Welding method, device and equipment for MiniLED and medium

Abstract

The invention relates to a welding method, device and equipment for a Mini LED and a medium. The method comprises the steps that welding characteristics of the Mini LED are obtained, a welding area of the Mini LED is divided based on the welding characteristics, and a plurality of welding subareas are generated; welding data and environment change data of the welding partitions are obtained, wherein the welding data comprise welding temperature, welding pressure, images, gas flow and welding time; based on the welding data and the environment change data, a pre-trained welding model is adopted for analysis, and a welding parameter adjusting strategy corresponding to the welding partition is generated; and based on the welding parameter adjusting strategy, the welding parameters of the Mini LED are correspondingly adjusted. The welding device has the effect of improving the welding quality.

Description

Technical Field

[0001] The present application relates to the technical field of microelectronic packaging, and in particular, to a welding method, device, equipment and medium for MiniLED. Background Art

[0002] Currently, with the increasingly wide application of MiniLED display technology, the welding technology of MiniLED circuit boards has become one of the key factors restricting its production efficiency and quality. Traditional welding control methods have problems such as uneven temperature distribution and unstable welding quality, resulting in easy occurrence of voids in solder joints, affecting the display effect and reliability of MiniLED. Existing temperature control systems mainly rely on preset temperature curves and material properties, but there is still a large room for optimization when facing high-density solder joints and complex processes.

[0003] The above-mentioned existing technical solutions have the following defects: During the welding process, the temperature in the solder joint area is easily affected by the substrate, environment and welding equipment parameters, resulting in uneven temperature distribution, thus affecting the welding quality, so there is room for improvement. Summary of the Invention

[0004] In order to improve the welding quality, the present application provides a welding method, device, equipment and medium for MiniLED.

[0005] The first invention object of the present application is achieved through the following technical solutions: A welding method for MiniLED, the welding method for MiniLED includes: Obtain the welding characteristics of MiniLED, and based on the welding characteristics, divide the welding area of the MiniLED to generate a plurality of welding partitions; Obtain the welding data and environmental change data of the welding partition, and the welding data includes welding temperature, welding pressure, image, gas flow rate and welding time; Based on the welding data and environmental change data, use a pre-trained welding model for analysis to generate a welding parameter adjustment strategy corresponding to the welding partition; Based on the welding parameter adjustment strategy, correspondingly adjust the welding parameters of the MiniLED.

[0006] By adopting the above technical solution, by obtaining the welding characteristics of MiniLED and dividing the welding area based on the welding characteristics to generate multiple welding zones, it can accurately reflect the spatial distribution and process condition differences of each part within the welding area, thereby providing a basis for the targeted regulation of parameters within subsequent zones; by collecting the welding data and environmental change data of the welding zones in real time, it can monitor the dynamic changes of key process parameters during the welding process in real time, thereby timely feedback the on-site status and ensure the stability of the welding process; by analyzing based on the welding data and environmental change data using a pre-trained welding model to generate a welding parameter adjustment strategy corresponding to the welding zones, it can automatically predict the optimal welding parameters for each zone, thereby achieving adaptive control and closed-loop optimization; by correspondingly adjusting the welding parameters of MiniLED based on the welding parameter adjustment strategy, it can adjust the welding process in real time, improve the solder joint quality and reduce defects such as voids, thereby improving the consistency and reliability of the product.

[0007] In one example, the present application can be further configured as: the obtaining of the welding characteristics of MiniLED includes: Using a camera module to collect the laser spot image and the welding area image of the welding area of the MiniLED, and extracting the spot feature information from the laser spot image through an image processing algorithm, where the spot feature information includes the edge, shape, and size parameters of the spot; Collecting the temperature data of the welding area of the MiniLED through an infrared thermal imager to generate a temperature distribution map; Using a solder joint detection algorithm to process the welding area image, extracting the solder joint feature information in the welding area image, and statistically analyzing the feature information to determine the solder joint density and solder joint distribution, where the solder joint characteristic information includes contour and position data; Fusing the spot feature information, temperature distribution map, solder joint density, and solder joint distribution to generate the welding characteristics.

[0008] By adopting the above technical solution, by using the camera module to collect the laser spot image and the welding area image of the MiniLED welding area, and by using the image processing algorithm to extract the edge, shape and size parameters of the spot, the geometric features of the laser spot can be accurately obtained, thus providing an accurate data basis for the spatial distribution analysis of the welding area; by using the thermal imager to collect the temperature data of the welding area and generate the temperature distribution map, the temperature gradient and temperature distribution of each area can be reflected in real time, so that the subsequent adjustment of temperature control is more scientific and precise; by using the solder joint detection algorithm to process the welding area image, extracting the contour and position data of the solder joints, and determining the solder joint density and distribution through statistical analysis, the quality and distribution characteristics of the solder joints can be quantified, thus providing an objective basis for the evaluation of the welding effect; by fusing the spot feature information, the temperature distribution map and the solder joint density and distribution, a comprehensive welding feature can be generated, and then comprehensive and reliable multi-dimensional data support can be provided for the subsequent fine division of the welding area and the intelligent parameter adjustment.

[0009] In one example of the present application, it can be further configured as: based on the welding features, dividing the welding area of the MiniLED to generate a plurality of welding partitions, including: Based on the spot feature information in the welding features, using the image segmentation algorithm to preliminarily divide the welding area to obtain a plurality of welding sub-areas; Based on the temperature distribution map, solder joint density and solder joint distribution in the welding features, using the k-means clustering algorithm to adjust the welding sub-areas to generate the welding partitions.

