; the present invention applies the algorithm to visual processing and obtains the optimal image through convolution calculation.

[0079] In step S2, analysis is performed based on the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots.

[0080] In the specific implementation process, the metal paste dot fullness is an indicator used to measure the actual filling degree of each metal paste dot in the image to be detected.

[0081] like Figure 2 As shown, in one embodiment, the method for obtaining the fullness of the metal paste dots in the above step S2 specifically includes steps S21 to S23, as follows:

[0082] S21: Identify the position data and size data of each metal paste dot according to the image to be detected.

[0083] S22: Evaluate and obtain the actual fullness of all the cream dots based on the image processing model.

[0084] S23: Compare the actual fullness with the standard fullness for similarity to obtain the fullness of the metal paste dots.

[0085] Among them, image processing models include edge detection, template matching or machine learning models, etc.; similarity comparison methods include Euclidean distance, cosine similarity, etc.

[0086] In one embodiment, to obtain the fullness of metal paste dots, firstly, the position and size data of each metal paste dot in the image to be detected are identified by image processing technology, and the center coordinates and size of each paste dot are recorded. Then, the actual fullness of each metal paste dot is evaluated based on the image processing model, and the filling condition of the paste dot is calculated by the image processing model and compared with the standard process requirements. Finally, the calculated actual fullness is compared with the preset standard fullness for similarity, and through this comparison, it is determined whether each paste dot meets the standard requirements, and a report is generated to indicate the paste dots that meet and do not meet the standards, and optimization suggestions are provided.

[0087] For example, during the production process, the bare die is coated with a layer of metal paste, and the metal paste dots on each bare die need to be inspected to ensure that their quality meets the specifications. A high-precision camera is used to obtain an image of the bare die to be inspected, and the image clearly shows the position and size of each metal paste dot. Through the image recognition algorithm, the system can accurately locate each metal paste dot and measure its diameter and shape. The system records detailed data for each paste dot, including its exact position on the bare die and the size information of the metal paste dot.

[0088] Next, the image processing model is used to evaluate the actual fullness of the identified metal paste dots. The image processing model analyzes the actual shape and size of each paste dot, compares it with the ideal state, and calculates the actual fullness of each paste dot. These data reflect the filling quality of each paste dot, such as whether there are voids or uneven coatings. Finally, the actual fullness data is compared with the preset standard fullness. The standard fullness is an ideal value set according to the manufacturing process requirements, which represents the fullness that the metal paste dots should achieve in a perfect state. By comparing the similarity between the actual fullness and the standard fullness, it is possible to evaluate whether the quality of each paste dot meets the requirements. If the actual fullness deviates from the standard value, the system will issue an alarm and provide further processing suggestions, such as adjusting the coating amount of the paste dot or recalibrating the production equipment.

[0089] In step S3, the pressure control parameters in the COB die bonding process are determined according to the process data package and the metal paste dot fullness.

[0090] It should be noted that the pressure control parameter is a numerical value used to adjust the pressure during the crystal bonding process. Combined with the metal paste point fullness data and the preset standard metal paste point fullness, determine whether the current metal paste point meets the requirements of the crystal bonding process. Among them, the process data package includes the standard metal paste point fullness required by the user. If the metal paste point fullness meets the requirements of the standard metal paste point fullness, the normal pressure control parameters are set; if the metal paste point fullness does not meet the requirements of the standard metal paste point fullness, the system will automatically adjust the pressure parameters to ensure that the fullness of the metal paste points meets the requirements. Exemplarily, when the metal paste point fullness is less than the standard metal paste point fullness, the system increases the pressure parameters; when the metal paste point fullness is greater than the standard metal paste point fullness, the system reduces the pressure parameters.

[0091] In step S4, detection is performed based on the image to be detected and a preset edge detection algorithm to obtain outer contour data of the image to be detected.

[0092] The outer contour data refers to the edge contour information of the metal paste dots in the image to be detected, and the contour includes the shape and boundary of each paste dot. The edge detection algorithm is a technology used to identify areas with significant brightness changes in an image and to extract edge information of the image.

