Surface mounting process of CT detector module
Patent Information
- Application Number
- CN202410635553.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-22
- Publication Date
- 2024-11-12
AI Technical Summary
The traditional CT detector module mounting process has problems such as low welding accuracy, poor reliability and low production efficiency. It is mainly due to the inaccurate welding position due to the insufficient accuracy of welding equipment or the inaccurate equipment calibration.
Using real-time image data processing technology, through denoising, grayscale processing, edge detection and feature extraction, the precise position and welding direction of the circuit board and components to be welded are determined, the quality of solder joints is evaluated in real time, defect classification and adjustment strategies are formulated to ensure that Welding accuracy and stability.
It improves welding accuracy and welding stability, ensures correct placement of electronic components, timely evaluates and deals with welding quality issues, improves production efficiency and product reliability, and reduces rework and failure rates.
Smart Images

Figure CN118926644A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of industrial automation, and in particular to a mounting process of a CT detector module. Background Art
[0002] CT (Computed Tomography) detectors are key components in CT equipment, used to receive attenuation signals after X-rays penetrate objects. With the development of science and technology, the demand for CT detectors is increasing, and the requirements for production efficiency and quality are also getting higher and higher. The traditional CT detector module mounting process has problems such as low mounting accuracy, poor reliability, and low production efficiency, which cannot meet the production needs of modern CT equipment.
[0003] Patent document with publication number CN113671553A discloses an X-ray detection array pixel unit, a manufacturing process of a double-layer scintillator array, and a double-layer energy spectrum CT detector based on the pixel unit. The pixel unit includes a top-layer scintillator pixel, a bottom-layer scintillator pixel, a thin-film light filter layer, and a photosensitive array. The sub-pixels of the top-layer and bottom-layer scintillator arrays are made of scintillator materials emitting different spectra. The thin-film light filter layer corresponds to the light output surface of each scintillator pixel. The pixels of the photosensitive array are divided into two independent sub-pixel areas, and their output signals correspond to the X-ray responses of the top-layer scintillator pixels and the bottom-layer scintillator pixels, respectively, so as to effectively obtain the energy spectrum information of the incident X-rays.
[0004] This shows that the following problem exists: in the prior art, the welding position is inaccurate because the welding equipment itself is not accurate enough or the equipment is not calibrated accurately. Summary of the invention
[0005] To this end, the present invention provides a mounting process for a CT detector module, so as to overcome the problem in the prior art that the welding position is inaccurate due to the insufficient precision of the welding equipment itself or the inaccurate calibration of the equipment.
[0006] To achieve the above object, the present invention provides a mounting process for a CT detector module, comprising:
[0007] Step S1, obtaining first real-time image data including a plurality of electronic components to be welded and a circuit board, determining the real-time position of the circuit board according to the first real-time image data, and determining the relative position and welding direction of the electronic components to be welded on the circuit board based on the real-time position of the circuit board as a reference point;
[0008] Step S2, welding the electronic component to be welded to the corresponding position of the circuit board according to the relative position and welding direction of the electronic component to be welded on the circuit board to obtain second real-time image data of the initial welding of the electronic component;
[0009] Step S3, obtaining a comparison result according to the comparison between the first real-time image data and the second real-time image data;
[0010] Step S4, analyzing the quality of the solder joints on the circuit board according to the comparison result to obtain an initial soldering quality, and comparing the initial soldering quality with a preset target soldering quality to determine whether the electronic component to be soldered is completely soldered to the circuit board;
[0011] Step S5, performing defect classification on the electronic components that are not completely soldered to the circuit board to obtain a classification result;
[0012] Step S6: formulate a corresponding adjustment strategy according to the classification result.
[0013] Further, the process of determining the real-time position of the circuit board according to the first real-time image data includes:
[0014] Performing denoising and grayscale processing on the first real-time image data to obtain a first processing result;
[0015] Identify the edge line of the first real-time image data according to the first processing result to obtain a recognition result;
[0016] Performing feature extraction of edge points and corner points on the first processing result according to the recognition result to obtain a first feature result;
[0017] Matching the first feature result with the feature points on the template to obtain a matching result;
[0018] The real-time position of the circuit board in the first real-time image is determined according to the matching result.
[0019] Furthermore, the process of determining the relative position and welding direction of the electronic component to be welded on the circuit board based on the real-time position of the circuit board as a reference point includes:
[0020] Extracting the shape, texture and edge features of the electronic component to be welded to obtain a second feature result;
[0021] Identify the electronic component to be welded according to the second feature result to obtain an identification result;
[0022] Calculating the coordinates of the electronic component to be welded in the first real-time image according to the recognition result;
[0023] Converting the image coordinate system of the first real-time image into a circuit board coordinate system according to the real-time position of the circuit board;
[0024] The relative position and welding direction of the electronic component to be welded on the circuit board are calculated according to the real-time position of the circuit board and the circuit board coordinate system.