[0010] By adopting the above technical solution, by using the image segmentation algorithm to preliminarily divide the welding area based on the spot feature information, the overall area can be divided into a plurality of welding sub-areas, thus providing a preliminary distribution basis for further optimizing the partition; by using the k-means clustering algorithm to adjust the preliminary division result based on the temperature distribution map and the solder joint density and distribution, the welding sub-areas can be further refined to form a plurality of welding partitions that are homogeneous in process conditions such as temperature and solder joint density, thus improving the welding consistency and quality stability within each partition.

[0011] In one example of the present application, it can be further configured as: obtaining the welding data and environmental change data of the welding partitions includes: Determining the real-time temperature data of the welding partitions from the temperature distribution map; Using a sensor group to collect the pressure data and protective gas flow data of the welding partitions, the sensor group includes a pressure sensor and a gas flow sensor; Performing recognition and analysis on the welding area image to obtain the solder joint morphology features of the welding partitions; Record the welding operation time of each stage of the welding zone to obtain the welding time of the welding zone; Obtain the environmental change data of the environment where the MiniLED is located through an environmental sensor, and the environmental change data includes environmental temperature change data and environmental humidity change data.

[0012] By adopting the above technical solutions, by determining the real-time temperature data of each welding zone from the temperature distribution map, the dynamic distribution of the temperature within the zone can be obtained, so as to timely adjust the heating parameters to maintain the optimal welding temperature; by using the sensor group to collect the pressure data and protective gas flow data of each zone, the pressure and gas protection status during the welding process can be monitored in real time, so as to ensure the stability of the welding environment; by identifying and analyzing the welding area image to extract the solder joint morphology characteristics, the actual formation of the solder joint can be monitored, so as to timely detect solder joint defects; by recording the welding operation time of each stage, the time parameters during the welding process can be accurately grasped, so as to ensure that each process stage is executed according to the predetermined duration; by obtaining the change data of the environmental temperature and humidity in the closed environment through the environmental sensor, the minute environmental fluctuations can be detected, so as to provide auxiliary information for real-time adjustment and ensure that the welding process is not affected by environmental anomalies.

[0013] In one example of the present application, it can be further configured that: before analyzing using the pre-trained welding model based on the welding data and environmental change data, the welding method for MiniLED further includes: Collect historical welding data, where the historical welding data includes historical welding temperature, welding pressure, image, gas flow, welding time data, and the corresponding welding effect; Preprocess the historical welding data to generate a model training set, and the preprocessing includes data normalization, denoising processing, and feature extraction; Based on the model training set, train using a convolutional neural network, and use the cross-validation method to optimize the parameters during the model training process to obtain the pre-trained welding model.

[0014] By adopting the above technical solutions, by collecting historical welding data, including historical welding temperature, welding pressure, image, gas flow, welding time, and the corresponding welding effect, a comprehensive historical data set can be established, so as to provide a rich reference basis for model training; by preprocessing the historical data, the accuracy and consistency of the data can be improved, so as to construct a high-quality model training set; by training using a convolutional neural network based on the training set and using the cross-validation method to optimize the parameters, a pre-trained welding model can be obtained, and this model can accurately map the non-linear relationship between the welding data and the welding effect, so as to provide a scientific basis for real-time welding parameter adjustment.

[0015] In one example, the present application can be further configured as follows: The method of analyzing the welding data and environmental change data and using a pre-trained welding model to generate a welding parameter adjustment strategy corresponding to the welding zone includes: Based on the environmental change data, using an incremental learning method, online parameter update and adjustment are performed on the pre-trained welding model to obtain an updated welding model; Input the welding data into the updated welding model for analysis to generate a welding parameter adjustment strategy corresponding to the welding zone.

[0016] By adopting the above technical solution, by analyzing the welding data and environmental change data collected in real time using a pre-trained welding model, and performing online parameter update and adjustment based on the environmental change data using an incremental learning method, the model can be made to adapt to small fluctuations in the environment in real time, thereby maintaining the accuracy of model prediction; by analyzing the real-time welding data using the updated model to generate a welding parameter adjustment strategy corresponding to each welding zone, the optimal welding parameters can be dynamically predicted, thereby realizing the adaptive optimization of the process and further improving the welding quality and consistency.

[0017] In one example, the present application can be further configured as follows: The welding method for MiniLED further includes: Compare and analyze the welding data and environmental change data with a preset safety threshold to generate a comparison and analysis result; If the comparison and analysis result exceeds the preset safety threshold, generate a fault warning signal, and extract the abnormal key parameters from the analysis result; Based on the abnormal key parameters, generate a fault report and transmit the fault report to the user terminal.

[0018] By adopting the above technical solution, by comparing and analyzing the welding data and environmental change data collected in real time with a preset safety threshold, abnormal changes during the welding process can be monitored in real time, and a fault warning signal can be generated in a timely manner, thereby preventing welding defects caused by abnormal parameters; by extracting abnormal key parameters from the comparison and analysis result and generating a detailed fault report, a basis for fault diagnosis can be provided for the operator, so that the maintenance personnel can respond and adjust quickly to ensure the stable operation of the system; by transmitting the fault report to the user terminal to realize remote monitoring and online intervention, the safety and reliability of the entire welding system can be improved, and data support can be provided for subsequent process improvement.