[0093] In one embodiment, a suitable edge detection algorithm is selected to identify the edges in the image. The algorithm processes the grayscale changes of the image and identifies the areas in the image where the brightness changes significantly, thereby extracting edge information. The preset edge detection algorithm is applied to the image to be detected to generate an edge image, which contains the outer contour data of the paste point. The extracted outer contour data is used for further analysis, such as calculating the area, position, shape, etc. of each paste point.

[0094] In step S5, the position deviation of the bare die and the area deviation of the metal paste dots are determined according to the outer contour data to obtain a deviation result, and movement compensation data for controlling the motor is generated based on the deviation result.

[0095] It should be noted that the position deviation of the bare die is the degree of deviation of the bare die from its ideal position, and the area deviation of the metal paste point is the difference between the actual area of ​​the metal paste point and the expected area. The deviation result includes the position deviation result and the area deviation result.

[0096] like Figure 3 As shown, in one embodiment, the method for obtaining the deviation result in the above S5 specifically includes steps S31 to S33, as follows:

[0097] S31: Determine the position data of the bare crystal and the area data of the metal paste dots according to the outer contour data.

[0098] In this embodiment, a grayscale extraction operation is performed on the image to be detected to obtain its grayscale information; the grayscale information of the image is compared with the preset standard grayscale information, and the data at the grayscale difference is corrected to the standard value, so as to determine the accurate grayscale data of the image; binarization processing is performed based on the accurate grayscale data to generate binary image data of the image to be detected. The metal paste dot data is extracted by analyzing the binary image data. With the metal paste dot data as the base point, a template image matching the image to be detected is generated, and the template image matching is performed to identify the center point position of the paste dot. According to the center point position of the paste dot and the actual bare crystal diameter contained in the process data packet, the area data of the metal paste dot is calculated; then, the actual position data of the bare crystal is obtained, and the center point position of the bare crystal corresponding to the bare crystal is determined from the actual position data; then, the center point position of the bare crystal is used as the positioning point, and the deviation between it and the center point position of the paste dot is identified, the relative position of the metal paste dot is determined, and the final position data of the metal paste dot is obtained.

[0099] It should be noted that grayscale information refers to the brightness value of each pixel in the image. Grayscale images are the standard form of black and white images, and the value of each pixel represents the brightness level from black to white, usually in the range of 0 to 255, where 0 represents black and 255 represents white. With grayscale information, the brightness and contrast of the image can be analyzed for further processing and analysis. Binarization is an image processing technique that converts a grayscale image into an image with only two colors (usually black and white). This process generates a binary image by setting a threshold value, setting pixel values ​​in the image greater than the threshold value to white, and pixel values ​​less than the threshold value to black. Binary image data is image data obtained after binarization, in which each pixel has only two possible values ​​(such as 0 and 1), representing black and white.

[0100] Preferably, the grayscale extraction operation is a process of converting a color image into a grayscale image. In this process, the color information of each pixel (usually composed of three color channels: red, green, and blue) is converted into a single grayscale value representing brightness. The grayscale extraction operation includes:

[0101] The grayscale extraction operation integrates the color information of these three channels into a single grayscale value by some method. The following formula is the grayscale conversion formula:

[0102] gray=0.299×red+0.587×green+0.114×blue

[0103] Among them, grey is the gray value, red is the red channel value, green is the green channel value, and blue is the blue channel value. The color value of each pixel is calculated according to the above formula, and the gray value obtained is usually in the range of 0 to 255, where 0 represents black, 255 represents white, and the values ​​between the two represent different gray levels.

[0104] For example, in a specific application, after taking images of the bare die to be inspected, grayscale extraction operations are performed on these images to obtain grayscale information for each pixel. These grayscale information are then compared with preset standard grayscale information, and any grayscale differences are corrected to the standard value to ensure that the grayscale data of the image is accurate. Then, these accurate grayscale data are used for binarization processing to convert the image into binary image data. This binary image data can clearly show the position and outline of the metal paste dots.

[0105] By analyzing the binary image data, the metal paste dot data, including the position and shape of each paste dot, is extracted. Based on these paste dot data, a template image matching the image to be detected is generated, and the template image matching algorithm is used to identify the center point of the paste dot.

