PCBA production line intelligent monitoring and closed-loop control system based on Internet of Things sensing and data driving

Abstract

The invention discloses a PCBA production line intelligent monitoring and closed-loop control system based on Internet of Things perception and data driving, and the system comprises a control terminal which is configured to execute the following operations: receiving an image comprising a printed circuit board assembly; determining whether the image comprises an edge identification mark; if it is determined that the image does not include the edge identification mark, the image obtaining device is commanded to adjust the position, and the image including the printed circuit board assembly is shot again; if it is determined that the image comprises the edge identification identifier, determining whether the image has a correct scaling based on the edge identification identifier; if it is determined that the image has the correct scaling, the image continues to be analyzed. According to the invention, the stability and accuracy of image detection can be improved, the partition differentiation parameter adjustment is realized, the defect tracing and process optimization capabilities are enhanced, and the production line yield and the production efficiency are improved.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) technology, and in particular to an intelligent monitoring and closed-loop control system for PCBA production lines based on IoT sensing and data-driven approaches. Background Technology

[0002] As the electronics manufacturing industry upgrades towards intelligent manufacturing, printed circuit board assembly (PCBA) production lines are gradually incorporating IoT sensing, industrial Ethernet, and edge computing to achieve online collection and interconnection of equipment status, process parameters, environmental, and quality data. Digital management based on systems such as Manufacturing Execution Systems (MES) and Supervisory Control and Data Acquisition (SCADA) is continuously improving. Data-driven methods are increasingly being applied in anomaly detection, process optimization, and quality traceability. Combining machine learning and digital twins enables more precise prediction and scheduling. Simultaneously, production line control is evolving from single-point control to cross-process collaboration and closed-loop regulation, promoting improvements in production efficiency, yield, and traceability.

[0003] Many anomaly detection methods rely on image data, but current image data preprocessing techniques still need improvement. Summary of the Invention

[0004] To achieve the above objectives, this invention provides an intelligent monitoring and closed-loop control system for printed circuit board assembly (PCBA) production lines based on IoT sensing and data-driven principles. The system includes a control terminal that acquires PCBA images and sequentially performs edge detection, scaling verification, sharpness verification, and orientation correction. This enables adaptive adjustment of the image acquisition device's position and focal length, ensuring consistency in the acquired images' scale, sharpness, and orientation. Furthermore, based on positional identification markers, positioning calibration is performed, and a template is overlaid on the correctly positioned PCBA to divide it into multiple regions. Defects in each region are identified, and different process adjustment strategies are applied to each region to form a closed-loop control. This improves the stability and accuracy of image detection, enables differentiated parameter tuning by region, enhances defect traceability and process optimization capabilities, and improves production line yield and efficiency.

[0005] This invention provides an intelligent monitoring and closed-loop control system for PCBA production lines based on IoT sensing and data-driven principles. The system includes a control terminal configured to perform the following operations: Receive images including printed circuit board assemblies; Determine whether the image contains edge detection markers; If it is determined that the image does not contain edge recognition markers, the image acquisition device is instructed to adjust its position and re-capture an image including the printed circuit board assembly; If the image is determined to contain edge detection markers, then the image is determined to have the correct scaling based on the edge detection markers. If the image is determined to be at the correct scaling, continue analyzing the image.

[0006] In a preferred embodiment, further image analysis includes: Identify sharpness identification marks in an image, wherein the sharpness identification marks are set on a printed circuit board assembly; Determining whether an image has the correct sharpness is based on sharpness identification markers; If it is determined that the image does not have the correct sharpness, the image acquisition device is instructed to adjust the focus and retake the image including the printed circuit board assembly; If the image is determined to have the correct sharpness, then proceed to identify the edges of the printed circuit board assembly in the image.

[0007] In a preferred embodiment, the control terminal is further configured to perform the following operations: Determine whether an image has the correct orientation based on the edges of the image and the edges of the printed circuit board assembly; If it is determined that the image is not correctly oriented, the image is oriented correctly and the location identification markers in the image are identified. If the image is determined to have the correct orientation, then location identification markers in the image are identified.