[0019] The above second invention object of the present application is achieved by the following technical solutions: A soldering device for MiniLED, the soldering device for MiniLED includes: A feature acquisition module, configured to acquire the soldering features of the MiniLED, and based on the soldering features, divide the soldering area of the MiniLED to generate a plurality of soldering partitions; A data acquisition module, configured to acquire the soldering data and environmental change data of the soldering partition, where the soldering data includes soldering temperature, soldering pressure, image, gas flow rate, and soldering time; A model analysis module, configured to analyze based on the soldering data and environmental change data by using a pre-trained soldering model to generate a soldering parameter adjustment strategy corresponding to the soldering partition; An adjustment module, configured to correspondingly adjust the soldering parameters of the MiniLED based on the soldering parameter adjustment strategy.

[0020] By adopting the above technical solution, by acquiring the soldering features of the MiniLED and dividing the soldering area based on the soldering features to generate a plurality of soldering partitions, it can accurately reflect the spatial distribution and process condition differences of each part in the soldering area, thereby providing a basis for the targeted regulation of parameters within subsequent partitions; by real-time collecting the soldering data and environmental change data of the soldering partition, it can monitor the dynamic changes of key process parameters during the soldering process in real time, thereby timely feedback the on-site state and ensure the stability of the soldering process; by analyzing based on the soldering data and environmental change data by using a pre-trained soldering model to generate a soldering parameter adjustment strategy corresponding to the soldering partition, it can automatically predict the optimal soldering parameters for each partition, thereby realizing adaptive control and closed-loop optimization; by correspondingly adjusting the soldering parameters of the MiniLED based on the soldering parameter adjustment strategy, it can adjust the soldering process in real time, achieve the improvement of the solder joint quality and the reduction of defects such as voids, thereby improving the consistency and reliability of the product.

[0021] The above object three of the present application is achieved by the following technical solution: A computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where when the processor executes the computer program, the steps of the above-mentioned soldering method for MiniLED are implemented.

[0022] The above object four of the present application is achieved by the following technical solution: A computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned soldering method for MiniLED are implemented.

[0023] In summary, the present application includes the following beneficial technical effects: 1. By obtaining the welding characteristics of MiniLEDs and dividing the welding area based on the welding characteristics to generate multiple welding partitions, the spatial distribution and process condition differences of each part within the welding area can be accurately reflected, thereby providing a basis for the targeted adjustment of parameters within subsequent partitions; by collecting the welding data and environmental change data of the welding partitions in real time, the dynamic changes of key process parameters during the welding process can be monitored in real time, thereby timely feedback on the on-site status and ensuring the stability of the welding process; 2. By analyzing based on the welding data and environmental change data using a pre-trained welding model to generate a welding parameter adjustment strategy corresponding to the welding partition, the optimal welding parameters for each partition can be automatically predicted, thereby achieving adaptive control and closed-loop optimization; by correspondingly adjusting the welding parameters of MiniLEDs based on the welding parameter adjustment strategy, the welding process can be adjusted in real time, achieving an improvement in the quality of solder joints and a reduction in defects such as voids, thereby improving the consistency and reliability of the product. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a flowchart of a welding method for MiniLEDs in an embodiment of the present application; Figure 2 is an implementation flowchart of step S10 in a welding method for MiniLEDs in an embodiment of the present application; Figure 3 is another implementation flowchart of step S10 in a welding method for MiniLEDs in an embodiment of the present application; Figure 4 is an implementation flowchart of step S20 in a welding method for MiniLEDs in an embodiment of the present application; Figure 5 is an implementation flowchart of step S30 in a welding method for MiniLEDs in an embodiment of the present application; Figure 6 is another implementation flowchart of step S30 in a welding method for MiniLEDs in an embodiment of the present application; Figure 7 is an implementation flowchart of a welding method for MiniLEDs in an embodiment of the present application; Figure 8 is a principle block diagram of a welding device for MiniLEDs in an embodiment of the present application; Figure 9 is a schematic diagram of a device in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] The present application will be further described in detail below with reference to the accompanying drawings.

[0026] In one embodiment, as Figure 1 shown, the present application discloses a welding method for MiniLEDs, which specifically includes the following steps: S10: Obtain the welding characteristics of the MiniLEDs. Based on the welding characteristics, divide the welding area of the MiniLEDs to generate multiple welding zones.

[0027] Specifically, by using a high-resolution camera module to collect the laser spot image and the overall welding area image of the MiniLED welding area, and at the same time combining the temperature data collected by the thermal imager, use image processing algorithms to perform edge detection, shape analysis and size measurement on the laser spot image. For example, it is detected that the spot diameter is about 5 mm, the major axis is about 6 mm, and the minor axis is about 4 mm. Then use the solder joint detection algorithm to extract the solder joint contour and position data from the overall welding area image, and perform statistical analysis on the extracted data (such as about 20 solder joints are detected per square centimeter). Finally, fuse the spot characteristics, temperature distribution and solder joint density information, perform a preliminary division of the overall area through an image segmentation algorithm, and then use a clustering algorithm (such as k-means clustering) to adjust the preliminary division result to generate multiple welding zones, so as to ensure that the process conditions within each zone are homogeneous and provide a data basis for subsequent precise parameter adjustment.

[0028] S20: Obtain the welding data and environmental change data of the welding zones. The welding data includes welding temperature, welding pressure, image, gas flow rate, and welding time.