[0106] Once the center point position of the metal paste dot is determined, the area data of the metal paste dot is calculated based on the actual bare die diameter in the process data package. For further calibration and verification, it is also necessary to obtain the actual position data of the bare die and determine the position of the center point of the bare die. Next, the center point position of the bare die is used as a positioning point and compared with the center point position of the paste dot to identify the deviation between the two. Finally, the relative position of the metal paste dot is determined and its final position data is obtained. These data will be used to adjust the production process to ensure the precise position of each metal paste dot, thereby improving the reliability and performance of the chip.

[0107] In one embodiment, obtaining the actual position data of the die includes: obtaining the image position data of the die according to the edge detection algorithm. and the metal paste dot area data, wherein, i From 1 to n, it represents the detected n image position data;

[0108] According to the image position data and the actual die diameter contained in the process data packet, the image position data is converted by the following formulas (1) and (2) to obtain the actual die position data:

[0109] (1)

[0110] (2)

[0111] The actual position data includes the actual width distance and the actual height distance. is the actual width distance, is the actual height distance, is the actual die diameter width, is the actual bare die diameter height, is the pixel width, is the pixel height.

[0112] S32: Subtract the position data from the standard position parameters in the process data packet to obtain a position deviation result.

[0113] It should be noted that the position deviation result is obtained by the following formula:

[0114] (1)

[0115] (2)

[0116] The position deviation result includes the width deviation result and the height deviation result, w is the width deviation result, h is the height deviation result, and the standard position parameters include the standard width distance Distance from standard height . Position data includes actual width distance Distance from actual height .

[0117] S33: Subtract the metal paste dot area data from the standard metal paste dot area in the process data packet to obtain the area deviation result.

[0118] It should be noted that the area deviation result is calculated by the following formula:

[0119]

[0120] Among them, Diviation is the metal paste dot area deviation, Actualarea is the actual metal paste dot area data, and area is the metal paste dot area data.

[0121] In step S6, a control instruction is generated according to the pressure control parameter and the movement compensation data to control the operation of the motor during the COB die bonding process.

[0122] It should be noted that the control instructions are commands used to adjust the operation of the motor during the COB die bonding process to achieve precise control and compensation. The pressure control parameters and movement compensation data are collected, input into the motor control algorithm, and the motor control algorithm is applied to convert these parameters into motor control instructions. The generated control instructions are sent to the motor system to control the operation of the motor during the COB die bonding process to ensure that the process requirements are met.

[0123] In order to facilitate the understanding of the present invention, some preferred embodiments of the present invention are further described below.

[0124] In one implementation, the method further includes:

[0125] According to the image to be detected, identifying outer contour edge points of a plurality of metal paste dots;

[0126] Determine the center point position of the metal paste dots according to the relative position relationship between the outer contour edge points of each two metal paste dots, and determine the degree of deviation of a single metal paste dot by calculating and comparing the center point position;

[0127] Generate overall error data based on the degree of deviation of all metal paste points;

[0128] Dividing the overall error data into units of graphic regions to generate regional error data;

[0129] The control instruction is updated according to the regional error data, so as to control the operation of the motor during the COB die bonding process according to the updated control instruction.

[0130] It should be noted that the overall error data is a total data set formed by summarizing the deviation degrees of all identified metal paste dots, and the regional error data is error data formed by dividing the overall error data into multiple graphic regions.

[0131] In one embodiment, to perform detection, first identify the outer contour edge points of multiple metal paste dots from the image to be detected, and then determine the center position of each metal paste dot based on the relative position relationship between these edge points. By comparing the actual center position with the expected position, the degree of deviation of each metal paste dot is calculated. Next, the deviation data of all paste points are summarized to generate overall error data. Subsequently, the overall error data is divided into several parts according to the graphic area in the image, and the error data of each area is calculated. Finally, the control instructions are updated according to these regional error data to adjust the operation of the motor during the COB die bonding process, thereby optimizing the production process.

[0132] Through this method, the error of metal paste dots can be fully and detailedly understood, so as to more effectively control and optimize production equipment and improve product quality during the COB die bonding process.

[0133] In summary, the present invention obtains the image to be detected and the process data packet, and uses the preset image processing model to analyze the image to determine the fullness of the metal paste dots. Based on this fullness and process data packet, the pressure control parameters in the COB die bonding process are set. At the same time, the preset edge detection algorithm is applied to obtain the outer contour data of the image to be detected, and the position deviation of the bare crystal and the area deviation of the metal paste dots are judged accordingly to generate corresponding motion compensation data. Finally, the pressure control parameters and the motion compensation data are combined to generate control instructions to adjust the working state of the motor in the COB die bonding process.