[0008] In a preferred embodiment, the control terminal is further configured to perform the following operations: Location identification markers are used to determine whether printed circuit board assemblies in images have the correct positions; If it is determined that the printed circuit board assembly in the image is not in the correct position, then adjust the position of the printed circuit board assembly in the image to the correct position; A template is overlaid on the printed circuit board assembly in the image with the correct position to divide the printed circuit board assembly in the image into multiple zones.

[0009] In a preferred embodiment, the plurality of zones includes a first zone and a second zone. The control terminal is also configured to perform the following operations: Identify the first defect in the first zone; Based on the first defect, the process for manufacturing printed circuit board assemblies is adjusted using the first method; Identify the second defect in the second zone; The process for manufacturing printed circuit board assemblies is adjusted using a second method based on the second defect, wherein the first method differs from the second method.

[0010] This invention provides a method for intelligent monitoring and closed-loop control of PCBA production lines based on IoT sensing and data-driven approaches. The method is executed by a control terminal and includes: Receive images including printed circuit board assemblies; Determine whether the image contains edge detection markers; If it is determined that the image does not contain edge recognition markers, the image acquisition device is instructed to adjust its position and re-capture an image including the printed circuit board assembly; If the image is determined to contain edge detection markers, then the image is determined to have the correct scaling based on the edge detection markers. If the image is determined to be at the correct scaling, continue analyzing the image.

[0011] In a preferred embodiment, further image analysis includes: Identify sharpness identification marks in an image, wherein the sharpness identification marks are set on a printed circuit board assembly; Determining whether an image has the correct sharpness is based on sharpness identification markers; If it is determined that the image does not have the correct sharpness, the image acquisition device is instructed to adjust the focus and retake the image including the printed circuit board assembly; If the image is determined to have the correct sharpness, then proceed to identify the edges of the printed circuit board assembly in the image.

[0012] In a preferred embodiment, the method further includes: Determine whether an image has the correct orientation based on the edges of the image and the edges of the printed circuit board assembly; If it is determined that the image is not correctly oriented, the image is oriented correctly and the location identification markers in the image are identified. If the image is determined to have the correct orientation, then location identification markers in the image are identified.

[0013] In a preferred embodiment, the method further includes: Location identification markers are used to determine whether printed circuit board assemblies in images have the correct positions; If it is determined that the printed circuit board assembly in the image is not in the correct position, then adjust the position of the printed circuit board assembly in the image to the correct position; A template is overlaid on the printed circuit board assembly in the image with the correct position to divide the printed circuit board assembly in the image into multiple zones.

[0014] In a preferred embodiment, the plurality of zones includes a first zone and a second zone. The methods also include: Identify the first defect in the first zone; Based on the first defect, the process for manufacturing printed circuit board assemblies is adjusted using the first method; Identify the second defect in the second zone; The process for manufacturing printed circuit board assemblies is adjusted using a second method based on the second defect, wherein the first method differs from the second method.