[0025] Further, the process of identifying the electronic component to be welded according to the second feature result to obtain an identification result includes:
[0026] Converting the second characteristic result into a numerical value to obtain a conversion result;
[0027] Use the transformation results as input parameters to build a recognition model;
[0028] The recognition model is trained to recognize the electronic component to be soldered.
[0029] Furthermore, the process of step S2 includes:
[0030] Planning the movement path of the welding head according to the relative position of the electronic components to be welded on the circuit board and the welding direction;
[0031] Use welding robots to perform welding operations according to the planned motion path;
[0032] The second real-time image data is acquired after welding is completed.
[0033] Furthermore, the process of step S3 includes:
[0034] Extracting features of edge, corner, shape and texture from the first real-time image data to obtain a third feature result;
[0035] Extracting features of edge, corner, shape and texture from the second real-time image data to obtain a fourth feature result;
[0036] Using a feature matching algorithm to match the third feature result and the fourth feature result to obtain a second matching result;
[0037] Calculate a transformation matrix of the first real-time image data relative to the second real-time image data according to the second matching result to obtain a first calculation result;
[0038] The similarity of the first real-time image data to the second real-time image data is evaluated according to the first calculation result to obtain an evaluation result.
[0039] Further, the process of evaluating the similarity of the first real-time image data with respect to the second real-time image data according to the first calculation result to obtain an evaluation result includes:
[0040] Calculating a pixel difference between the transformed first real-time image data and the second real-time image data to obtain a second calculation result;
[0041] Using the SSIM index to evaluate the structural similarity between the first real-time image data and the second real-time image data to obtain a second evaluation result;
[0042] Calculating mutual information between the first real-time image data and the second real-time image data to obtain a third calculation result;
[0043] A weight value is assigned to each of the second calculation result, the second evaluation result and the third calculation result, and a weighted average value is calculated according to each of the weight values.
[0044] Further, the process of comparing the initial welding quality with a preset target welding quality to determine whether the electronic component to be welded is completely welded to the circuit board includes:
[0045] If the initial welding quality is greater than or equal to the target welding quality, the electronic component to be welded is completely welded to the circuit board;
[0046] If the initial welding quality is less than the target welding quality, the electronic component to be welded is not completely welded to the circuit board.
[0047] Furthermore, the process of classifying defects of electronic components that are not completely soldered to the circuit board to obtain classification results includes:
[0048] identifying a defect of the electronic component that is not completely soldered to the circuit board according to the second real-time image data;
[0049] Marking the defect to obtain a marking result;
[0050] According to the marking results, the defects are divided into three categories: solder joint missing, solder deficiency and position shift.
[0051] Furthermore, the process of formulating a corresponding adjustment strategy according to the classification result includes:
[0052] For the defects of missing solder joints, a strategy of re-welding or repair welding is adopted;
[0053] For the defect of insufficient solder, adjust the welding parameters or increase the amount of solder;
[0054] For the position deviation defect, calibrate the position of the electronic component or adjust the position of the welding head.
[0055] Compared with the prior art, the beneficial effect of the present invention is that the present invention determines the position of the circuit board and the electronic components to be welded through real-time image data, which can improve the accuracy of welding and ensure that the electronic components are correctly placed on the corresponding positions of the circuit board. Determining the welding direction of the electronic components to be welded helps to ensure the stability and welding quality of the welding process. By comparing the real-time image data, the quality of the solder joints can be evaluated in time to ensure the quality of the welding process. Defect classification of electronic components that are not completely welded to the circuit board helps to quickly identify and handle problems and improve production efficiency. Formulating corresponding adjustment strategies based on the defect classification results can solve problems in a targeted manner and improve the reliability of the welding process.
[0056] In particular, denoising can remove noise from the image, and grayscale processing simplifies the image, which helps improve the quality and accuracy of subsequent image processing. Edge detection can reduce the erroneous identification of non-relevant features, thereby improving accuracy. Extracting feature points helps improve the accuracy of the matching step. Feature matching compares the extracted feature points with the feature points on the template to find the best match, and uses the matching results to calculate the coordinates of the circuit board in the image, thereby achieving accurate positioning.
[0057] In particular, by extracting the shape, texture and edge features of the electronic components to be soldered, the component type and position can be identified more accurately. Calculating the coordinates of the components in the first image helps optimize the soldering path and improve soldering efficiency. Converting the image coordinate system to the circuit board coordinate system enables accurate correspondence between the image information and the actual circuit board layout, which helps improve soldering accuracy and ensures that components are correctly installed.
[0058] In particular, by converting the second feature result into a numerical value, the dimension difference between the features is eliminated, so that different features can be compared and processed at the same scale. The recognition model can process image data in real time and detect welding quality problems in time.
