A PCB board dynamic burning status monitoring method and system

Abstract

The present invention discloses a method and system for monitoring the dynamic burning status of a PCB board, which relates to the field of PCB board burning process monitoring and image processing technology. The application can perform multi-dimensional monitoring and abnormality identification on key operating states during the burning process, and realizes full process coverage from image acquisition, path analysis, contact state evaluation to electrical fluctuation monitoring. The system adopts image algorithms combined with crimping behavior modeling technology, which can not only accurately identify mechanical anomalies such as probe crimping path offset, contact inconsistency, and timing mismatch, but also detect electrical anomalies such as micro-contact resistance changes, thereby enhancing monitoring accuracy and real-time performance. By constructing an abnormal risk time anchor point and accessing blockchain evidence storage, data credibility and subsequent traceability are guaranteed. At the same time, a matching locking and alarm mechanism is provided to achieve automatic isolation and closed-loop management of equipment anomalies, effectively improving the consistency of burning quality, equipment operation stability, and maintenance efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of PCB board burning process monitoring and image processing, and in particular to a PCB board dynamic burning state monitoring method and system. Background Art

[0002] In electronics manufacturing, PCB boards undergo firmware burning (programming) before shipment to ensure accurate functional logic and system configuration. This process, typically performed on automated burning equipment, involves several key steps, including station positioning, board placement, probe crimping, power-on programming, and resetting. The probe assembly must precisely contact the corresponding pads on the PCB to establish a signal path and transmit data. With increasing product precision and diverse board configurations, the burning process places increasing demands on both equipment operation and probe contact accuracy.

[0003] However, existing technologies for status monitoring during the programming process still primarily rely on electrical connectivity determination, continuity testing, or manual visual inspection. They lack dynamic monitoring methods based on image processing, behavioral modeling, and microscopic feature analysis, making it difficult to promptly detect subtle deviations and potential faults in the programming process. For example, slight deviations in the probe crimping path, incomplete compression of some pin arrays, mismatches between contact timing and power-on control, and especially sudden fluctuations in contact status (such as instantaneous resistance changes caused by microjitter), misalignment between crimping time and power-on timing, and asymmetric crimping of individual pin arrays due to wear or contamination are all "short-term effects" or "microscopic anomalies" that are difficult to capture using traditional methods. These issues not only affect programming consistency and yield, but also increase the burden on equipment maintenance and reduce system automation and stability. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method and system for monitoring the dynamic burning status of a PCB board to solve the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for monitoring the dynamic burning status of a PCB board, comprising the following steps:

[0006] Step 1: Identify and number multiple key operating units of the burning station; use the industrial camera to collect the image sequence set of the i-th operating unit within the burning cycle t, identify the micro-angle deviation of the probe's pressing path based on the image processing algorithm, and calculate the positioning deviation coefficient of the i-th operating unit within the burning cycle t and misconnection anomaly index and evaluating, and when there is a risk of abnormal positioning, generating a first positioning risk record and triggering a first alarm signal;

[0007] Step 2: Based on the image sequence of the i-th operation unit within the burning cycle t, identify the probe compression height change in the continuous image frames and calculate the contact consistency index of the i-th operation unit And evaluate, if there is a risk of abnormal contact of the probes in the same group, generate a second risk record and trigger a second alarm signal;

[0008] Step 3: Extract the image frame at the moment of contact from the image sequence set of the i-th operation unit within the burning cycle t, and identify and obtain the compression deformation change rate of the k-th probe in the f-th frame. , Brightness change of the needle tip reflection area , the edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame After determining the contact occurrence time, collect the contact action occurrence time Power-on control time , calculate the contact timing offset index of the i-th operating unit And evaluate, when it is determined that there is a risk of abnormal contact timing, generate a third abnormal timing risk record and trigger a third alarm signal;

[0009] Step 4: After determining the contact occurrence time in step 3, collect the micro-voltage jitter value sequence, combine it with the image sequence set, and calculate the micro-contact impedance anomaly coefficient of the i-th operating unit within the burning cycle t And evaluate, when it is determined that there is a micro-contact abnormality, generate a fourth micro-contact abnormality risk record and trigger a fourth alarm signal.

[0010] Preferably, step one includes:

[0011] S11, classify and identify multiple key operating units in the burning station, register them by number, and generate an operating unit index table;

[0012] The operation unit includes: a burning needle bed assembly, a clamping mechanism and a feeding robot arm or a handling module; each operation unit number is bound to the control system address and task scheduling interface, and recorded in the equipment console database;

[0013] S12. When receiving the burning task instruction, bind the PCB board number to be burned with the assigned operation unit, and generate a board burning operation binding record table, the record fields including the operation unit number, board number, task number and planned station coordinates.

[0014] Preferably, step one further comprises:

[0015] S13, before each burning cycle t starts, the industrial camera set above or on the side of the burning needle bed is used to collect dynamic image frames of the needle bed crimping process, and the image collection frequency is not less than 30 frames / second to form an image sequence set of the i-th operation unit in the burning cycle t. ;

[0016] S14, using the image detection algorithm to identify the probe arrangement and the board pad alignment area in the first image, using the standard crimping path vector As a benchmark, identify the path offset angle in each frame image Offset from position , calculate the positioning offset coefficient of the i-th operation unit in the burning cycle t :

[0017] S15, synchronously calculating the misconnection abnormality index of the i-th operation unit in the programming cycle t based on the relative position image of the probe needle tip and the board pad collected in the programming cycle t of the i-th operation unit :

[0018] S16, preset positioning offset threshold Dth and misconnection offset threshold CYth, if or If any one of the following conditions is met, it is determined that the i-th operation unit has a risk of abnormal probe crimping positioning during the current programming cycle. At this time, the first positioning risk record is recorded, including: operation unit number, programming task number, offset image frame ID, offset angle and displacement value, and the first alarm signal is triggered immediately.