[0015] Compared with existing technologies, this invention has the following advantages: It proposes an intelligent monitoring and closed-loop control method for PCBA production lines based on IoT sensing and data-driven approaches. This method is executed by a control terminal and works in conjunction with an image acquisition device. It establishes an automatic judgment and closed-loop adjustment mechanism for imaging effectiveness from the source, overcoming problems such as strong dependence on imaging prerequisites, insufficient robustness of two-point calibration, difficulty in suppressing scale and distortion errors, and frequent rework due to a lack of imaging quality control in existing technologies. Specifically, this invention first receives an image including printed circuit board assemblies and automatically determines whether the image contains edge recognition markers. When no edge recognition markers are detected, the control terminal directly commands the image acquisition device to adjust its position and re-capture the image, thus avoiding errors introduced by continuing subsequent calculations when the circuit board is not in the effective field of view or edge information is missing. After detecting edge recognition markers, this invention further verifies the image scaling ratio based on the edge recognition markers. When the scaling ratio does not meet the preset requirements, it does not proceed to subsequent analysis but instead re-acquires an image that meets the correct scaling ratio, reducing the unreliability of coordinate transformation caused by excessive scale differences. Subsequently, when continuing image analysis, this invention identifies sharpness identification markers set on the printed circuit board assembly and determines whether the image has correct sharpness. When the sharpness is insufficient, the control terminal commands the image acquisition device to adjust the focus and retake the image, completing the focus loop before entering edge recognition and position calculation, thereby avoiding unstable edge extraction and the spread of matching errors caused by out-of-focus or blurry images. Furthermore, this invention determines whether the image orientation is correct based on the image edges and the edges of the printed circuit board assembly. If the orientation is incorrect, the image is first adjusted to the correct orientation before identifying the position identification marker, ensuring that subsequent positioning is performed under a unified directional reference. Based on the position identification markers, it determines whether the printed circuit board assembly is in the correct position; if not, its position is adjusted to the correct position, achieving standardized correction of orientation and position, reducing subsequent analysis failures caused by out-of-range poses or positional offsets. Finally, the present invention divides the printed circuit board assembly into multiple areas by overlaying a template on an image that meets the correct position, and identifies defects in each area. For example, a first defect is identified in the first area and the processing technology is adjusted using a first method, and a second defect is identified in the second area and the processing technology is adjusted using a second method. The first method and the second method are different, thereby forming a differentiated closed-loop process adjustment mechanism for different areas and different defect types. This avoids the lack of adaptability caused by relying on a single correction or single processing strategy, and achieves more stable intelligent monitoring and closed-loop control of the production line. Attached Figure Description

[0016] Figure 1 This is a system logic block diagram of one embodiment of the present invention.

[0017] Figure 2This is a flowchart of a method executed by a control terminal according to an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of an embodiment of the present invention, including a printed circuit board assembly.

[0019] Figure 4 This is a schematic diagram of an image excluding edge recognition markers, according to an embodiment of the present invention.

[0020] Figure 5 This is a schematic diagram of a poorly sharp image according to an embodiment of the present invention.

[0021] Figure 6 This is a schematic diagram of an image of a printed circuit board assembly with misalignment, according to an embodiment of the present invention.

[0022] Figure 7 This is a schematic diagram of an image showing a location identification marker, according to an embodiment of the present invention.

[0023] Figure 8 This is a schematic diagram of an embodiment of the present invention after a template has been added to an image. Detailed Implementation

[0024] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings, but it should be understood that the scope of protection of the present invention is not limited to the specific embodiments.

[0025] The prior art CN114612417B proposes an image calibration method for PCB boards, which combines its Figure 3-5The technical solution described is as follows: Two reference points are selected in the circuit board image captured by the platform, and the corresponding reference points are matched in the AOI scan image to calculate the translation and rotation relationship of the circuit board. When the scales of the two images are inconsistent, the scaling factor is calculated by the ratio of the distance between the two reference points, and coordinate transformation is performed accordingly to adjust the circuit board in the image. Our research found that this type of adjustment method still has obvious defects in engineering applications: First, this method is highly dependent on the imaging prerequisites, such as requiring that the circuit board is not upside down, the angle deviation must be limited to a certain range, and the imaging parameters must be pre-adjusted to make the size of the circuit board similar to the AOI image. Once the angle range is exceeded, the circuit board is flipped / inverted, or the scale difference is too large, the two-point calibration link is prone to failure or introduces systematic errors. Second, the robustness of two-point calibration to the selection and matching of reference points is insufficient. If the reference point selection is inaccurate or the matching is incorrect (e.g., affected by device occlusion, reflection, noise, or similar local textures), the translation, rotation, and scaling parameters may be locked by overall errors, resulting in global drift, rotation, or... Third, the scaling factor is only approximated by the ratio of the distance between two points, which implies an ideal imaging assumption. It is difficult to effectively resist factors such as perspective distortion, lens distortion, platform height changes or board warping that may exist in real production lines, resulting in inconsistent correction effects in different areas. Fourth, the method lacks closed-loop constraints on image quality. It does not effectively judge and control the image sharpness, edge integrity and feature recognizability before calibration. When the image has out-of-focus, reflection or motion blur, it may still enter the calculation and verification process, resulting in repeated rework or reliance on manual confirmation. In addition, if an iterative "projection-shooting-correction-repetition" correction strategy is adopted, it may further bring about convergence uncertainty and cycle fluctuation problems, making it difficult to meet the requirements of stable and efficient production line applications.