[0134] Through the above technical solution, the image processing model is first used to accurately analyze the fullness of the metal paste dots, and then determine the appropriate pressure control parameters. Then, the outer contour data is obtained through the edge detection algorithm, the position deviation and area deviation are identified, and the compensation data is calculated. These compensation data are combined with the pressure control parameters to generate effective control instructions, optimize the working state of the motor, and thus achieve accurate target positioning. The present invention significantly improves the target positioning accuracy in the COB solid crystal process through precise image processing and deviation compensation mechanism, thereby reducing production defects caused by positioning errors and improving production stability and overall production quality.

[0135] The following describes the working process of the present invention by taking a relatively common scenario as an example. Please also refer to the working scenario schematic diagram.

[0136] In the COB (Chip on Board) bonding process of a certain LED packaging factory, the bonding machine is equipped with an advanced visual processing system to ensure precise operation. The core of this system is a positioning method based on machine vision, and its workflow is as follows:

[0137] First, the die bonder captures images of the bare die and metal paste dots during the die bonding process using a light source and a high-resolution camera installed below the device. These image data are input into the system along with the standard process parameters required for the die bonding process. The process data package contains key parameters such as the standard fullness of the metal paste dots and the standard position of the bare die, which are essential for subsequent analysis and adjustment.

[0138] Next, the system processes the acquired image using a pre-trained convolutional neural network (CNN) model. This process accurately analyzes the actual fullness of the metal paste dots, that is, the degree to which each metal paste dot fills the image. The model evaluates the actual fullness and compares it with the standard fullness to determine whether it meets the process requirements.

[0139] According to the actual fullness of the metal paste dots, the system determines the appropriate pressure control parameters. If the fullness of the metal paste dots meets the standard, the system will set normal pressure parameters; if the fullness does not meet the requirements, the system will automatically adjust the pressure to ensure that the fullness of the metal paste dots meets the standard.

[0140] After that, the system uses the preset edge detection algorithm to extract the outer contour data of the metal paste dots from the image. These contour data include the shape and boundary information of the metal paste dots for further analysis. By analyzing these contour data, the system can determine the position deviation of the bare die and the area deviation of the metal paste dots, and generate relevant deviation results.

[0141] Based on these deviation results, the system generates movement compensation data for the control motor. This data is used to adjust the motor of the die bonder to correct the actual position of the bare die and the actual area of ​​the metal paste dot to align them with the expected standard.

[0142] Finally, the system generates control instructions based on the pressure control parameters and motion compensation data to control the motor in the die bonder in real time. These instructions ensure that the pressure and position adjustment during the die bond process meet the process requirements, thereby improving product quality and production efficiency.

[0143] Through this machine vision positioning method, every link in the die bonding process is precisely controlled, significantly improving the automation level of the production line and the consistency of the products.

[0144] Reference Figure 4 The second embodiment of the present invention provides a machine vision positioning system in a COB die bonding process, comprising:

[0145] The data acquisition module 101 is used to acquire the image to be detected and the process data packet;

[0146] An analysis module 102 is used to analyze the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots;

[0147] A parameter determination module 103, for determining a pressure control parameter in a COB die bonding process according to the process data packet and the metal paste dot fullness;

[0148] A detection module 104 is used to perform detection based on the image to be detected and a preset edge detection algorithm to obtain outer contour data of the image to be detected;

[0149] A deviation module 105 is used to determine the position deviation of the bare die and the area deviation of the metal paste dots according to the outer contour data, obtain a deviation result, and generate movement compensation data for controlling the motor based on the deviation result;

[0150] The instruction module 106 is used to generate a control instruction to control the operation of the motor during the COB die bonding process according to the pressure control parameter and the movement compensation data.

[0151] Preferably, the analysis module comprises:

[0152] A data recognition module, used to recognize the position data and size data of each metal paste dot according to the image to be detected;

[0153] A fullness evaluation module, used for evaluating the actual fullness of all cream dots based on the image processing model;

[0154] The similarity comparison module is used to perform a similarity comparison between the actual fullness and the standard fullness to obtain the fullness of the metal paste dots.