[0019] Preferably, step 2 includes:

[0020] S21, in S12, the image sequence set of the i-th operation unit in the burning cycle t In the middle, the key image frame where the probe is about to be fully pressed and maintain a stable contact state is extracted, which is the middle and rear frame and is recorded as the stable frame interval. ;

[0021] S22, in the stable frame interval Extract the compressed height value set of the same group of probes in the i-th operation unit within the burning cycle t ;in, to represents the compression height identified by the 1st to the mth probes in the same group of probes in the i-th operation unit;

[0022] S23, calculate the contact consistency index of the probes in the same group of the i-th operation unit , used to measure the overall dispersion of compression height;

[0023] S24, preset consistency assessment threshold Cth, if , it is determined that there is a risk of abnormal contact of the same group of probes in the current burning cycle of the i-th operation unit, a second risk record is generated, and a second alarm signal is immediately triggered.

[0024] Preferably, step three includes:

[0025] S31. Before each programming cycle begins, the control system outputs a power-on control signal and records the control output timestamp. And bind it with the operation unit number i and the burning task number, and write it into the task log table. The fields include: task number, operation unit number, control power-on time and power-on signal flag;

[0026] S32, extracting the image frame at the time of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and identifying and obtaining the compression deformation change rate of the k-th probe in the f-th frame , Brightness change of the needle tip reflection area , the edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame ;

[0027] And preset the compression variable mutation threshold Hth, reflection change threshold Lth, edge gradient mutation threshold Gth, and minimum stable contact time threshold Tmin;

[0028] S33. When any two or more of the following four conditions are recognized, it is determined that the contact occurs, including: Condition 1 is a sudden change in the probe compression deformation, indicating a significant jump in the compression degree between the previous and next frames: Condition 2: The change in brightness or area of ​​the needle tip reflective area exceeds the threshold, indicating a change in the optical reflection state: Condition 3: The edge of the contact area between the probe and the PCB changes dramatically, and the gradient field in the contact area in the image changes suddenly: Condition 4: The duration of the crimping state reaches the stability threshold. Starting from frame f, the probe maintains a stable crimping state for more than the set time: ;

[0029] S34: Determine the contact occurrence time based on S33, extract the image frame from the contact occurrence time, and obtain the contact action occurrence time Power-on control time , calculate the contact timing offset index of the i-th operating unit :

[0030] S35, set the timing offset threshold Tth, when , it is determined that the i-th operating unit in the current cycle has a contact timing abnormality risk, marked as the third timing abnormality risk record of the i-th operating unit in cycle t, and the third alarm signal is immediately triggered.

[0031] Preferably, the S32 specifically includes:

[0032] S321, identifying and obtaining the rate of change of the compression deformation of the k-th probe in the f-th frame The specific method is:

[0033] S3201, extract the image frame at the moment of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and extract the probe compression height of the f-th frame and the previous frame and :

[0034] S3202, calculate the probe compression height of the f-th frame and the previous frame and The absolute value of the compression deformation change rate of the kth probe in the fth frame is obtained ;

[0035] S322, identifying and obtaining the brightness change of the tip reflective area of ​​the k-th probe in the f-th frame The specific method is:

[0036] S3221, extracting the image frame at the moment of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and presetting the high reflection grayscale threshold , to identify the high reflective area of ​​the needle tip :

[0037]

[0038] in, Represents the pixel grayscale in frame f;

[0039] S3222, calculate the average brightness of the k-th probe in the f-th frame :

[0040] S3223, calculate the brightness change of the tip reflection area of ​​the k-th probe in the f-th frame :

[0041] S323, identifying and obtaining the edge gradient amplitude of the image of the contact area between the kth probe and the PCB :

[0042] S3231, extracting the image frame at the time of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and identifying the image coordinates of the k-th probe in the f-th frame , construct a rectangular region of interest centered at the coordinates , the size is set to W×H, where W is the width and H is the height, covering the probe contact PCB area;

[0043] S3232, in the rectangular region of interest In the example, Sobel operator is used to calculate all pixel points. , calculate its grayscale gradient amplitude ;

[0044] S3233, in the rectangular region of interest The average grayscale gradient amplitude of all pixels is taken to obtain the average gradient energy of the kth probe contact area in the frame :

[0045] S3234, combined with the average gradient energy of the k-th probe contact area , calculate the edge gradient amplitude of the image of the kth probe contact area with the PCB .

[0046] Preferably, step four includes:

[0047] S41: When S33 determines the conditions of the contact occurrence moment, extract the image frame from the contact occurrence moment and obtain the contact action occurrence time Power-on control time During the process, the contact voltage micro-jitter signal is collected synchronously. From the moment the probe group contacts the PCB board in the i-th operation unit to the burning data transmission stage, the instantaneous voltage waveform is collected at a frequency of ≥10kHz, and the micro-voltage jitter value sequence is recorded: , to Indicates the voltage value from the 1st sampling time to the Bth sampling time;

[0048] S42, calculating the peak-to-peak amplitude normalized value of the micro-voltage jitter sequence of the i-th operation unit within the programming cycle t according to the micro-voltage jitter value sequence

[0049] S43. Calculate the normalized value of the standard deviation of the jitter sequence based on the micro-voltage jitter value sequence. ;