[0026] Example 1: Figure 1This is a system logic block diagram of one embodiment of the present invention. As shown in the figure, the intelligent monitoring and closed-loop control system 100 for PCBA production lines based on IoT sensing and data driving of the present invention includes an image acquisition subsystem 105, a control terminal 110, and a process execution and feedback subsystem 115 connected to the production line equipment. The image acquisition subsystem 105 includes an image acquisition device 1051, a position adjustment mechanism 1052, and a focus adjustment mechanism 1053, used to capture images of the PCBA and output images containing the PCBA. The control terminal 110 is communicatively connected to the image acquisition device 1051 (the image acquisition device may be a camera), used to receive images and execute a multi-level verification and correction process based on edge recognition markers, sharpness recognition markers, and position recognition markers. When the scaling ratio, sharpness, orientation, or position of the image is detected to be inconsistent with the preset requirements, a position adjustment command or focus adjustment command is sent to the image acquisition device 1051 to complete the re-capture. After the image meets the acquisition conditions, the control terminal 110 overlays a template on the PCBA image based on the positioning results to divide it into multiple regions, performs regional defect identification, and generates corresponding process adjustment commands according to the defect type or defect characteristics of different regions, which are sent to the production line process execution equipment 120 to achieve closed-loop adjustment of processing parameters. At the same time, the process execution and feedback subsystem 115 acquires the adjustment results and production line status information of the process execution equipment 120, and transmits the adjusted process parameters and status information back to the control terminal 110 for recording, tracing and subsequent data-driven optimization, thereby constructing a closed-loop control link of "perception-analysis-decision-execution-feedback".

[0027] Example 2: Figure 2 This is a flowchart illustrating a method executed by a control terminal according to an embodiment of the present invention. As shown, the control terminal is configured to perform the following steps: Step 205: Receive an image including the printed circuit board assembly; an example of an image including the printed circuit board assembly can be found here. Figure 3 Those skilled in the art should understand that, in order to protect the commercial interests of our clients, Figure 3 The circuit board diagram in the image is not a real circuit board, but rather a diagram obtained by modifying a real circuit board to a certain extent. Figure 3 The image edge 305 is illustrated in the diagram. It can be understood that in the actual image, the gap between the image edge 305 (the image edge can also be referred to as the image boundary) and the printed circuit board assembly should be the worktable that supports the printed circuit board assembly. Step 210: Determine whether the image contains edge detection markers; further as... Figure 3 As shown, Figure 3The image shows an edge recognition mark 310, which is set on the worktable carrying the printed circuit board assembly. Therefore, when the camera photographs the printed circuit board assembly, it also captures the edge recognition mark 310 on the worktable in the image. Step 215: If it is determined that the image does not contain edge recognition markers, the image acquisition device is instructed to adjust its position and re-capture an image including the printed circuit board assembly. In one example, the camera may move up and down due to an externally applied force. It should also be noted that the camera is shooting the printed circuit board assembly from above. Therefore, if the camera moves upward, the camera is farther away from the printed circuit board assembly, and the proportion of the printed circuit board assembly in the image decreases. Conversely, if the camera moves downward, the proportion of the printed circuit board assembly in the image increases. Figure 4 This diagram illustrates how the proportion of printed circuit board assemblies in the image increases as the camera moves downwards, such as... Figure 4 As shown, edge 405 of the image substantially coincides with the edge of the printed circuit board assembly, therefore in Figure 4 Edge detection markers could not be identified in the image; it is understandable that... Figure 4 Based on the image, the control terminal will send instructions to the image acquisition subsystem, which will then move the camera upwards according to the instructions so that the camera moves away from the printed circuit board assembly. In the subsequently captured image, edge recognition marks may be identified. Step 220: If it is determined that the image contains edge detection markers, then determine whether the image has the correct scaling ratio based on the edge detection markers; still using Figure 3 For example, it can be pre-set that if the edge detection marker is 20 pixels away from the left edge of the image, the image has the correct scaling (that is, it indicates that the distance between the camera and the printed circuit board assembly is correct). Understandably, if the edge detection marker is more than 20 pixels away from the left edge of the image, it indicates that the camera is too far away from the printed circuit board assembly, and the camera should be instructed to move toward the printed circuit board assembly. Conversely, if the edge detection marker is less than 20 pixels away from the left edge of the image, it indicates that the camera is too close to the printed circuit board assembly, and the camera should be instructed to move away from the printed circuit board assembly. Step 225: If the image is determined to have the correct scaling, continue analyzing the image.