[0029] Specifically, use a sensor group to collect the process data in each welding zone in real time. Among them, the temperature distribution map output by the thermal imager is digitally processed to obtain the temperature data of each zone. For example, the real-time temperature of a certain zone is 180°C ± 2°C. At the same time, use a high-precision pressure sensor to collect the local pressure data within the zone, such as 0.8 MPa ± 0.05 MPa, and use a gas flow sensor to record the protective gas flow data in real time, such as 1.2 L / min ± 0.1 L / min, and collect welding image data through a high-speed camera module to record the dynamics of the solder joints. At the same time, record the operation time of each stage such as heating, heat preservation and cooling during the welding process. In addition, monitor the temperature and humidity changes in the closed environment through an environmental sensor. For example, the environmental temperature is maintained at 25°C ± 1°C, and the humidity is 50% ± 5%. All these real-time data are filtered and normalized to form a multi-dimensional state data vector as the basis for subsequent process analysis.

[0030] S30: Based on the welding data and environmental change data, use a pre-trained welding model for analysis to generate a welding parameter adjustment strategy corresponding to the welding zones.

[0031] Specifically, the real-time data of the welded zones and the environmental change data collected are used as inputs, and a pre-trained welding model is utilized for analysis. This model is constructed based on a convolutional neural network architecture and obtains a non-linear mapping relationship through training on historical welding data and welding effects. For example, the distinction between a welding effect score above 90 points and below 80 points. Feature extraction and parameter prediction are performed in combination with the real-time data, and at the same time, cross-validation and incremental learning methods are adopted for online parameter update to output the optimal welding parameter adjustment strategy corresponding to each welded zone. This strategy covers multiple key parameters such as temperature, pressure, welding time, and gas flow rate. For example, adjusting the temperature up by 2°C or down by 3°C, extending the holding time by 0.5 seconds, etc., thereby achieving process self-adaptive adjustment.

[0032] S40: Correspondingly adjust the welding parameters of the MiniLED based on the welding parameter adjustment strategy.

[0033] Specifically, according to the generated welding parameter adjustment strategy, the estimated optimal welding parameters are sent to the welding device in real time through the control interface to dynamically adjust the temperature, pressure, welding time, and gas flow rate during the MiniLED welding process. For example, adjusting the welding temperature from the original set 180°C to 182°C, the welding pressure from 0.8 MPa to 0.83 MPa, and extending the heating time to 3.2 seconds, thereby precisely controlling the welding process, ensuring stable and consistent solder joint quality, effectively reducing the welding defect rate, and improving product reliability and production efficiency.

[0034] By adopting the above technical solutions, by obtaining the welding characteristics of the MiniLED and dividing the welding area based on the welding characteristics to generate multiple welded zones, the spatial distribution and process condition differences of each part within the welding area can be accurately reflected, thereby providing a basis for the targeted regulation of parameters within subsequent zones; by collecting the welding data and environmental change data of the welded zones in real time, the dynamic changes of key process parameters during the welding process can be monitored in real time, thereby timely feedback on the on-site state and ensuring a stable welding process; by analyzing the welding data and environmental change data using a pre-trained welding model to generate the welding parameter adjustment strategy corresponding to the welded zones, the optimal welding parameters for each zone can be automatically predicted, thereby achieving self-adaptive control and closed-loop optimization; by correspondingly adjusting the welding parameters of the MiniLED based on the welding parameter adjustment strategy, the welding process can be adjusted in real time to improve the solder joint quality and reduce defects such as voids, thereby improving product consistency and reliability.

[0035] In one embodiment, as Figure 2 shown, in step S10, that is, obtaining the welding characteristics of the MiniLED, specifically includes: S11: Use the imaging module to collect the laser spot image and the welding area image of the MiniLED welding area, and extract the spot feature information from the laser spot image through an image processing algorithm. The spot feature information includes the edge, shape, and size parameters of the spot.

[0036] Specifically, use a high-resolution industrial camera fixedly installed above the welding area to collect the laser spot image and the overall welding area image of the MiniLED welding area under preset appropriate lighting conditions. During collection, use autofocus and exposure control to ensure clear images. Then input the collected images into the preprocessing module, eliminate noise through grayscale conversion and filtering, then use the Canny edge detection algorithm to extract the edges of the laser spot image, and then combine the contour detection algorithm to perform closed contour fitting on the extracted edges. Calculate the size and shape of the spot by measuring the geometric parameters of the closed contour (such as diameter, major axis, and minor axis). For example, the measured value of the spot diameter is about 5 millimeters, and further use morphological operations (such as dilation and erosion) to optimize the edge data, so as to accurately obtain the spot feature information, which includes the edge, shape, and size parameters of the spot.

[0037] S12: Use a thermal imager to collect the temperature data of the MiniLED welding area and generate a temperature distribution map.

[0038] Specifically, use a thermal imager installed around the welding area to collect the infrared temperature data of the MiniLED welding area in real time. The thermal imager converts the thermal radiation signal during the collection process and generates a thermal image. By reading the temperature values of each pixel point in the thermal image, use digital image processing technology to generate a temperature distribution map. The color or gray level in the image represents the temperature gradient. For example, the high-temperature area is displayed as red or white, and the low-temperature area is displayed as blue or black, so as to accurately reflect the temperature distribution in the welding area.

[0039] S13: Use a solder joint detection algorithm to process the welding area image, extract the solder joint feature information in the welding area image, and perform statistical analysis on the feature information to determine the solder joint density and solder joint distribution. The solder joint characteristic information includes contour and position data.