[0050] S44, combining the peak-to-peak amplitude normalized value of the micro-voltage jitter sequence of the i-th operation unit within the programming cycle t and the normalized value of the standard deviation of the jitter sequence , calculate the comprehensive voltage fluctuation index of the i-th operating unit during the programming cycle t :

[0051] S45, identifying and obtaining the pinpoint compression integrity index of the i-th operation unit from the image sequence set of the i-th operation unit within the burning period t and needle array regularity index , and calculate the image penalty factor of the i-th job unit in the burning cycle t ;

[0052] S46: Combine the comprehensive voltage fluctuation index of the i-th operating unit obtained in S44 and S45 within the programming cycle t and the image penalty factor of the i-th job unit in the burning cycle t , calculate the micro-contact impedance anomaly coefficient of the i-th operating unit during the programming cycle t ;

[0053] S47, preset impedance abnormality threshold Zth, if , it is determined that the i-th operating unit has a micro-contact abnormality risk in the current burning cycle, and the fourth alarm signal is immediately triggered.

[0054] Preferably, S45 specifically includes:

[0055] S451, from the image sequence set of the i-th operation unit in the burning period t, identify and obtain the pinpoint compression integrity index of the i-th operation unit The specific steps are:

[0056] Assume that the compression depth of the mth probe identified in the crimping state frame is , the compression depth target is , and the acceptance deviation is preset to ,satisfy The crimping is judged to be up to standard, and the needle tip compression integrity index of the i-th operation unit is calculated. ;

[0057] S452: Identify and obtain the needle arrangement regularity index of the i-th operating unit from the image sequence set of the i-th operating unit within the burning cycle t. The specific steps are:

[0058] The nominal center pixel coordinate of each probe head is , perform linear fitting on the set of points arranged horizontally or vertically on the probe, and obtain the needle arrangement neatness index of the i-th operation unit .

[0059] Preferably, the method further includes step 5, collecting the first positioning risk record, the second risk record, the third timing abnormality risk record, and the fourth micro-contact abnormality risk record generated in steps 1 to 4; and binding them to the corresponding burning task number to form the abnormality risk time anchor point of the i-th operation unit in the burning cycle;

[0060] Count the total number of the first positioning risk record, the second risk record, and the third time series abnormal risk record to form the first abnormal risk anchor point number; count the total number of abnormal risk time anchor points of the fourth micro-contact abnormal risk record to form the second abnormal risk anchor point number;

[0061] And submit it through the blockchain node interface for chain operation. After the chain is successfully uploaded, if the number of first abnormal risk anchor points of the i-th work unit is greater than 3 times or the number of second abnormal risk anchor points is greater than 5 times, the work station lock, inspection mark or process review will be automatically triggered.

[0062] A PCB board dynamic burning status monitoring system, comprising:

[0063] The operation unit identification module is used to classify and identify multiple key operation units within the burning station, register them by number, and generate an operation unit index table; the operation unit includes: a burning needle bed assembly, a clamping mechanism, and a feeding robot arm or a handling module; each operation unit number is bound to the control system address and task scheduling interface, and recorded in the equipment console database;

[0064] The image acquisition module is used to capture dynamic image frames of the needle bed crimping process before each burning cycle t by using an industrial camera set above or on the side of the burning needle bed. The image acquisition frequency is not less than 30 frames per second, forming an image sequence set of the i-th operation unit within the burning cycle t;

[0065] Positioning or misconnection status analysis module, used to identify the micro-angle deviation of the probe pressing path based on image processing algorithm, and calculate the positioning deviation coefficient of the i-th operation unit within the burning cycle t and misconnection anomaly index and evaluating, and when there is a risk of abnormal positioning, generating a first positioning risk record and triggering a first alarm signal;

[0066] The contact consistency analysis module is based on the image sequence of the i-th operation unit within the burning cycle t, identifies the change in the probe compression height in the continuous image frames, and calculates the contact consistency index of the i-th operation unit And evaluate, if there is a risk of abnormal contact of the probes in the same group, generate a second risk record and trigger a second alarm signal;

[0067] The contact timing analysis module is used to extract the image frame at the time of contact from the image sequence set of the i-th operating unit within the burning cycle t, and identify and obtain the compression deformation change rate of the k-th probe in the f-th frame. , Brightness change of the needle tip reflection area , the edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame After determining the contact occurrence time, collect the contact action occurrence time Power-on control time , calculate the contact timing offset index of the i-th operating unit And evaluate, when it is determined that there is a risk of abnormal contact timing, generate a third abnormal timing risk record and trigger a third alarm signal;

[0068] The micro-contact impedance analysis module is used to collect the micro-voltage jitter value sequence after determining the contact occurrence moment in the contact timing analysis module, and calculate the micro-contact impedance anomaly coefficient of the i-th operating unit within the burning cycle t in combination with the image sequence set. and evaluating, and when it is determined that there is a micro-contact abnormality, generating a fourth micro-contact abnormality risk record and triggering a fourth alarm signal;

[0069] The risk time anchor point statistics module collects the generated first positioning risk record, second risk record, third time sequence abnormal risk record and fourth micro-contact abnormal risk record; and binds them to the corresponding burning task number to form the abnormal risk time anchor point of the i-th operation unit in the burning cycle, and submits them for chain operation through the blockchain node interface. The total number of the first positioning risk record, the second risk record, and the third time sequence abnormal risk record is counted to form the first abnormal risk anchor point number, and the total number of abnormal risk time anchor points of the fourth micro-contact abnormal risk record is counted to form the second abnormal risk anchor point number;

[0070] The operation unit locking module is used to automatically trigger the workstation locking, inspection mark or process review when the number of first abnormal risk anchor points of the i-th operation unit is greater than 3 times or the number of second abnormal risk anchor points is greater than 5 times.