[0059] In particular, by accurately planning the movement path of the soldering head, it can be ensured that the soldering head accurately reaches the position of the electronic component to be soldered, thereby reducing human errors. Precise control of the soldering head helps to achieve high-quality soldering, improve the reliability of solder joints and the performance of the circuit board. The pre-planned movement path can reduce unnecessary movement of the soldering head on the circuit board and increase the soldering speed.
[0060] In particular, by using a feature matching algorithm to match the features of the first real-time image data with the features of the second real-time image data, the key point correspondence between the first real-time image data and the second real-time image data can be ensured, and a high matching accuracy can be maintained even when there is a certain amount of deformation or noise in the image. The transformation matrix calculated based on the feature matching result can accurately describe the position and posture transformation of the first real-time image data relative to the second real-time image data. By evaluating the similarity between the first real-time image data and the second real-time image data, the welding quality can be quantitatively judged.
[0061] In particular, the accuracy and reliability of the evaluation can be improved by calculating pixel differences, evaluating the SSIM index, and calculating mutual information. By calculating and evaluating image similarity in real time, the quality of the welding process can be monitored in a timely manner and fed back to the control system in real time so that the welding parameters can be adjusted in time to ensure the welding quality.
[0062] In particular, initial soldering quality greater than or equal to the target soldering quality means that the solder joint has sufficient strength and stability to ensure that the electrical and mechanical connection between the electronic component and the circuit board is reliable. By identifying insufficient soldering quality, soldering parameters can be adjusted in time or bad solder joints can be repaired to prevent unqualified products from leaving the production line.
[0063] In particular, through real-time monitoring and defect classification, problems in the welding process can be discovered and corrected in a timely manner, ensuring that the welding quality meets the standards. The need for rework and re-welding can be reduced, and production efficiency can be improved. Timely detection and treatment of defects can reduce the number of poorly functioning products and improve the overall reliability of products. Equipment failures and safety risks caused by welding defects can be reduced.
[0064] In particular, the signal quality and image clarity of CT detectors can be improved by re-soldering or repairing missing solder joints to ensure the integrity of the circuit. Adjust soldering parameters or increase the amount of solder to fill insufficient solder to prevent signal leakage or loss and ensure stable performance of the detector. Calibrate component placement or adjust soldering head position to reduce mechanical instability caused by component offset. Improve the overall reliability of CT detector modules through precise soldering and positioning, reduce failure rates and maintenance costs. Quick-response defect classification and adjustment strategies help reduce production line downtime and improve production efficiency. Reduce the number of rework and re-soldering, save production costs, and increase output. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] Figure 1 A schematic diagram of a process flow of a CT detector module mounting process provided by an embodiment of the present invention;
[0066] Figure 2A schematic diagram of a process flow of a mounting process for determining the real-time position of the circuit board according to the first real-time image data in a CT detector module provided in an embodiment of the present invention;
[0067] Figure 3 A schematic diagram of the process flow of the mounting process of step S3 in the CT detector module provided in an embodiment of the present invention;
[0068] Figure 4 A logic determination diagram of a mounting process in which the initial welding quality is compared with a preset target welding quality to determine whether the electronic component to be welded is completely welded to the circuit board in a CT detector module provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0069] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0070] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.
[0071] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.
[0072] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0073] See also Figure 1 As shown, the mounting process of the CT detector module provided by the embodiment of the present invention includes:
[0074] Step S1, obtaining first real-time image data including a plurality of electronic components to be welded and a circuit board, determining the real-time position of the circuit board according to the first real-time image data, and determining the relative position and welding direction of the electronic components to be welded on the circuit board with the real-time position of the circuit board as a reference point;
[0075] Step S2, welding the electronic component to be welded to the corresponding position of the circuit board according to the relative position and welding direction of the electronic component to be welded on the circuit board to obtain second real-time image data of the initial welding of the electronic component;
[0076] Step S3, obtaining a comparison result according to the comparison between the first real-time image data and the second real-time image data;
[0077] Step S4, analyzing the quality of the solder joints on the circuit board according to the comparison result to obtain an initial soldering quality, and comparing the initial soldering quality with a preset target soldering quality to determine whether the electronic component to be soldered is completely soldered to the circuit board;
[0078] Step S5, performing defect classification on the electronic components that are not completely soldered to the circuit board to obtain a classification result;
[0079] Step S6: formulate a corresponding adjustment strategy according to the classification result.