[0028] Example 3: In Example 3, the image analysis continues, including: The image identifies sharpness identification markers, wherein the sharpness identification markers are disposed on a printed circuit board assembly; in one example, an example of a sharpness identification marker can be found here. Figure 3 , Figure 3 The image shows the clarity identification mark 315. Figure 5 This is a schematic diagram of a poorly sharp image according to an embodiment of the present invention. It is understood that the reduced image sharpness may be due to a camera focus error. (Comparison...) Figure 5 and Figure 3 It can be seen that, Figure 5 The image is obviously compared to Figure 3 The image clarity is lower (also known as a "blurry" image); in Figure 5 In the image, the clarity identification mark 505, which is blurred, is clearly more accurate than... Figure 3 The normal sharpness identification mark 315 in the image is wider, which can be used to determine whether the image sharpness is correct. To solve the sharpness problem, an intuitive idea might be to directly rely on the edge position or width of the parts on the printed circuit board assembly to judge the image sharpness. However, in our research, we found that the edges and dimensions of corresponding parts on different circuit boards are not completely consistent (that is, the dimensions of each part have certain allowable tolerances). Therefore, under the premise that the part dimensions themselves are not fixed, it is basically not feasible to judge the image sharpness by relying on the edge position or width of the parts on the printed circuit board assembly. Therefore, we proposed to use a sharpness identification mark to judge the sharpness. Determining whether an image has the correct sharpness is based on sharpness identification markers; If it is determined that the image does not have the correct sharpness, the image acquisition device is instructed to adjust the focus and retake the image including the printed circuit board assembly; If the image is determined to have the correct sharpness, then proceed to identify the edges of the printed circuit board assembly in the image.

[0029] Example 4: In Example 4, the control terminal is further configured to perform the following steps: Determining whether an image has the correct orientation is based on the edges of the image and the edges of the printed circuit board assembly; such as... Figure 6 As shown, the edge 605 of the printed circuit board assembly is clearly not parallel to the edge of the image. In this case, the printed circuit board assembly should be considered to have incorrect orientation. It should be noted that in the system of this invention, the camera itself does not have degrees of freedom for forward, backward, left, or right movement, thus causing… Figure 6 The reason for the incorrect orientation of the printed circuit board assembly shown is that the printed circuit board assembly was not correctly moved to the designated shooting position; If it is determined that the image is not correctly oriented, the image is oriented correctly and the location identification markers in the image are identified; in one example, if such as Figure 6In the case shown, the printed circuit board assembly in the image can be adjusted by rotating its orientation without re-capturing the image. If the image is determined to have the correct orientation, then location identification markers in the image are identified.