[0071] The present invention provides a method and system for monitoring the dynamic burning status of a PCB board. It has the following beneficial effects:

[0072] (1) This method and system for monitoring the dynamic burning status of PCB boards uses an industrial camera to capture images of the entire probe crimping process. By combining image processing with a path modeling algorithm, it can extract the offset angle of the probe crimping path and its position error relative to the pad in real time, forming a positioning offset coefficient and a misconnection anomaly index. Compared with traditional alignment methods, this solution can identify "micro-misconnections" caused by clamping, module jitter, or probe arrangement errors within a millisecond cycle, effectively avoiding burning failures or PCB pad damage caused by mis-pressing.

[0073] By analyzing fluctuations in probe compression height across consecutive frames, the contact consistency index for each pin array is calculated to identify issues such as asynchronous crimping, individual probe failure, or insufficient pin contact depth. Compared to static inspection, this method can capture contact variability at the image level, enabling early detection of "local anomalies" within the probe array and improving programming quality stability.

[0074] By using image recognition to identify the moment of contact and comparing it with the power-on control signal timestamp, and calculating the contact timing offset index, it is possible to accurately determine whether there are asynchronous phenomena such as "power-on before contact" or "power-on delayed by contact", thereby avoiding the risks of abnormal circuit connection and disconnection, abnormal data burning or hardware damage caused by operation timing mismatch.

[0075] This method, for the first time, introduces the fusion calculation of image features (such as compression integrity and needle array uniformity) and micro-voltage jitter signals to extract the micro-contact impedance anomaly coefficient. It can effectively identify instantaneous contact resistance changes caused by contact contamination, needle tip oxidation, jitter instability, etc., and make up for the "contact sub-health" problem that traditional voltage monitoring cannot detect, thereby improving the burning success rate.

[0076] The present invention can realize full-process, full-element, and full-time domain monitoring of the probe-board contact behavior without interfering with the existing burning process, greatly improving the automation, precision control, and abnormal response capabilities of the PCB burning process, and is suitable for the field of high-density, high-reliability electronic assembly.

[0077] (2) This method and system for monitoring the dynamic burning status of PCB boards achieves high-precision dynamic monitoring of key details of the burning process by integrating multiple modules such as image acquisition, path offset recognition, crimping consistency evaluation, timing analysis and micro-contact resistance monitoring. The system can not only identify potential problems such as probe bias, inconsistent crimping depth, and contact timing misalignment, but also collect micro-voltage jitter at the moment of contact in real time to further identify micro-contact anomalies. Combined with the blockchain node's chain mechanism, a traceable "time anchor point" is formed for various abnormal risk records to improve data credibility. At the same time, a statistical mechanism is introduced to set a trigger threshold. Once the number of abnormalities exceeds the limit, the operation unit can be automatically locked to prevent bad boards from flowing into subsequent links, thereby improving the burning consistency, equipment stability and production line intelligence level, and effectively reducing maintenance costs and failure risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 This is a schematic diagram of the steps of a method for monitoring the dynamic burning status of a PCB board according to the present invention;

[0079] Figure 2 The present invention is a schematic flow chart of a PCB board dynamic burning status monitoring system. DETAILED DESCRIPTION

[0080] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0081] Example 1

[0082] See also Figure 1 The present invention provides a method for monitoring the dynamic burning status of a PCB board, comprising the following steps:

[0083] Step 1: Identify and number multiple key operating units of the burning station; use the industrial camera to collect the image sequence set of the i-th operating unit within the burning cycle t, identify the micro-angle deviation of the probe's pressing path based on the image processing algorithm, and calculate the positioning deviation coefficient of the i-th operating unit within the burning cycle t and misconnection anomaly index and evaluating, and when there is a risk of abnormal positioning, generating a first positioning risk record and triggering a first alarm signal;

[0084] Step 2: Based on the image sequence of the i-th operation unit within the burning cycle t, identify the probe compression height change in the continuous image frames and calculate the contact consistency index of the i-th operation unit And evaluate, if there is a risk of abnormal contact of the probes in the same group, generate a second risk record and trigger a second alarm signal;

[0085] Step 3: Extract the image frame at the moment of contact from the image sequence set of the i-th operation unit within the burning cycle t, and identify and obtain the compression deformation change rate of the k-th probe in the f-th frame. , Brightness change of the needle tip reflection area , the edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame After determining the contact occurrence time, collect the contact action occurrence time Power-on control time , calculate the contact timing offset index of the i-th operating unit And evaluate, when it is determined that there is a risk of abnormal contact timing, generate a third abnormal timing risk record and trigger a third alarm signal;

[0086] Step 4: After determining the contact occurrence time in step 3, collect the micro-voltage jitter value sequence, combine it with the image sequence set, and calculate the micro-contact impedance anomaly coefficient of the i-th operating unit within the burning cycle t And evaluate, when it is determined that there is a micro-contact abnormality, generate a fourth micro-contact abnormality risk record and trigger a fourth alarm signal.

[0087] In this embodiment, an industrial camera captures images of the entire probe crimping process. Combined with image processing and path modeling algorithms, the offset angle of the probe crimping path and its positional error relative to the pad are extracted in real time, generating a positioning offset coefficient and misconnection anomaly index. Compared to traditional alignment methods, this solution can identify "micro-misconnections" caused by clamping, module jitter, or probe placement errors within milliseconds, effectively preventing programming failures or PCB pad damage caused by misalignment.