[0080] Specifically, first, real-time image data of the circuit board and the electronic components to be welded are obtained by high-precision imaging equipment. The real-time position of the circuit board is determined by an image processing algorithm, and the relative position and welding direction of the electronic components to be welded on the circuit board are calculated with the real-time position of the circuit board as the reference point. According to the calculated relative position and welding direction, the welding operation is performed to weld the electronic components to the corresponding positions on the circuit board. During the welding process, the real-time image data after welding is obtained at the same time for comparison with the initial image data. The image data before the initial welding is compared with the image data after welding to obtain the comparison result. The purpose of the comparison is to check whether the welding process causes the position change of the electronic components or the formation of solder joints. The comparison results are analyzed to evaluate the quality of the solder joints on the circuit board. The initial welding quality is compared with the preset target welding quality to determine whether the electronic components are completely welded to the circuit board. If it is found that the quality of the solder joints does not meet the standard or the electronic components are not completely welded, the next step of defect classification and adjustment strategy formulation will be entered. Defect classification is performed on the electronic components that are not completely welded to the circuit board to distinguish different types of defects such as missing solder joints, insufficient solder or position offset. Develop corresponding adjustment strategies for different defect types, such as re-welding, adjusting welding parameters, or calibrating component placement. Adjust welding parameters based on the developed adjustment strategies.
[0081] Specifically, by determining the position of the circuit board and the electronic components to be soldered through real-time image data, the accuracy of soldering can be improved to ensure that the electronic components are correctly placed on the corresponding positions of the circuit board. Determining the soldering direction of the electronic components to be soldered helps to ensure the stability and quality of the soldering process. By comparing real-time image data, the quality of solder joints can be evaluated in a timely manner to ensure the quality of the soldering process. Defect classification of electronic components that are not fully soldered to the circuit board helps to quickly identify and handle problems and improve production efficiency. Formulating corresponding adjustment strategies based on the defect classification results can solve problems in a targeted manner and improve the reliability of the soldering process.
[0082] Specifically, if Figure 2 As shown, the process of determining the real-time position of the circuit board according to the first real-time image data includes:
[0083] Performing denoising and grayscale processing on the first real-time image data to obtain a first processing result;
[0084] Identify the edge line of the first real-time image data according to the first processing result to obtain a recognition result;
[0085] Performing feature extraction of edge points and corner points on the first processing result according to the recognition result to obtain a first feature result;
[0086] Matching the first feature result with the feature points on the template to obtain a matching result;
[0087] The real-time position of the circuit board in the first real-time image is determined according to the matching result.
[0088] Specifically, the image is denoised using a median filter to remove random noise from the image. The image is converted into a grayscale image to obtain a first processing result. The edge lines in the image are detected using a Canny edge detector. The main edge lines in the image are identified, which may represent the outer boundaries of the circuit board. Feature points, such as edge points and corner points, are extracted from the edge detection results. The extracted feature points are matched with the feature points on the known template using the SI FT algorithm. Template matching is performed using the NCC algorithm to search for the area in the image that is most similar to the template. Template matching can provide information about the position and rotation of the circuit board in the image. Based on the matching results, the coordinates of the circuit board in the image are calculated. Assume that the matching result is: x=100, y=200, which means that the position of the circuit board in the image is (100, 200).
[0089] Specifically, denoising can eliminate noise in the image, and grayscale processing simplifies the image, which helps improve the quality and accuracy of subsequent image processing. Edge detection can reduce the erroneous identification of non-relevant features, thereby improving accuracy. Extracting feature points helps improve the accuracy of the matching step. Feature matching compares the extracted feature points with the feature points on the template to find the best match, and uses the matching results to calculate the coordinates of the circuit board in the image, thereby achieving precise positioning.
[0090] Specifically, the process of determining the relative position and welding direction of the electronic component to be welded on the circuit board based on the real-time position of the circuit board as a reference point includes:
[0091] Extracting the shape, texture and edge features of the electronic component to be welded to obtain a second feature result;
[0092] Identify the electronic component to be welded according to the second feature result to obtain an identification result;
[0093] Calculating the coordinates of the electronic component to be welded in the first real-time image according to the recognition result;
[0094] Converting the image coordinate system of the first real-time image into a circuit board coordinate system according to the real-time position of the circuit board;
[0095] The relative position and welding direction of the electronic component to be welded on the circuit board are calculated according to the real-time position of the circuit board and the circuit board coordinate system.
[0096] Specifically, a Canny edge detector is used to extract the edge of the electronic component, a local binary pattern (LBP) or a Gabor filter is used to analyze the texture, and a Hu invariant moment is used to describe the shape to obtain a second feature result. According to the second feature result, a machine learning algorithm or a deep learning model is used to identify the electronic component to be welded. According to the recognition result, the coordinates of the electronic component to be welded in the first image are calculated. This is usually performed based on the image coordinate system. For example, an image processing algorithm can be used to find the center point or specific key points of the component and give the positions of these points in the image. The image coordinate system of the first image is converted to the circuit board coordinate system using a homography matrix or an affine transformation. According to the real-time position of the circuit board and the circuit board coordinate system, the relative position and welding direction of the electronic component to be welded on the circuit board are calculated. For example, the welding direction can be determined by calculating the distance and angle between the center point of the component and the center point of the circuit board.