[0155] Preferably, the deviation module comprises:

[0156] A data determination module, used to determine the position data of the bare die and the area data of the metal paste dots according to the outer contour data;

[0157] A position deviation module, used for performing subtraction between the position data and the standard position parameters in the process data packet to obtain a position deviation result;

[0158] The area deviation module is used to obtain an area deviation result by subtracting the metal paste dot area data from the standard metal paste dot area in the process data packet.

[0159] Preferably, the data determination module includes:

[0160] A grayscale extraction module, used for performing a grayscale extraction operation on the image to be detected to obtain grayscale information of the image to be detected;

[0161] A grayscale comparison module is used to compare the grayscale information of the image to be detected with the preset grayscale information, correct the data at the grayscale difference to the standard value, and determine the accurate grayscale data of the image to be detected;

[0162] A binarization processing module, used for performing binarization processing according to the accurate grayscale data to obtain binary image data of the image to be detected;

[0163] A binary image module, used to obtain metal paste dot data according to the binary image data;

[0164] An image matching module is used to generate a template image matching the image to be detected based on the metal paste dot data, and perform template image matching to identify the center point position of the metal paste dot data;

[0165] An area data module, used to obtain metal paste dot area data according to the center point position of the paste dot and the actual bare die diameter contained in the process data packet;

[0166] A center point module, used to obtain actual position data of the bare die, and determine the bare die center point position corresponding to the bare die from the actual position data;

[0167] The position data module is used to use the center point of the bare crystal as the positioning point, identify the deviation between the center point of the bare crystal and the center point of the paste point, obtain the relative position of the bare crystal to the metal paste point, and determine the position data of the metal paste point.

[0168] It should be noted that the machine vision positioning system in a COB die bonding process provided in an embodiment of the present invention is used to execute all process steps of a machine vision positioning method in a COB die bonding process of the above embodiment. The working principles and beneficial effects of the two correspond one to one, and thus will not be elaborated upon.

[0169] The embodiment of the present invention further provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a machine vision positioning program in a COB die bonding process. When the processor executes the computer program, the steps in the above-mentioned machine vision positioning method embodiments in the COB die bonding process are implemented, such as Figure 1 Alternatively, when the processor executes the computer program, the functions of each module / unit in the above-mentioned device embodiments are implemented, such as the detection module.

[0170] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the electronic device.

[0171] The electronic device may be a computing device such as a desktop computer, a notebook, a PDA, and a smart tablet. The electronic device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. The electronic device may include more or fewer components than the above components, or may combine certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0172] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the electronic device, and uses various interfaces and lines to connect various parts of the entire electronic device.

[0173] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the electronic device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0174] Wherein, if the module / unit integrated in the electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0175] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0176] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A machine vision positioning method in a COB die bonding process, characterized in that: The method comprises: Obtain the image to be inspected and the process data packet; Analyze the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots; Determine the pressure control parameters in the COB die bonding process according to the process data package and the fullness of the metal paste dots; Perform detection based on the image to be detected and a preset edge detection algorithm to obtain outer contour data of the image to be detected; Determine the position deviation of the bare die and the area deviation of the metal paste dots according to the outer contour data, obtain a deviation result, and generate movement compensation data for controlling the motor based on the deviation result; Generate a control instruction to control the operation of the motor during the COB die bonding process according to the pressure control parameter and the movement compensation data; The step of analyzing the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots includes: Identify the position data and size data of each metal paste dot according to the image to be detected; Based on the image processing model, the actual fullness of all the cream dots is evaluated; Comparing the actual fullness with the standard fullness for similarity to obtain the fullness of the metal paste dots; The step of determining the position deviation of the bare die and the area deviation of the metal paste dots according to the outer contour data to obtain the deviation result includes: Determine the position data of the bare die and the area data of the metal paste dots according to the outer contour data; Subtracting the position data from the standard position parameters in the process data packet to obtain a position deviation result; Subtracting the metal paste dot area data from the standard metal paste dot area in the process data packet to obtain an area deviation result; The deviation results include position deviation results and area deviation results.