[0088] By analyzing fluctuations in probe compression height across consecutive frames, the contact consistency index for each pin array is calculated to identify issues such as asynchronous crimping, individual probe failure, or insufficient pin contact depth. Compared to static inspection, this method can capture contact variability at the image level, enabling early detection of "local anomalies" within the probe array and improving programming quality stability.

[0089] By using image recognition to identify the moment of contact and comparing it with the power-on control signal timestamp, and calculating the contact timing offset index, it is possible to accurately determine whether there are asynchronous phenomena such as "power-on before contact" or "power-on delayed by contact", thereby avoiding the risks of abnormal circuit connection and disconnection, abnormal data burning or hardware damage caused by operation timing mismatch.

[0090] This method, for the first time, introduces the fusion calculation of image features (such as compression integrity and needle array uniformity) and micro-voltage jitter signals to extract the micro-contact impedance anomaly coefficient. It can effectively identify instantaneous contact resistance changes caused by contact contamination, needle tip oxidation, jitter instability, etc., and make up for the "contact sub-health" problem that traditional voltage monitoring cannot detect, thereby improving the burning success rate.

[0091] The present invention can realize full-process, full-element, and full-time domain monitoring of the probe-board contact behavior without interfering with the existing burning process, greatly improving the automation, precision control, and abnormal response capabilities of the PCB burning process, and is suitable for the field of high-density, high-reliability electronic assembly.

[0092] Example 2

[0093] This embodiment is explained in Example 1. Specifically, step 1 includes:

[0094] S11, classify and identify multiple key operating units in the burning station, register them by number, and generate an operating unit index table;

[0095] The operation unit includes: a burning needle bed assembly, a clamping mechanism and a feeding robot arm or a handling module; each operation unit number is bound to the control system address and task scheduling interface, and recorded in the equipment main console database; it helps to improve the system's management consistency and scheduling clarity for multiple types of operation units.

[0096] S12. When receiving a programming task instruction, the PCB board number to be programmed is bound to the assigned work unit, and a board programming job binding record table is generated. The record fields include the work unit number, board number, task number, and planned workstation coordinates. This helps track the actual execution unit of the task and prevents operations caused by equipment misassignment, number confusion, etc.

[0097] Example 3

[0098] This embodiment is explained in Example 1. Specifically, step 1 also includes:

[0099] S13, before each burning cycle t starts, the industrial camera set above or on the side of the burning needle bed is used to collect dynamic image frames of the needle bed crimping process, and the image collection frequency is not less than 30 frames / second to form an image sequence set of the i-th operation unit in the burning cycle t. ;

[0100] S14, using the image detection algorithm to identify the probe arrangement and the board pad alignment area in the first image, using the standard crimping path vector As a benchmark, identify the path offset angle in each frame image Offset from position , calculate the positioning offset coefficient of the i-th operation unit in the burning cycle t :

[0101]

[0102] in, represents the offset angle between the probe trajectory and the standard path in the jth frame, represents the two-dimensional pixel displacement of the needle tip relative to the standard crimping starting point in the jth frame; n represents the number of image frames, and Indicates the weight coefficient; if the probe crimping process requires higher directional consistency (for example, it needs to be pressed down at a fixed angle), then set On the contrary, if the direction can be biased and the position is more critical, then Can be smaller, the initial weight is: , , which ensures that the offset position is offset are more fully considered, improving the sensitivity and reliability of risk identification.

[0103] S15, synchronously calculating the misconnection abnormality index of the i-th operation unit in the programming cycle t based on the relative position image of the probe needle tip and the board pad collected in the programming cycle t of the i-th operation unit :

[0104]

[0105] Where N represents the total number of probe tip and corresponding pad pairs measured; represents the actual three-dimensional coordinate of the k-th probe tip, Represents the standard three-dimensional coordinates corresponding to the k-th pad; represents the Euclidean distance;

[0106] S16, preset positioning offset threshold Dth and misconnection offset threshold CYth, if or If any one of the following conditions is met, it is determined that the i-th operation unit has a risk of abnormal probe crimping positioning during the current programming cycle. At this time, the first positioning risk record is recorded, including: operation unit number, programming task number, offset image frame ID, offset angle and displacement value, and the first alarm signal is triggered immediately.

[0107] The positioning offset threshold Dth and misconnection offset threshold CYt are determined based on the mechanical design tolerances and maximum allowable error range of the programming equipment and probe assembly, combined with the manufacturer's technical specifications. These parameters are typically obtained through factory testing and calibration of the equipment. Based on production line experience and feedback from quality engineers, and according to actual programming yield and defect levels, the thresholds are adjusted to achieve a balance between early warning and false alarms, ensuring timely detection of anomalies while preventing frequent false alarms that impact production efficiency.

[0108] In this embodiment, the whole process of needle bed crimping can be captured in real time through S13-S16 before the start of burning, and the positioning offset coefficient and misconnection abnormality index are calculated with reference to the standard path vector. When any index exceeds the threshold, a positioning risk record is generated and the first alarm signal is triggered; it is convenient to find the probe angle deviation or XY position displacement abnormality within the millisecond time window of the needle bed pressing down, so as to avoid the board pad being mispressed or leaked; synchronously comparing the three-dimensional coordinates of the needle tip and the pad, it can timely identify misconnection problems such as the overall offset and flip placement of the single-row needle array, and improve the alignment accuracy; the risk information carries the image frame ID and the angle / displacement, which can provide direct evidence for subsequent tracing and rapid reproduction, shortening the fault location time; the early alarm stops the device before the pad is damaged or re-burned multiple times, reducing board damage and tooling wear, and stabilizing the process yield.