[0097] Specifically, by extracting the shape, texture, and edge features of the electronic components to be soldered, the component type and position can be more accurately identified. Calculating the coordinates of the components in the first image helps optimize the soldering path and improve soldering efficiency. Converting the image coordinate system to the circuit board coordinate system enables accurate correspondence between the image information and the actual circuit board layout, which helps improve soldering accuracy and ensures that components are correctly installed.
[0098] Specifically, the process of identifying the electronic component to be welded according to the second feature result to obtain an identification result includes:
[0099] Converting the second characteristic result into a numerical value to obtain a conversion result;
[0100] Use the transformation results as input parameters to build a recognition model;
[0101] The recognition model is trained to recognize the electronic component to be soldered.
[0102] Specifically, the extracted shape, texture, and edge features are converted into numerical representations. For example, a feature extraction algorithm can be used to directly output feature vectors, or features can be converted into numerical values using a manual coding method. A machine learning algorithm is selected, such as a support vector machine (SVM), random forest, convolutional neural network (CNN), or other deep learning models. The converted feature results are used as input parameters to build a recognition model. The architectural parameters of the model, such as the number of layers, number of neurons, activation functions, etc., are determined, which may involve multiple trials and errors to find the best configuration. The recognition model is trained using a labeled image dataset. During the training process, the model learns how to identify different electronic components from the feature vectors. A cross-validation technique is applied to evaluate the performance of the model, and the model parameters are adjusted to optimize the results. The performance of the model is evaluated on an independent test set, using metrics such as accuracy, recall, and F1 score. If the model performance does not meet the requirements, the architectural parameters of the model are adjusted and the model is rebuilt. The trained model is applied to image data from the actual welding process. The electronic components in the image are recognized in real time and the recognition results are output.
[0103] Specifically, by converting the second feature result into a numerical value, the dimension difference between the features is eliminated, so that different features can be compared and processed at the same scale. The recognition model can process image data in real time and detect welding quality problems in time.
[0104] Specifically, the process of step S2 includes:
[0105] Planning the movement path of the welding head according to the relative position of the electronic components to be welded on the circuit board and the welding direction;
[0106] Use welding robots to perform welding operations according to the planned motion path;
[0107] The second real-time image data is acquired after welding is completed.
[0108] In this embodiment, it is assumed that there is a simple circuit board with a resistor and a capacitor that need to be welded. A circuit design software is used to generate a design file (such as a PCB file) for the circuit board. The position information of the resistor and the capacitor is extracted from the design file, assuming that the resistor is located in the upper left corner of the circuit board and the capacitor is located in the lower right corner. Determine the position and welding direction of the electronic components that the welding head needs to contact. Assume that the welding head has a rotation function and adjust the welding direction as needed. According to the position and welding direction of the electronic components, calculate the path that the welding head needs to move. Determine each key point that the welding head needs to reach, including the starting point, welding point, and end point. Assume that the planned path is that the welding head moves to the resistor position first and then to the capacitor position. Convert the motion path of the welding head into a specific motion instruction. Assume that the motion instruction includes parameters such as the position (X coordinate, Y coordinate) of the welding head, the rotation angle, and the welding time. Use a welding robot to perform the welding operation according to the planned motion path: Start the welding robot and let it move the welding head according to the planned motion path. After the welding head reaches the predetermined position, activate the welding function, such as heating the welding head or spraying welding material. Keep the welding head at the predetermined position for a period of time to ensure that the solder is fully melted and connects the electronic components and the circuit board. After welding is completed, the welding point is photographed using a camera to obtain second real-time image data.
[0109] Specifically, by accurately planning the motion path of the soldering head, it can be ensured that the soldering head accurately reaches the position of the electronic component to be soldered, thereby reducing human errors. Precise control of the soldering head helps to achieve high-quality soldering, improve the reliability of solder joints and the performance of the circuit board. The pre-planned motion path can reduce unnecessary movement of the soldering head on the circuit board and increase the soldering speed.
[0110] Specifically, if Figure 3 As shown, the process of step S3 includes:
[0111] Step S31, extracting the features of edge, corner point, shape and texture from the first real-time image data to obtain a third feature result;
[0112] Step S32, extracting the features of edge, corner point, shape and texture from the second real-time image data to obtain a fourth feature result;
[0113] Step S33, using a feature matching algorithm to match the third feature result with the fourth feature result to obtain a second matching result;
[0114] Step S34, calculating a transformation matrix of the first real-time image data relative to the second real-time image data according to the second matching result to obtain a first calculation result;
[0115] Step S35: evaluating the similarity of the first real-time image data with respect to the second real-time image data according to the first calculation result to obtain an evaluation result.