[0040] Specifically, the collected welding area image is used as input, and the image is processed by a preset solder joint detection algorithm. This algorithm combines local threshold segmentation and template matching techniques. First, the image is binarized to distinguish the solder joints from the background. Then, a contour extraction algorithm (such as edge detection combined with region filling) is used to extract the contours of the solder joints. Subsequently, the positions and areas of the extracted solder joint contours are statistically analyzed. By setting thresholds for area and shape parameters, noise is filtered, and finally, the solder joint density and solder joint distribution are determined. The solder joint characteristic information includes the contours and position information of each solder joint. For example, the density is reflected by calculating the coordinates and distribution ranges of the center points of each solder joint. When 20 solder joints are detected per square centimeter, the center position and boundary information of each solder joint are recorded simultaneously, thus forming solder joint characteristic information, including the contour and position data of the solder joints, that is, the solder joint density is 20 / cm 2 。

[0041] S14: Fuse the spot feature information, temperature distribution map, solder joint density, and solder joint distribution to generate welding features.

[0042] Specifically, the spot feature information, temperature distribution map, solder joint density, and solder joint distribution are fused. Multivariate statistical analysis methods are used for data fusion. Each data source is fused and calculated after being processed according to predefined weights and normalization. For example, the correlation between each data is extracted through principal component analysis (PCA) to generate a comprehensive welding feature vector. For example, the weight of the spot feature is set to 40%, the weight of the temperature distribution is set to 30%, and the weight of the solder joint density is set to 30%. Finally, a comprehensive welding feature vector is generated. This vector can comprehensively describe the optical, thermal, and solder joint distribution characteristics of the MiniLED welding area, providing a unified and accurate data basis for subsequent welding area division and process parameter adjustment.

[0043] In one embodiment, as Figure 3 shown, in step S10, based on the welding features, the welding area of the MiniLED is divided to generate multiple welding partitions, specifically including: S15: Based on the spot feature information in the welding features, use an image segmentation algorithm to preliminarily divide the welding area to obtain multiple welding sub-areas.

[0044] Specifically, based on the obtained spot feature information, an image segmentation algorithm is used to preliminarily divide the welding area. The algorithm divides the areas with similar spot features in the image into multiple welding sub-areas by setting appropriate thresholds and edge connection rules. For example, under the condition that the spot diameter is about 5 mm, the overall area is divided into several sub-areas to reflect the spatial distribution density of the spots, thus providing a preliminary division basis for subsequent area adjustment.

[0045] S16: Based on the temperature distribution map, solder joint density, and solder joint distribution in the welding features, the k-means clustering algorithm is used to adjust the welding sub-regions to generate welding partitions.

[0046] Specifically, based on the generated temperature distribution map and the determined solder joint density and distribution, the k-means clustering algorithm is used to adjust the welding sub-regions obtained from the preliminary division. The algorithm constructs a feature vector according to the temperature mean, solder joint density, and distribution characteristics of each sub-region, and presets the number of clustering clusters (for example, set to 4 or 5) to cluster each sub-region, and finally generates welding partitions with relatively uniform temperature and solder joint density, thereby improving the consistency of process conditions within each partition.

[0047] In one embodiment, as Figure 4 shown, in step S20, that is, obtaining the welding data and environmental change data of the welding partition, specifically including: S21: Determine the real-time temperature data of the welding partition from the temperature distribution map.

[0048] Specifically, read the temperature values of all pixels within each welding partition from the temperature distribution map generated by the thermal imager, and use statistical analysis methods to calculate the average temperature and standard deviation of the partition. For example, the average temperature of a certain partition is calculated to be 180°C, and the standard deviation is 2°C, so as to determine the real-time temperature data of the partition.

[0049] S22: Use the sensor group to collect the pressure data and shielding gas flow data of the welding partition. The sensor group includes a pressure sensor and a gas flow sensor.

[0050] Specifically, use the pressure sensor installed in the welding partition to continuously collect local pressure data, and at the same time use the gas flow sensor to record the flow data of the shielding gas (such as nitrogen or argon) in real time. For example, the pressure of a certain partition is measured to be 0.8 MPa, and the gas flow is 1.2 L / min. These data are summarized as the key process parameters within the partition through the data acquisition interface.

[0051] S23: Perform recognition and analysis on the welding area image to obtain the solder joint morphology characteristics of the welding partition.

[0052] Specifically, through the recognition and analysis of the collected welding area image, use the convolutional neural network or template matching method to detect the solder joint edges and contours, and then extract the solder joint morphology characteristics. For example, the average diameter of the solder joint is calculated to be 3 mm, and the center coordinates and boundary shapes of each solder joint are recorded, so as to obtain the solder joint morphology characteristics of the welding partition.

[0053] S24: Record the welding operation time of each stage of the welding partition to obtain the welding time of the welding partition.

[0054] Specifically, by starting a timer during the welding process, record the duration of each stage such as heating, heat preservation, and cooling in each welding zone. For example, record that the heating stage is 3 seconds, the heat preservation stage is 1 second, and the cooling stage is 2 seconds. Integrate these time data into welding time parameters to ensure that each process stage is strictly executed according to the preset duration.

[0055] S25: Obtain environmental change data of the environment where the MiniLED is located through an environmental sensor. The environmental change data includes environmental temperature change data and environmental humidity change data.

[0056] Specifically, continuously monitor the environmental temperature and humidity in the closed welding chamber through a high-precision environmental sensor installed in the welding environment, and use the digital signals output by the sensor to record and analyze the fluctuations of the environmental temperature and humidity in real time. For example, it is detected that the environmental temperature is maintained at 25°C ± 1°C and the humidity is 50% ± 5%, so as to obtain the environmental change data of the environment where the MiniLED is located and provide auxiliary information for real-time model adjustment.