[0109] Example 4

[0110] This embodiment is explained in Example 1. Specifically, step 2 includes:

[0111] S21, in S12, the image sequence set of the i-th operation unit in the burning cycle t In the middle, the key image frame where the probe is about to be fully pressed and maintain a stable contact state is extracted, which is the middle and rear frame and is recorded as the stable frame interval. ;

[0112] S22, in the stable frame interval Extract the compressed height value set of the same group of probes in the i-th operation unit within the burning cycle t ;in, to represents the compression height identified by the 1st to the mth probes in the same group of probes in the i-th operation unit;

[0113] S23, calculate the contact consistency index of the probes in the same group of the i-th operation unit , which is used to measure the overall discreteness of the compression height. The formula is as follows:

[0114]

[0115] in, Represents the set of compressed height values ​​of the same group of probes in the i-th operation unit during the programming cycle t The standard deviation of Represents the set of compressed height values ​​of the same group of probes in the i-th operation unit during the programming cycle t The average value of ∈[0,1], the closer it is to 1, the more consistent the contact;

[0116] Inconsistent probe compression can cause some probes to not be pressed enough (the compression height is too shallow), which can lead to failure to establish reliable contact with the corresponding pads, resulting in data writing failure, test failure, or even programming interruption on some signal pins; some probes are pressed too deep, while others are not pressed to the bottom, which leads to concentrated pressure distribution; mechanical life is accelerated, causing needle spring failure, needle body bending, and needle sleeve deformation; secondary problems such as probe offset or scratching of PCB pads are likely to occur; temporary voltage differences are generated between multiple data bits, power bits, or GND bits due to asynchronous contact; each crimp height is different, abnormal behavior is unstable, and the problem of "test OK but customer test OK" occurs.

[0117] S24, preset consistency assessment threshold Cth, if , it is determined that there is a risk of abnormal contact of the same group of probes in the current burning cycle of the i-th operation unit, a second risk record is generated, and a second alarm signal is immediately triggered.

[0118] The source of the consistency assessment threshold Cth is: by statistically analyzing the consistency index of the same group of probe compression heights in a large number of qualified burning samples, obtaining their average value and standard deviation, and setting the threshold by referring to the 5th percentile or the lower limit of the standard deviation to divide the boundary between "abnormal" and "normal". The mean is 0.92 and the standard deviation is 0.03, so the threshold Cth can be set to 0.85 or 0.88.

[0119] The alarm signal promptly intercepts the burning action of poor crimping, prevents the outflow of unqualified boards, improves batch consistency and release reliability, avoids the premature scrapping of individual probes due to overpressure, extends the life of the probe assembly, and reduces maintenance costs;

[0120] In this embodiment, by quantitatively evaluating the consistency of the compression height of probes within a group, abnormal fluctuations and deviations during the probe crimping process can be effectively identified, thereby promptly detecting potential contact risks. Specifically, the contact consistency index, calculated by the ratio of the standard deviation to the mean, reflects the overall dispersion of the compression heights, ensuring that each probe within the probe group is subjected to uniform and stable force, thereby avoiding poor signal contact due to insufficient compression of individual probes, or mechanical damage and secondary defects due to excessive compression.

[0121] This method provides timely warnings for various potential hazards caused by inconsistent probe compression, including data write failures, programming interruptions, shortened mechanical lifespan, and board pad damage, thereby improving the system's monitoring accuracy and response speed. By setting appropriate consistency thresholds and incorporating an alarm mechanism, the quality control level of the programming process is significantly improved, ensuring the consistency and reliable release of PCB board batches. Furthermore, it effectively extends the lifespan of the probe assembly, reduces maintenance frequency and costs, and achieves economical and stable equipment operation and maintenance.

[0122] Example 5

[0123] This embodiment is explained in Example 1. Specifically, step three includes:

[0124] S31. Before each programming cycle begins, the control system outputs a power-on control signal and records the control output timestamp. And bind it with the operation unit number i and the burning task number, and write it into the task log table. The fields include: task number, operation unit number, control power-on time and power-on signal flag;

[0125] S32, extracting the image frame at the time of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and identifying and obtaining the compression deformation change rate of the k-th probe in the f-th frame , Brightness change of the needle tip reflection area , the edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame ;

[0126] And preset the compression variable mutation threshold Hth, reflection change threshold Lth, edge gradient mutation threshold Gth, and minimum stable contact time threshold Tmin;

[0127] The compression deformation mutation threshold, Hth, is specifically determined to determine whether the probe compression depth has "mutated" between frames. This threshold is often used to identify contact moments or mis-pressing. Data on the rate of change in compression deformation between frames is extracted from thousands of sets of compression images to identify the difference between normal contact and non-contact phases. For example, if the total compression stroke of the probe is 1.0 mm and the allowable deformation change rate is 10%, Hth can be set to 0.1. In failure records, sudden changes in the compression curve are typically greater than 0.12 or 0.15, so Hth can be set to 0.12.

[0128] The reflection change threshold (Lth) is specifically used to determine the sudden brightness change caused by the change in the reflective area at the moment of probe contact, which is an important optical signal of the contact point state change. By comparing the frames before and after contact using the grayscale histogram, the average brightness difference in the reflective area is typically between 20 and 40. The relative grayscale change (Lth) is preset based on the industrial camera's resolution, lighting consistency, and other factors, typically at 10% to 15% of the original grayscale value.

[0129] The edge gradient mutation threshold Gth is derived from a sudden change in the gradient amplitude of the contact area image, which usually indicates that the probe pressure has caused a significant change in the edge clarity of the area. Using the Sobel or Laplacian operator, we test the edge recognition of the contact area, record the difference in average gradient energy under different states, and calculate the rate of change of the mean gradient before and after contact. Generally, a sudden change is considered when the average gradient in the 10px to 20px area exceeds Gth (e.g., Gth = 15).