[0116] Specifically, edge detection algorithms (such as Canny detectors), corner detection algorithms (such as Harris corner detection), shape descriptors, and texture analysis are used to extract edge, corner, shape, and texture features from the first real-time image data and the second image data to obtain a third feature result and a fourth feature result. A feature matching algorithm, such as SI FT, SURF, or ORB, is applied to compare the degree of matching between the first set of feature results and the second set of feature results. The RANSAC algorithm is used to filter out mismatches and improve the accuracy of matching. Based on the feature matching results, a homography matrix or a similarity transformation is used to calculate the transformation matrix of the first real-time image data relative to the second real-time image data. A method for minimizing errors, such as the Levenberg-Marquardt algorithm, is applied to optimize the calculation of the transformation matrix. The structural similarity index (SSIM) or other similarity indicators are used to evaluate the similarity between the transformed first real-time image data and the second real-time image data. The similarity evaluation can be a numerical value or a classification result, such as high similarity, medium similarity, or low similarity.
[0117] Specifically, by using a feature matching algorithm to match the features of the first real-time image data with the features of the second real-time image data, the key point correspondence between the first real-time image data and the second real-time image data can be ensured, and a high matching accuracy can be maintained even when there is a certain amount of deformation or noise in the image. The transformation matrix calculated based on the feature matching results can accurately describe the position and posture transformation of the first real-time image data relative to the second real-time image data. By evaluating the similarity between the first real-time image data and the second real-time image data, the welding quality can be quantitatively judged.
[0118] Specifically, the process of evaluating the similarity of the first real-time image data with respect to the second real-time image data according to the first calculation result to obtain an evaluation result includes:
[0119] Calculating a pixel difference between the transformed first real-time image data and the second real-time image data to obtain a second calculation result;
[0120] Using the SSIM index to evaluate the structural similarity between the first real-time image data and the second real-time image data to obtain a second evaluation result;
[0121] Calculating mutual information between the first real-time image data and the second real-time image data to obtain a third calculation result;
[0122] A weight value is assigned to each of the second calculation result, the second evaluation result and the third calculation result, and a weighted average value is calculated according to each of the weight values.
[0123] Specifically, the transformed first real-time image data and the second real-time image data are compared pixel by pixel, and the difference of each corresponding pixel is calculated. The difference of all pixel points is squared and summed to obtain the total pixel difference, which can be used to measure the degree of difference between the first real-time image data and the second real-time image data. The second calculation result can be a numerical value representing the total amount of pixel difference or the average value of pixel difference. The calculation formula of SSIM index is used to compare the mean, variance and covariance of the first real-time image data and the second real-time image data, and calculate their structural similarity values. The range of SSIM value is -1 to 1, close to 1 means that the images are very similar, and close to -1 means that the images are very different. The joint probability distribution and marginal probability distribution of the first real-time image data and the second real-time image data are calculated, and then the calculation formula of mutual information is used to obtain the mutual information value. According to the requirements and experience of welding process, a weight value is assigned to each indicator. The weight value can be set according to the importance of the indicator. For example, if the pixel difference is more important in evaluating welding quality, a larger weight value can be assigned. The calculation result of each indicator is multiplied by its corresponding weight value. All products are added together and then divided by the sum of the weights to obtain a weighted average value. The weighted average value can be used to comprehensively evaluate the similarity between the first real-time image data and the second real-time image data.
[0124] Specifically, the accuracy and reliability of the evaluation can be improved by calculating pixel differences, evaluating the SSIM index, and calculating mutual information. By calculating and evaluating image similarity in real time, the quality of the welding process can be monitored in a timely manner and fed back to the control system in real time so that the welding parameters can be adjusted in time to ensure the welding quality.
[0125] Specifically, if Figure 4 As shown, the process of comparing the initial welding quality with the preset target welding quality to determine whether the electronic component to be welded is completely welded to the circuit board includes:
[0126] If the initial welding quality is greater than or equal to the target welding quality, the electronic component to be welded is completely welded to the circuit board;
[0127] If the initial welding quality is less than the target welding quality, the electronic component to be welded is not completely welded to the circuit board.
[0128] Specifically, the preset target welding quality refers to the quality standard of the solder joints expected to be achieved during the welding process, which is a collection of parameters and conditions for evaluating the quality of solder joints. In this embodiment, the target welding quality is that there is no visible cold solder joint, cold solder joint or solder bridging phenomenon. Under the specified test conditions, the solder joint can withstand a certain tensile force without falling off.
[0129] Specifically, if the initial soldering quality is greater than or equal to the target soldering quality, it means that the solder joint has sufficient strength and stability to ensure that the electrical and mechanical connection between the electronic component and the circuit board is reliable. By identifying insufficient soldering quality, the soldering parameters can be adjusted in time or the bad solder joints can be repaired to prevent unqualified products from leaving the production line.
[0130] Specifically, the process of classifying defects of electronic components that are not completely soldered to the circuit board to obtain classification results includes:
[0131] identifying a defect of the electronic component that is not completely soldered to the circuit board according to the second real-time image data;
[0132] Marking the defect to obtain a marking result;
[0133] According to the marking results, the defects are divided into three categories: solder joint missing, solder deficiency and position shift.