[0057] In one embodiment, as Figure 5 shown, before step S30, that is, before analyzing using a pre-trained welding model based on welding data and environmental change data, this welding method for MiniLEDs further includes: S301: Collect historical welding data, which includes historical welding temperature, welding pressure, images, gas flow rate, welding time data, and the corresponding welding effects.

[0058] Specifically, automatically extract historical welding temperature data, welding pressure data, welding area images, protective gas flow rate data, and welding time data recorded in each welding stage from historical production records and process monitoring logs. At the same time, combine the finally detected welding effects (such as solder joint quality scores, void ratios, etc.), and use the data interface to summarize a historical data set containing multi-dimensional parameters and corresponding welding effects from the data warehouse. This data set can include, for example, 1000 groups of data records collected under the same process conditions in the past six months, so as to provide sufficient and representative historical samples for subsequent model training.

[0059] S302: Preprocess the historical welding data to generate a model training set. The preprocessing includes data normalization, denoising processing, and feature extraction.

[0060] Specifically, when preprocessing the collected historical welding data, first perform normalization on each numerical data item to convert data with different dimensions into standard values between 0 and 1. Then, use a filtering algorithm (such as median filtering or Kalman filtering) to remove data noise. Next, use edge detection and feature extraction algorithms to extract key features from the image data, and perform annotation processing on the extracted data. Finally, integrate the processed numerical data and image feature data into a unified set of feature vectors to form a training set for model training. This training set ensures high data quality and clear feature expression, thereby providing reliable input for model construction.

[0061] S303: Based on the model training set, use a convolutional neural network for training, and use the cross-validation method for parameter optimization during the model training process to obtain a pre-trained welding model.

[0062] Specifically, when using a convolutional neural network for training based on the training set, first input the processed image feature data into the convolutional layer for local feature extraction. For example, extract the solder joint edge and texture features through 3 layers of convolution and pooling operations. Then, fuse the output of the convolutional layer with the normalized temperature, pressure, gas flow rate, and time data in the fully connected layer, and use the backpropagation algorithm to optimize the network weights. At the same time, use the cross-validation method to iteratively tune the model parameters during the training process to ensure that the model performs stably on the validation set, and finally obtain a pre-trained welding model that can accurately predict the non-linear mapping relationship between the welding effect and process parameters.

[0063] In one embodiment, as Figure 6 shown, in step S30, that is, based on the welding data and environmental change data, use the pre-trained welding model for analysis to generate a welding parameter adjustment strategy corresponding to the welding zone, which specifically includes: S31: Based on the environmental change data, use the incremental learning method to perform online parameter update and adjustment on the pre-trained welding model to obtain an updated welding model.

[0064] Specifically, based on the real-time collected environmental change data (such as the minute fluctuation data of environmental temperature and humidity), use the incremental learning method to perform online parameter update and adjustment on the pre-trained welding model. During the data inflow process, use the newly collected data to perform local re-training on the model. For example, every 5 minutes, fuse the latest collected environmental data and the corresponding welding data into the model, and adjust the network parameters through the gradient update strategy, so that the model can adapt to the minute changes of parameters such as temperature and humidity in a closed environment in real time and maintain the prediction accuracy.

[0065] S32: Input the welding data into the updated welding model for analysis to generate a welding parameter adjustment strategy corresponding to the welding zone.

[0066] Specifically, the welding data collected in real time is input into the welding model updated by incremental learning for analysis, including the real-time temperature, pressure, gas flow rate, welding time, and the morphological characteristics of the solder joints obtained through image recognition, such as the contour, size, and position of the solder joints. The model performs feature mapping and non-linear calculations on the input data, uses the pre-defined multi-layer non-linear mapping in the model to perform feature extraction and data fusion calculations on the fused features, and outputs the welding parameter adjustment strategies corresponding to each welding zone. For example, it is predicted that a certain zone needs to increase the welding temperature by 2°C, increase the welding pressure by 0.03 MPa, extend the heat preservation time by 0.5 seconds, or adjust the gas flow rate by 0.1 L / min, so as to provide precise guidance for subsequent real-time process control.

[0067] In one embodiment, as Figure 7 shown, this welding method for MiniLEDs further includes: S50: Compare and analyze the welding data and environmental change data with the preset safety thresholds to generate a comparison and analysis result.

[0068] Specifically, compare and analyze the welding data collected in real time and the environmental change data with the preset safety thresholds. First, set the safety upper and lower limits of each key parameter. For example, the temperature safety range is 175°C to 185°C, and the pressure safety range is 0.75 MPa to 0.85 MPa. Then, use statistical analysis methods to calculate the deviation between the real-time data and the preset thresholds to generate a comparison and analysis result. This result can show whether the real-time data of each welding zone is within the safety range and mark the degree of deviation. For example, if the temperature of a certain zone is detected to be 188°C, it deviates 3°C from the preset upper limit, thus providing data support for fault warning.

[0069] S60: If the comparison and analysis result exceeds the preset safety threshold, generate a fault warning signal and extract the abnormal key parameters from the analysis result.

[0070] Specifically, when the comparison and analysis result shows that the real-time data exceeds the preset safety threshold, immediately trigger the fault warning program. This program generates a fault warning signal by calculating the percentage and duration of the deviation of each key parameter from the threshold, and at the same time extracts the abnormal key parameters from the analysis result, such as abnormal values of temperature, pressure, etc., and records the relevant data and timestamps to form a detailed fault warning report. This report can clearly indicate the abnormal parameters and possible process deviations, providing a basis for rapid response for on-site technicians.

[0071] S70: Generate a fault report based on the abnormal key parameters and transmit the fault report to the user terminal.