2. The machine vision positioning method in the COB die bonding process according to claim 1, characterized in that: The step of determining the position data of the bare die and the area data of the metal paste dots according to the outer contour data includes: Performing a grayscale extraction operation on the image to be detected to obtain grayscale information of the image to be detected; Comparing the grayscale information of the image to be detected with the preset grayscale information, correcting the data at the grayscale difference to the standard value, and determining the accurate grayscale data of the image to be detected; Performing binarization processing according to the accurate grayscale data to obtain binary image data of the image to be detected; Obtaining metal paste dot data according to the binary image data; Taking the metal paste dot data as a base point, generating a template image matching the image to be detected, and performing template image matching to identify the center point position of the paste dot of the metal paste dot data; According to the center point position of the paste point and the actual bare die diameter contained in the process data packet, the area data of the metal paste point is obtained; Acquire actual position data of the bare die, and determine the position of the bare die center point corresponding to the bare die from the actual position data; Taking the center point of the bare crystal as the positioning point, the deviation between the center point of the bare crystal and the center point of the paste point is identified to obtain the relative position of the bare crystal to the metal paste point and determine the position data of the metal paste point.

3. The machine vision positioning method in the COB die bonding process according to claim 2, characterized in that: The obtaining of the actual position data of the bare die includes: The image position data of the bare die is obtained by detecting the edge detection algorithm and the metal paste dot area data, wherein, i From 1 to n, it represents the detected n image position data; According to the image position data and the actual die diameter contained in the process data packet, the image position data is converted by the following formulas (1) and (2) to obtain the actual die position data: （1） （2） The actual position data includes the actual width distance and the actual height distance. is the actual width distance, is the actual height distance, is the actual die diameter width, is the actual bare die diameter height, is the pixel width, is the pixel height.

4. The machine vision positioning method in the COB die bonding process according to claim 1, characterized in that: The subtraction of the metal paste dot area data from the standard metal paste dot area in the process data packet to obtain the area deviation result includes: The area deviation result is calculated by the following formula: Among them, Diviation is the metal paste dot area deviation, Actualarea is the actual metal paste dot area data, and area is the metal paste dot area data.

5. The machine vision positioning method in the COB die bonding process according to claim 1, characterized in that: The step of acquiring the image to be detected comprises: The original image data obtained by the image acquisition device is input into the pre-trained convolution model to obtain the image to be detected; The training process of the convolutional model includes: A training set is constructed based on historical image data and historical images to be detected, and the training set is input into the initial convolution model for iterative calculation. When the loss parameter value is less than the preset loss threshold, the training is determined to be completed, and a trained convolution model is obtained; the convolution model includes a convolution layer and a deconvolution layer.

6. The machine vision positioning method in the COB die bonding process according to claim 1, characterized in that: The method further comprises: According to the image to be detected, identifying outer contour edge points of a plurality of metal paste dots; Determine the center point position of the metal paste dots according to the relative position relationship between the outer contour edge points of each two metal paste dots, and determine the degree of deviation of a single metal paste dot by calculating and comparing the center point position; Generate overall error data based on the degree of deviation of all metal paste points; Dividing the overall error data into units of graphic regions to generate regional error data; The control instruction is updated according to the regional error data, so as to control the operation of the motor during the COB die bonding process according to the updated control instruction.

7. A machine vision positioning system in a COB die bonding process, used to implement the machine vision positioning method in a COB die bonding process according to any one of claims 1 to 6, characterized in that: The system comprises: A data acquisition module is used to acquire the image to be detected and the process data packet; An analysis module, used for analyzing the image to be detected and a preset image processing model to obtain the fullness of the metal paste dots; A parameter determination module, used to determine the pressure control parameters in the COB die bonding process according to the process data package and the metal paste dot fullness; A detection module, used to perform detection based on the image to be detected and a preset edge detection algorithm to obtain the outer contour data of the image to be detected; A deviation module, used for determining the position deviation of the bare die and the area deviation of the metal paste dots according to the outer contour data, obtaining a deviation result, and generating movement compensation data for controlling the motor based on the deviation result; The instruction module is used to generate a control instruction to control the operation of the motor during the COB die bonding process according to the pressure control parameter and the movement compensation data.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the machine vision positioning method in the COB die bonding process as described in any one of claims 1 to 6.

Images