[0134] Specifically, the second real-time image data is analyzed using image processing software to identify the location and properties of incompletely soldered electronic components. The identified defects may include missing solder joints, insufficient solder, solder joint offset, cold welding, pores, cracks, etc. The identified defects are marked on the image for further analysis and processing. For example, for the marking of missing solder joints, the defect of missing solder joints is marked in red, the defect of insufficient solder is marked in yellow, and the defect of position offset is marked in green. For the identified missing solder joint defects, the area of missing solder joints is circled by software tools. The position coordinates of the missing solder joints, as well as their specific positions and welding directions on the circuit board are recorded. For the defect of insufficient solder, the insufficient part of the solder contact area with the component and the circuit board is determined by image analysis. The degree of insufficient solder is quantified by calculating the ratio of the solder contact area to the theoretical contact area. For example, the solder contact area on a solder joint is only 80% of the theoretical contact area. For the defect of position offset, the deviation of the actual position of the component relative to the circuit board from the theoretical position is determined by image analysis. Measure the distance and direction of the offset, for example, a component is offset 0.5 mm relative to the circuit board and is 5 degrees to the right.
[0135] Specifically, through real-time monitoring and defect classification, problems in the welding process can be discovered and corrected in a timely manner to ensure that welding quality meets standards. The need for rework and re-welding can be reduced, and production efficiency can be improved. Timely detection and treatment of defects can reduce the number of poorly functioning products and improve the overall reliability of products. Equipment failures and safety risks caused by welding defects can be reduced.
[0136] Specifically, the process of formulating a corresponding adjustment strategy according to the classification result includes:
[0137] For the defects of missing solder joints, a strategy of re-welding or repair welding is adopted;
[0138] For the defect of insufficient solder, adjust the welding parameters or increase the amount of solder;
[0139] For defects such as position shift, calibrate the position of the electronic component or adjust the position of the welding head.
[0140] In this embodiment, it is assumed that during the production of the CT detector module, the following defects are detected by an automatic optical inspection (AOI) system:
[0141] Missing solder joints: There are 10 solder joints that are not properly soldered, and these solder joints are distributed on 5 different components.
[0142] Insufficient solder defects: There are 8 solder joints with insufficient solder, which are also distributed on 5 different components.
[0143] Position shift defect: The positions of three components were shifted by approximately 0.5 mm to 1 mm.
[0144] To address these deficiencies, the following adjustment strategies are adopted:
[0145] Missing solder joints: For these 10 solder joints, a re-soldering strategy was adopted. A welding robot was used to re-solder these solder joints to ensure that the welding quality met the standards. After re-soldering, an AOI system was used to inspect again to ensure that all solder joints were properly soldered.
[0146] Insufficient solder defects: For these 8 solder joints, the first real-time image data and the second real-time image data will adjust the welding parameters and increase the amount of solder. Adjust the parameters of the welding robot and increase the amount of solder spray to ensure that the solder can fully cover the solder joints. After adjustment, use the AOI system to inspect again to ensure that the amount of solder meets the requirements.
[0147] Position deviation defects: For these three components with position deviation, the electronic component position will be calibrated or the position of the soldering head will be adjusted. Use the calibration tool to fine-tune the component position to ensure that it is accurately placed on the PCB board. Adjust the position of the soldering head to ensure that the components can be accurately aligned during the soldering process. After adjustment, use the AOI system to verify whether the component position is correct.
[0148] Specifically, by re-soldering or repairing missing solder joints, the integrity of the circuit is ensured, thereby improving the signal quality and image clarity of the CT detector. Adjusting welding parameters or increasing the amount of solder to fill insufficient solder prevents signal leakage or loss and ensures stable performance of the detector. Calibrate component placement or adjust the position of the soldering head to reduce mechanical instability caused by component offset. Through precise welding and positioning, the overall reliability of the CT detector module is improved, and the failure rate and maintenance cost are reduced. Quick-response defect classification and adjustment strategies help reduce production line downtime and improve production efficiency. Reducing the number of rework and re-soldering saves production costs and improves output.
[0149] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.
[0150] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A mounting process for a CT detector module, characterized in that: include: Step S1, obtaining first real-time image data including a plurality of electronic components to be welded and a circuit board, determining the real-time position of the circuit board according to the first real-time image data, and determining the relative position and welding direction of the electronic components to be welded on the circuit board with the real-time position of the circuit board as a reference point; Step S2, welding the electronic component to be welded to the corresponding position of the circuit board according to the relative position and welding direction of the electronic component to be welded on the circuit board to obtain second real-time image data of the initial welding of the electronic component; Step S3, obtaining a comparison result according to the comparison between the first real-time image data and the second real-time image data; Step S4, analyzing the quality of the solder joints on the circuit board according to the comparison result to obtain an initial soldering quality, and comparing the initial soldering quality with a preset target soldering quality to determine whether the electronic component to be soldered is completely soldered to the circuit board; Step S5, performing defect classification on the electronic components that are not completely soldered to the circuit board to obtain a classification result; Step S6: formulate a corresponding adjustment strategy according to the classification result.