[0072] Specifically, based on the abnormal key parameters extracted from the fault warning, a fault report is generated by using a preset fault diagnosis rule and a data matching algorithm. The fault report includes abnormal parameter values, the amplitude of deviation from the safety threshold, the duration, and the type of welding defects that may be caused. The fault report is transmitted to the monitoring platform of the user terminal through a data interface. For example, in the case where it is detected that the temperature continuously exceeds the safety upper limit and the pressure drops abnormally, a detailed fault report is generated and sent to the remote terminal so that the operator can take corrective measures and perform maintenance in a timely manner to ensure the safety and stability of the welding process.

[0073] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0074] In one embodiment, a welding device for MiniLED is provided. The welding device for MiniLED corresponds one-to-one with the welding method for MiniLED in the above embodiment. As Figure 8 shown, the welding device for MiniLED includes a feature acquisition module, a data acquisition module, a model analysis module, and an adjustment module. The detailed description of each functional module is as follows: The feature acquisition module is used to acquire the welding features of MiniLED. Based on the welding features, the welding area of MiniLED is divided to generate multiple welding zones; The data acquisition module is used to acquire the welding data and environmental change data of the welding zone. The welding data includes welding temperature, welding pressure, image, gas flow rate, and welding time; The model analysis module is used to analyze based on the welding data and environmental change data by using a pre-trained welding model to generate a welding parameter adjustment strategy corresponding to the welding zone; The adjustment module is used to correspondingly adjust the welding parameters of MiniLED based on the welding parameter adjustment strategy.

[0075] Optionally, the feature acquisition module includes: The image acquisition sub-module is used to collect the laser spot image and the welding area image of the welding area of MiniLED by using a camera module, and extract the spot feature information from the laser spot image through an image processing algorithm. The spot feature information includes the edge, shape, and size parameters of the spot; The temperature acquisition sub-module is used to collect the temperature data of the welding area of MiniLED by using a thermal imager to generate a temperature distribution map; The image processing sub-module is used to process the welding area image by using the solder joint detection algorithm, extract the solder joint feature information in the welding area image, and perform statistical analysis on the feature information to determine the solder joint density and solder joint distribution. The solder joint characteristic information includes contour and position data; The fusion sub-module is used to perform fusion processing on the spot feature information, temperature distribution map, solder joint density and solder joint distribution to generate welding features.

[0076] The preliminary division sub-module is used to preliminarily divide the welding area by using the image segmentation algorithm based on the spot feature information in the welding features to obtain multiple welding sub-areas; The area adjustment sub-module is used to adjust the welding sub-areas by using the k-means clustering algorithm based on the temperature distribution map, solder joint density and solder joint distribution in the welding features to generate welding partitions.

[0077] Optionally, the data acquisition module includes: The temperature determination sub-module is used to determine the real-time temperature data of the welding partition from the temperature distribution map; The sensing detection sub-module is used to collect the pressure data and shielding gas flow data of the welding partition by using a sensor group, and the sensor group includes a pressure sensor and a gas flow sensor; The recognition and analysis sub-module is used to perform recognition and analysis on the welding area image to obtain the solder joint morphology features of the welding partition; The time recording sub-module is used to record the welding operation time of each stage of the welding partition to obtain the welding time of the welding partition; The environment monitoring sub-module is used to obtain the environmental change data of the environment where the MiniLED is located through an environmental sensor, and the environmental change data includes environmental temperature change data and environmental humidity change data.

[0078] Optionally, the welding device for MiniLED further includes: The historical data collection module is used to collect historical welding data, and the historical welding data includes historical welding temperature, welding pressure, image, gas flow, welding time data and the corresponding welding effect; The data preprocessing module is used to preprocess the historical welding data to generate a model training set, and the preprocessing includes data normalization, denoising processing and feature extraction; The model training module is used to train by using a convolutional neural network based on the model training set, and use the cross-validation method to optimize the parameters during the model training process to obtain a pre-trained welding model.

[0079] Optionally, the model analysis module includes: An incremental learning sub-module, which is used to perform online parameter update and adjustment on a pre-trained welding model by using an incremental learning method based on environmental change data, so as to obtain an updated welding model; An update analysis sub-module, which is used to input welding data into the updated welding model for analysis, and generate a welding parameter adjustment strategy corresponding to the welding partition.

[0080] Optionally, the welding device for MiniLED further includes: A comparison and analysis module, which is used to compare and analyze welding data and environmental change data with a preset safety threshold to generate a comparison and analysis result; An early warning module, which is used to generate a fault warning signal if the comparison and analysis result exceeds the preset safety threshold, and extract abnormal key parameters from the analysis result; A report generation module, which is used to generate a fault report based on the abnormal key parameters and transmit the fault report to the user side.

[0081] For the specific limitations of a welding device for MiniLED, reference can be made to the limitations of a welding method for MiniLED in the above text, which will not be elaborated here. Each module in the above welding device for MiniLED can be implemented in whole or in part by software, hardware and their combination. The above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0082] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 9 shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store historical welding data, and the historical welding data includes historical welding temperature, welding pressure, image, gas flow rate, welding time data and the corresponding welding effect. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it realizes a welding method for MiniLED.

[0083] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: Obtain the welding characteristics of the MiniLED, and based on the welding characteristics, divide the welding area of the MiniLED to generate multiple welding partitions; Obtain the welding data and environmental change data of the welding partitions. The welding data includes welding temperature, welding pressure, image, gas flow rate, and welding time; Based on the welding data and environmental change data, use a pre-trained welding model for analysis to generate a welding parameter adjustment strategy corresponding to the welding partitions; Based on the welding parameter adjustment strategy, correspondingly adjust the welding parameters of the MiniLED.