[0130] The source of the minimum stable contact time threshold Tmin is: the minimum time the probe must maintain a stable crimping state from the moment of contact to ensure correct electrical contact. For example, the stable crimping delay before programming is designed to be ≥30m, and some high-reliability probes require a stable electrical signal channel establishment time of ≥20ms. Programming failures often correspond to unstable crimping, repeated bouncing, etc., and the time is less than 15ms.

[0131] S33. When any two or more of the following four conditions are recognized, it is determined that the contact occurs, including: Condition 1 is a sudden change in the probe compression deformation, indicating a significant jump in the compression degree between the previous and next frames: Condition 2: The change in brightness or area of ​​the needle tip reflective area exceeds the threshold, indicating a change in the optical reflection state: Condition 3: The edge of the contact area between the probe and the PCB changes dramatically, and the gradient field in the contact area in the image changes suddenly: Condition 4: The duration of the crimping state reaches the stability threshold. Starting from frame f, the probe maintains a stable crimping state for more than the set time: ;

[0132] S34: Determine the contact occurrence time based on S33, extract the image frame from the contact occurrence time, and obtain the contact action occurrence time Power-on control time , calculate the contact timing offset index of the i-th operating unit :

[0133]

[0134] S35, set the timing offset threshold Tth, when , it is determined that the i-th operating unit in the current cycle has a contact timing abnormality risk, marked as the third timing abnormality risk record of the i-th operating unit in cycle t, and the third alarm signal is immediately triggered.

[0135] The source of the timing offset threshold Tth threshold is the maximum allowable offset between the probe compaction time and the power-on control time. Too early or too late may cause programming failure or physical damage. The average offset of a normal operating unit is generally 15ms to 25ms.

[0136] Some ICs require a delay of 10ms or more before powering on after crimping to prevent transient arcing. Programming failures with a timing offset exceeding 40ms are a significant occurrence. Therefore, the timing offset threshold, Tth, is set to 35ms or 40ms.

[0137] This embodiment uses image processing and control timing to accurately identify the actual contact moment between the probe and the PCB board, and compares it with the power-on signal time output by the control system, quantifying the degree of timing matching between contact and power-on, thereby constructing a contact timing offset index, effectively solving the problems of burning failure, signal loss or data errors caused by the mismatch between "logical power-on" and "physical contact" in traditional methods.

[0138] Compared with the method that relies solely on electrical continuity judgment, this method combines multiple key features in the image, such as compression amount, reflection changes, edge gradient mutations, and crimping state continuity. By setting multi-threshold criteria, it can more stably identify the moment when the probe contacts actually complete contact, and has strong environmental adaptability and anti-interference capabilities.

[0139] The specific beneficial effects are: it can identify abnormal behaviors such as premature power-on (voltage is applied before contact) or delayed power-on (power is applied after contact); reduce the burning failure rate caused by contact-power-on asynchrony; realize closed-loop monitoring of the probe crimping action and the control system output behavior, and improve the timing consistency and reliability of the automated burning process.

[0140] Example 6

[0141] This embodiment is explained in Example 5. Specifically, the S32 includes:

[0142] S321, identifying and obtaining the rate of change of the compression deformation of the k-th probe in the f-th frame The specific method is:

[0143] S3201, extract the image frame at the moment of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and extract the probe compression height of the f-th frame and the previous frame and :

[0144]

[0145] in, Indicates the calibration distance from the needle bed reference plane to the camera, represents the depth distance from the kth probe tip to the camera in frame f;

[0146] S3202: Calculate the compression deformation rate of the kth probe in the fth frame using the following formula: :

[0147]

[0148] S322, identifying and obtaining the brightness change of the tip reflective area of ​​the k-th probe in the f-th frame The specific method is:

[0149] S3221, extracting the image frame at the moment of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and presetting the high reflection grayscale threshold , to identify the high reflective area of ​​the needle tip :

[0150]

[0151] in, Represents the pixel grayscale in frame f;

[0152] S3222, calculate the average brightness of the k-th probe in the f-th frame :

[0153]

[0154] S3223, the brightness change of the tip reflection area of ​​the k-th probe in the f-th frame :

[0155]

[0156] in, represents the brightness change of the tip reflection area of ​​the k-th probe in the previous frame f-1;

[0157] S323, identifying and obtaining the edge gradient amplitude of the image of the contact area between the kth probe and the PCB :

[0158] S3231, extracting the image frame at the time of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and identifying the image coordinates of the k-th probe in the f-th frame , construct a rectangular region of interest centered at the coordinates , the size is set to W×H, where W is the width and H is the height, covering the probe contact PCB area;

[0159] S3232, in the rectangular region of interest In the example, Sobel operator is used to calculate all pixel points. , calculate its grayscale gradient amplitude:

[0160]

[0161] in, Represents the grayscale gradient along the x direction obtained by the horizontal Sobel convolution calculation;

[0162] Represents the grayscale gradient along the y direction obtained by Sobel convolution in the vertical direction;

[0163] S3233, in the rectangular region of interest In the frame, the gray gradient amplitude of all pixels is averaged to get the k The average gradient energy of the probe contact area is:

[0164]

[0165] S3234, calculate the edge gradient amplitude of the image of the contact area between the kth probe and the PCB :

[0166]

[0167] in, Indicates the edge gradient amplitude of the image of the contact area between the k-th probe and the PCB in the previous frame f-1.