[0030] Example 5: In Example 5, the control terminal is further configured to perform the following steps: Location identification markers are used to determine whether printed circuit board assemblies in an image are in the correct position; an example of a location identification marker can be found here. Figure 7 ,like Figure 7 As shown, the position identification mark can be the fixing hole at the upper left corner of the printed circuit board assembly. Since the size and position of such fixing holes on each circuit board are stable, they can be used as position identification marks. In one example, since the errors that may occur in the scaling ratio, sharpness, tilt, etc. of the printed circuit board assembly in the image have been eliminated through the aforementioned steps, in embodiment 5, it can be determined whether the printed circuit board assembly in the image has the correct position based on the actual coordinates and the predetermined coordinates of the position identification mark. For example, if the actual coordinates and the predetermined coordinates of the position identification mark are consistent, it is determined that the printed circuit board assembly in the image has the correct position. Figure 7 In the example, it is clear that the actual coordinates of the location identification marker and the predetermined coordinates of the location identification marker are inconsistent. In this case, the printed circuit board assembly in the image can be translated without re-capturing the image, thereby solving the problem of the incorrect position of the printed circuit board assembly. It should be noted that in the system of the present invention, the camera itself does not have the freedom of forward, backward, left, and right movement, thus causing... Figure 7 The reason for the incorrect position of the printed circuit board assembly shown is that the printed circuit board assembly was not correctly moved to the designated shooting position; If it is determined that the printed circuit board assembly in the image is not in the correct position, then adjust the position of the printed circuit board assembly in the image to the correct position; A template is overlaid on the printed circuit board assembly in the image with the correct position to divide the printed circuit board assembly in the image into multiple zones. Figure 8 This is a schematic diagram illustrating the addition of a template to an image according to an embodiment of the present invention. Figure 8As shown, after overlaying a template onto the printed circuit board assembly in the image, three regions are observed on the assembly, each outlined by a red box. The template is determined based on the fact that defects appearing within the boxes are usually of the same type, and the process parameters causing this same defect are typically a fixed set of parameters. Based on this region division, the system can perform defect identification, feature extraction, and statistical modeling for each region, limiting the mapping of defect samples and their associated process parameters to the region level. This reduces the heterogeneity and noise interference of machine learning training data, and improves the convergence speed and generalization ability of the model in defect type discrimination and causal parameter correlation analysis.

[0031] Multiple zones, including Zone 1 and Zone 2, wherein the control terminal is further configured to perform the following steps: Identify the first defect in the first area; image-based defect identification is a well-known prior art, for example, prior art CN116258682A discloses a method for identifying defects on PCBA based on images; The process for manufacturing printed circuit board assemblies is adjusted based on a first defect using a first method. Adjusting the process based on defects can be done using well-known large-scale model training methods. For example, taking solder paste printing quality control as an example, the training process is roughly as follows: The system acquires multiple batches of production data during the solder paste printing process and establishes a mapping relationship between defect types and process parameter adjustment strategies. The production data includes at least: images of the pad areas after solder paste printing, solder paste printing defect types obtained from image recognition, the corresponding batch's solder paste printing process parameter settings, and subsequent detection results. The subsequent detection results are used to characterize whether the process parameter adjustments are effective. Solder paste printing defect types include at least insufficient solder, excessive solder, solder paste misalignment, solder spikes, and collapse defects.

[0032] To construct the aforementioned mapping relationship, the system labels the pad area images with defect types. Labeling methods include manual labeling, automatic labeling by an existing defect identification system, and manual correction of automatic labeling results, ensuring that each pad area image corresponds to a clear defect type label. Simultaneously, the system records the process parameter adjustment results associated with the defect samples. These process parameter adjustment results include at least: the type of process parameter adjusted, the direction and magnitude of the adjustment, and the subsequent defect improvement on the circuit board. When the corresponding defect is significantly reduced or disappears after adjustment, the process parameter adjustment operation is marked as effective for that defect type, thus forming a "defect-adjustment-effect" associated sample set that can be used for training.

[0033] Based on the sample set, the system trains a mapping model between defect types and process parameter adjustment strategies. The defect type serves as the model input, the process parameter adjustment strategy as the model output, and the defect improvement effect as feedback for model training and optimization. The mapping model can be implemented using rule-based models, statistical models, or machine learning models, and its specific structure is not limited, as long as it can output the corresponding process parameter adjustment strategy based on the input defect type. The process parameter adjustment strategy should at least include the type of process parameter to be adjusted, the direction of adjustment, and the range of adjustment. For example, when the defect type is a solder shortage defect, the system can output an adjustment strategy to adjust the scraper pressure or scraper speed; when the defect type is a solder excess defect, the system can output an adjustment strategy to adjust the scraper angle or solder paste release gap.