[0084] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: Obtain the welding characteristics of the MiniLED, and based on the welding characteristics, divide the welding area of the MiniLED to generate multiple welding partitions; Obtain the welding data and environmental change data of the welding partitions. The welding data includes welding temperature, welding pressure, image, gas flow rate, and welding time; Based on the welding data and environmental change data, use a pre-trained welding model for analysis to generate a welding parameter adjustment strategy corresponding to the welding partitions; Based on the welding parameter adjustment strategy, correspondingly adjust the welding parameters of the MiniLED.

[0085] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0086] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above.

[0087] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A soldering method for MiniLED, characterized in that, The described welding method for MiniLEDs includes: Obtain the welding characteristics of the MiniLEDs. Based on the welding characteristics, divide the welding area of the MiniLEDs to generate multiple welding partitions; Obtain the welding data and environmental change data of the welding partitions. The welding data includes welding temperature, welding pressure, image, gas flow rate, and welding time; Based on the welding data and environmental change data, use a pre-trained welding model for analysis to generate a welding parameter adjustment strategy corresponding to the welding partitions; Based on the welding parameter adjustment strategy, correspondingly adjust the welding parameters of the MiniLEDs.

2. The soldering method for MiniLED according to claim 1, wherein The obtaining of the welding characteristics of the MiniLEDs includes: Use a camera module to collect the laser spot image and the welding area image of the welding area of the MiniLEDs, and extract the spot feature information from the laser spot image through an image processing algorithm. The spot feature information includes the edge, shape, and size parameters of the spot; Collect the temperature data of the welding area of the MiniLEDs through an infrared thermal imager to generate a temperature distribution map; Use a solder joint detection algorithm to process the welding area image, extract the solder joint feature information in the welding area image, and perform statistical analysis on the feature information to determine the solder joint density and solder joint distribution. The solder joint characteristic information includes contour and position data; Perform fusion processing on the spot feature information, temperature distribution map, solder joint density, and solder joint distribution to generate the welding characteristics.

3. The soldering method for MiniLED according to claim 2, wherein, The dividing of the welding area of the MiniLEDs based on the welding characteristics to generate multiple welding partitions includes: Based on the spot feature information in the welding characteristics, use an image segmentation algorithm to preliminarily divide the welding area to obtain multiple welding sub-areas; Based on the temperature distribution map, solder joint density, and solder joint distribution in the welding characteristics, use the k-means clustering algorithm to adjust the welding sub-areas to generate the welding partitions.

4. A soldering method for MiniLED according to claim 3, wherein, The obtaining of the welding data and environmental change data of the welding partitions includes: Determine the real-time temperature data of the welding partitions from the temperature distribution map; Use a sensor group to collect the pressure data and protective gas flow rate data of the welding partitions. The sensor group includes a pressure sensor and a gas flow sensor; Perform recognition and analysis on the welding area image to obtain the solder joint morphology characteristics of the welding partitions; Record the welding operation time of each stage of the welding partitions to obtain the welding time of the welding partitions; Obtain the environmental change data of the environment where the MiniLEDs are located through an environmental sensor. The environmental change data includes environmental temperature change data and environmental humidity change data.

5. A soldering method for MiniLED according to claim 1, characterized in that, Before the analysis using the pre-trained welding model based on the welding data and environmental change data, the welding method for MiniLEDs further includes: Collect historical welding data. The historical welding data includes historical welding temperature, welding pressure, image, gas flow rate, welding time data, and the corresponding welding effect; Preprocess the historical welding data to generate a model training set, where the preprocessing includes data normalization, denoising, and feature extraction; Based on the model training set, use a convolutional neural network for training, and use the cross-validation method to optimize parameters during the model training process to obtain the pre-trained welding model.

6. The MiniLED welding method according to claim 1, wherein: The method for analyzing based on the welding data and environmental change data and using the pre-trained welding model to generate the welding parameter adjustment strategy corresponding to the welding zone includes: Based on the environmental change data, use the incremental learning method to perform online parameter update and adjustment on the pre-trained welding model to obtain an updated welding model; Input the welding data into the updated welding model for analysis to generate the welding parameter adjustment strategy corresponding to the welding zone.

7. The MiniLED welding method according to claim 1, wherein: The welding method for MiniLED further includes: Compare and analyze the welding data and environmental change data with a preset safety threshold to generate a comparison and analysis result; If the comparison and analysis result exceeds the preset safety threshold, generate a fault warning signal, and extract the abnormal key parameters from the analysis result; Based on the abnormal key parameters, generate a fault report and transmit the fault report to the user terminal.

8. A soldering device for MiniLED, characterized in that, The welding device for MiniLED includes: A feature acquisition module for acquiring the welding features of MiniLED, and based on the welding features, dividing the welding area of MiniLED to generate multiple welding zones; A data acquisition module for acquiring the welding data and environmental change data of the welding zone, where the welding data includes welding temperature, welding pressure, image, gas flow rate, and welding time; A model analysis module for analyzing based on the welding data and environmental change data and using the pre-trained welding model to generate the welding parameter adjustment strategy corresponding to the welding zone; An adjustment module for correspondingly adjusting the welding parameters of MiniLED based on the welding parameter adjustment strategy.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the welding method for MiniLED according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the welding method for MiniLED according to any one of claims 1 to 7.

Images