2. The mounting process of the CT detector module according to claim 1, characterized in that: The process of determining the real-time position of the circuit board according to the first real-time image data includes: Performing denoising and grayscale processing on the first real-time image data to obtain a first processing result; Identify the edge line of the first real-time image data according to the first processing result to obtain a recognition result; Performing feature extraction of edge points and corner points on the first processing result according to the recognition result to obtain a first feature result; Matching the first feature result with the feature points on the template to obtain a matching result; The real-time position of the circuit board in the first real-time image is determined according to the matching result.
3. The mounting process of the CT detector module according to claim 2, characterized in that: The process of determining the relative position and welding direction of the electronic component to be welded on the circuit board based on the real-time position of the circuit board as a reference point includes: Extracting the shape, texture and edge features of the electronic component to be welded to obtain a second feature result; Identify the electronic component to be welded according to the second feature result to obtain an identification result; Calculating the coordinates of the electronic component to be welded in the first real-time image according to the recognition result; Converting the image coordinate system of the first real-time image into a circuit board coordinate system according to the real-time position of the circuit board; The relative position and welding direction of the electronic component to be welded on the circuit board are calculated according to the real-time position of the circuit board and the circuit board coordinate system.
4. The mounting process of the CT detector module according to claim 3, characterized in that: The process of identifying the electronic component to be welded according to the second feature result to obtain an identification result includes: Converting the second characteristic result into a numerical value to obtain a conversion result; Use the transformation results as input parameters to build a recognition model; The recognition model is trained to recognize the electronic component to be soldered.
5. The mounting process of the CT detector module according to claim 4, characterized in that: The process of step S2 includes: Planning the movement path of the welding head according to the relative position of the electronic components to be welded on the circuit board and the welding direction; Use welding robots to perform welding operations according to the planned motion path; The second real-time image data is acquired after welding is completed.
6. The mounting process of the CT detector module according to claim 5, characterized in that: The process of step S3 includes: Extracting features of edge, corner, shape and texture from the first real-time image data to obtain a third feature result; Extracting features of edge, corner, shape and texture from the second real-time image data to obtain a fourth feature result; Using a feature matching algorithm to match the third feature result and the fourth feature result to obtain a second matching result; Calculate a transformation matrix of the first real-time image data relative to the second real-time image data according to the second matching result to obtain a first calculation result; The similarity of the first real-time image data to the second real-time image data is evaluated according to the first calculation result to obtain an evaluation result.
7. The mounting process of the CT detector module according to claim 6, characterized in that: The process of evaluating the similarity of the first real-time image data to the second real-time image data according to the first calculation result to obtain an evaluation result includes: Calculating a pixel difference between the transformed first real-time image data and the second real-time image data to obtain a second calculation result; Using the SSIM index to evaluate the structural similarity between the first real-time image data and the second real-time image data to obtain a second evaluation result; Calculating mutual information between the first real-time image data and the second real-time image data to obtain a third calculation result; A weight value is assigned to each of the second calculation result, the second evaluation result and the third calculation result, and a weighted average value is calculated according to each of the weight values.
8. The mounting process of the CT detector module according to claim 7, characterized in that: The process of comparing the initial welding quality with the preset target welding quality to determine whether the electronic component to be welded is completely welded to the circuit board includes: If the initial welding quality is greater than or equal to the target welding quality, the electronic component to be welded is completely welded to the circuit board; If the initial welding quality is less than the target welding quality, the electronic component to be welded is not completely welded to the circuit board.
9. The mounting process of the CT detector module according to claim 8, characterized in that: The process of defect classification of electronic components that are not completely soldered to the circuit board to obtain classification results includes: identifying a defect of the electronic component that is not completely soldered to the circuit board according to the second real-time image data; Marking the defect to obtain a marking result; According to the marking results, the defects are divided into three categories: solder joint missing, solder deficiency and position shift.
10. The mounting process of the CT detector module according to claim 9, characterized in that: The process of formulating corresponding adjustment strategies according to the classification results includes: For the defects of missing solder joints, a strategy of re-welding or repair welding is adopted; For the defect of insufficient solder, adjust the welding parameters or increase the amount of solder; For defects such as position shift, calibrate the position of the electronic component or adjust the position of the welding head.
Citation Information
Patent Citations
Flexible welding robot system and welding method thereof
CN110524581A
Industrial robot instruction sending system
CN117564557A
Solder inspection method
JP2014102206A
Registration apparatus and registering mehtod for surface mounting process
TW200905428A
Cited By
Circuit board automatic scoring method and device, electronic equipment and storage medium
CN120746994A