[0034] During the production operation phase, the system performs real-time analysis of the solder pad area image after solder paste printing to determine the solder paste printing defect type of the current circuit board. The defect type is then input into a trained mapping model to obtain the corresponding process parameter adjustment strategy. Subsequently, the system issues process parameter adjustment instructions to the solder paste printing equipment according to the adjustment strategy, causing subsequent circuit boards to use the adjusted process parameters for solder paste printing. To form a closed-loop feedback, after parameter adjustment, the system again acquires images of the solder pad area of ​​subsequent circuit boards printed using the adjusted process parameters and determines whether the corresponding defect type has been improved. If the defect is determined to be improved, the defect type and process parameter adjustment strategy are recorded as effective strategies; if the defect is determined not to be improved, it is recorded as an invalid strategy or its weight is reduced. Based on the adjustment results, the mapping model is iteratively updated, enabling the model to prioritize historically validated process parameter adjustment strategies in subsequent operations, thereby achieving data-driven closed-loop optimization and automatic parameter tuning of process parameters.

[0035] Identify the second defect in the second zone; Based on the second defect, a second method is used to adjust the process for manufacturing printed circuit board assemblies, where the first method differs from the second method. The control terminal can trigger differentiated process adjustment strategies for defects in different regions: for example, the first method is used to adjust the first set of process parameters associated with the first defect identified in the first region, and the second method is used to adjust the second set of process parameters associated with the second defect identified in the second region. This transforms process parameter tuning from "globally unified adjustment" to targeted closed-loop control at the "region-defect-parameter group" granularity, reducing irrelevant parameter disturbances and the risk of misadjustment, improving defect elimination efficiency and parameter tuning stability, and facilitating the formation of a traceable regional-level defect-parameter knowledge base, further supporting subsequent self-learning optimization and production line yield improvement.

[0036] It should be understood that the above description of the specific embodiments of the method and apparatus of the present invention in conjunction with the accompanying drawings is only for illustrating the technical concept and implementation of the present invention, and does not constitute a limitation on the scope of protection of the present invention. Those skilled in the art, after understanding the technical solution of the present invention, can make various changes, substitutions, or combinations to the technical features in the embodiments without departing from the spirit and substance of the present invention, such as making equivalent substitutions for method steps or making corresponding adjustments to the functional units in the apparatus; all such changes, substitutions, or combinations should be considered to fall within the scope of protection of the present invention.

[0037] Furthermore, the functional modules, units, or device components described in this specification can be implemented as independent hardware, or through software programs, firmware, or a combination of hardware and software. The above division of functional modules is merely for the purpose of understanding the technical solution of this invention and is not intended to limit the specific structure or implementation method. In other embodiments, relevant functional modules can be merged, split, or reconfigured according to the actual needs of the communication system without affecting the implementation of the technical solution of this invention.

[0038] In the embodiments described in this specification, the order of the steps is merely illustrative and is not intended to limit the method flow of the present invention. Unless explicitly stated or logically required to be performed in a specific order, those skilled in the art may adjust the order of the steps or perform some steps in parallel according to specific communication scenarios, system architectures, or implementation conditions, and such modifications should not be considered as a departure from the essence of the present invention.

[0039] Furthermore, in this specification and claims, unless otherwise expressly defined, the terms "comprising," "including," and variations thereof should be interpreted in an open-ended manner, meaning "including but not limited to." Additionally, expressions such as "first," "second," etc., are only used to distinguish different technical features or objects and do not indicate limitations in order, quantity, or importance, nor should they be construed as limiting the technical solution of this invention.

Claims

1. A PCBA production line intelligent monitoring and closed-loop control system based on IoT sensing and data-driven methods, the system including a control terminal configured to perform the following operations: Receive images including printed circuit board assemblies; Determine whether the image contains edge recognition markers; If it is determined that the image does not contain the edge recognition marker, the image acquisition device is instructed to adjust its position and re-capture an image including the printed circuit board assembly; If it is determined that the image includes the edge recognition marker, then it is determined whether the image has the correct scaling ratio based on the edge recognition marker; If it is determined that the image has the correct scaling ratio, then the analysis of the image continues.

2. The system according to claim 1, wherein, Further analysis of the image includes: Identify sharpness identification marks in the image, wherein the sharpness identification marks are disposed on the printed circuit board assembly; The image is determined to have the correct sharpness based on the sharpness identification marker. If it is determined that the image does not have the correct sharpness, the image acquisition device is instructed to adjust the focus and retake the image including the printed circuit board assembly; If the image is determined to have the correct sharpness, the identification of the edges of the printed circuit board assembly in the image continues.

3. The system according to claim 2, wherein, The control terminal is also configured to perform the following operations: The image is determined to have the correct orientation based on the edges of the image and the edges of the printed circuit board assembly. If it is determined that the image does not have the correct orientation, the orientation of the image is adjusted to the correct orientation and the position identification mark in the image is identified; If the image is determined to have the correct orientation, then the location identification marker in the image is identified.

4. The system according to claim 3, wherein, The control terminal is also configured to perform the following operations: The location identification marker is used to determine whether the printed circuit board assembly in the image has the correct position; If it is determined that the printed circuit board assembly in the image is not in the correct position, then the position of the printed circuit board assembly in the image is adjusted to the correct position; A template is overlaid on the printed circuit board assembly in the image with the correct position to divide the printed circuit board assembly in the image into multiple zones.

5. The system according to claim 4, wherein, The multiple zones include Zone 1 and Zone 2. The control terminal is also configured to perform the following operations: Identify the first defect in the first region; Based on the first defect, the process for manufacturing printed circuit board assemblies is adjusted using a first method; Identify the second defect in the second region; The process for manufacturing printed circuit board assemblies is adjusted using a second method based on the second defect, wherein the first method differs from the second method.

6. A method for intelligent monitoring and closed-loop control of PCBA production lines based on IoT sensing and data-driven approaches, wherein the method is executed by a control terminal, and the method includes: Receive images including printed circuit board assemblies; Determine whether the image contains edge recognition markers; If it is determined that the image does not contain the edge recognition marker, the image acquisition device is instructed to adjust its position and re-capture an image including the printed circuit board assembly; If it is determined that the image includes the edge recognition marker, then it is determined whether the image has the correct scaling ratio based on the edge recognition marker; If it is determined that the image has the correct scaling ratio, then the analysis of the image continues.

7. The method according to claim 6, wherein, Further analysis of the image includes: Identify sharpness identification marks in the image, wherein the sharpness identification marks are disposed on the printed circuit board assembly; The image is determined to have the correct sharpness based on the sharpness identification marker. If it is determined that the image does not have the correct sharpness, the image acquisition device is instructed to adjust the focus and retake the image including the printed circuit board assembly; If the image is determined to have the correct sharpness, the identification of the edges of the printed circuit board assembly in the image continues.

8. The method according to claim 7, wherein, The method further includes: The image is determined to have the correct orientation based on the edges of the image and the edges of the printed circuit board assembly. If it is determined that the image does not have the correct orientation, the orientation of the image is adjusted to the correct orientation and the position identification mark in the image is identified; If the image is determined to have the correct orientation, then the location identification marker in the image is identified.

9. The method according to claim 8, wherein, The method further includes: The location identification marker is used to determine whether the printed circuit board assembly in the image has the correct position; If it is determined that the printed circuit board assembly in the image is not in the correct position, then the position of the printed circuit board assembly in the image is adjusted to the correct position; A template is overlaid on the printed circuit board assembly in the image with the correct position to divide the printed circuit board assembly in the image into multiple zones.

10. The method according to claim 9, wherein, The multiple zones include Zone 1 and Zone 2. The method further includes: Identify the first defect in the first region; Based on the first defect, the process for manufacturing printed circuit board assemblies is adjusted using a first method; Identify the second defect in the second region; The process for manufacturing printed circuit board assemblies is adjusted using a second method based on the second defect, wherein the first method differs from the second method.

Images