[0168] In this embodiment, the rate of change of the compression deformation of the kth probe in the fth frame is , Brightness change of the needle tip reflection area and the edge gradient amplitude of the probe and PCB contact area image The beneficial effects are calculated as:

[0169] Promote the improvement of positioning accuracy, the three-dimensional depth difference can capture the micro-displacement of the crimping moment in milliseconds, avoid "insufficient pressure" or "over-pressure damage to the disk", and the brightness change of the needle tip reflective area Reflects the difference in reflection when the needle tip contacts or leaves the pad, providing optical confirmation complementary to the mechanical signal; the edge gradient amplitude of the image of the contact area between the probe and the PCB Detecting the clear outline of the needle tip and pad boundary during formation can effectively distinguish real contact from out-of-focus and stains; it is used to improve the recognition sensitivity of hidden faults such as "power on before contact", "partial floating needle" or "tilted crimping", and reduce false alarms and missed detections.

[0170] Example 7

[0171] This embodiment is explained in Example 1. Specifically, step 4 includes:

[0172] S41: When S33 determines the conditions of the contact occurrence moment, extract the image frame from the contact occurrence moment and obtain the contact action occurrence time Power-on control time During the process, the contact voltage micro-jitter signal is collected synchronously. From the moment the probe group contacts the PCB board in the i-th operation unit to the burning data transmission stage, the instantaneous voltage waveform is collected at a frequency of ≥10kHz, and the micro-voltage jitter value sequence is recorded: , to Indicates the voltage value from the 1st sampling time to the Bth sampling time;

[0173] This sequence is used to subsequently calculate the jitter amplitude and frequency characteristics in the micro-contact state;

[0174] S42, calculating the peak-to-peak amplitude normalized value of the micro-voltage jitter sequence of the i-th operation unit within the programming cycle t according to the micro-voltage jitter value sequence :

[0175]

[0176] in, The preset maximum allowable voltage jitter peak-to-peak amplitude reference value;

[0177] S43. Calculate the normalized value of the standard deviation of the jitter sequence based on the micro-voltage jitter value sequence. :

[0178]

[0179] in, It is the preset maximum allowable voltage jitter standard deviation reference value;

[0180] S44, combining the peak-to-peak amplitude normalized value of the micro-voltage jitter sequence of the i-th operation unit within the programming cycle t and the normalized value of the standard deviation of the jitter sequence , calculate the comprehensive voltage fluctuation index of the i-th operating unit during the programming cycle t :

[0181]

[0182] In the formula, Represents the peak-to-peak amplitude normalized value of the micro-voltage jitter sequence for the i-th operating unit within the programming cycle t weight, Indicates the normalized value of the jitter sequence standard deviation weight, and are constants, and , Indicates that and Weight, calculate the comprehensive voltage fluctuation index of the i-th operation unit within the burning cycle t ;

[0183] and Source of weight: In poor crimping, peak-to-peak mutations often indicate intermittent contact or short circuit, and high standard deviation indicates large contact fluctuations. The two have slightly different effects: if the system has high requirements for instantaneous contact reliability (such as concurrent fever recording, more sensitive signals), the weight should be increased. If you are concerned about overall stability (such as long-term continuous communication), the standard deviation weight Higher; initial range is: ∈[0.5,0.7], ∈[0.3,0.5].

[0184] S45, identifying and obtaining the pinpoint compression integrity index of the i-th operation unit from the image sequence set of the i-th operation unit within the burning period t and needle array regularity index The image penalty factor of the i-th job unit in the burning cycle t is calculated by the following formula ;

[0185]

[0186] in, and Indicates the "abnormality" or "defect degree" of the corresponding indicator, and the average of the two (divided by 2) is obtained , represents the comprehensive degree of image abnormality. The larger the value, the more serious the probe compression or arrangement abnormality detected in the image;

[0187] S46: Combine the comprehensive voltage fluctuation index of the i-th operating unit obtained in S44 and S45 within the programming cycle t and the image penalty factor of the i-th job unit in the burning cycle t , calculate the micro-contact impedance anomaly coefficient of the i-th operating unit during the programming cycle t :

[0188]

[0189] In the formula, and are constants, and , Indicates that and Weight, calculate the micro-contact impedance anomaly coefficient of the i-th operating unit within the burning cycle t ;

[0190] S47, preset impedance abnormality threshold Zth, if , it is determined that the i-th operating unit has a micro-contact abnormality risk during the current burning cycle, specifically a comprehensive contact abnormality in the mechanical and electrical dimensions; and the fourth alarm signal is immediately triggered.

[0191] and The source of weight: reflects the contribution of "electrical signal jitter" and "image abnormality" to contact reliability. If the system uses high-precision image analysis (such as needle tip detection accuracy less than 0.1mm), the image reliability is higher and the weight can be appropriately increased. If the image system is easily disturbed (such as reflection, blur, angle error), the electrical signal is more accurate, so the ; The initial range is: ∈[0.4,0.6], ∈[0.4,0.6].

[0192] The source of the impedance anomaly threshold Zth, collecting the micro-contact impedance anomaly coefficients of normal and abnormal cycles The value distribution is used to extract the mean and lower limit of abnormality; it is set to the mean of the normal distribution + 2 times the standard deviation, or the median of the minimum Z value in the abnormal samples.

[0193] In this embodiment, the comprehensive voltage fluctuation index Reflects the probe-pad instantaneous conduction quality; image penalty factor Characterize mechanical defects such as insufficient tip compression and skewed arrangement. Shape micro-contact impedance anomaly coefficient through weight fusion It can not only identify false fluctuations caused by electrical noise, but also detect hidden poor contact caused by pure mechanical mismatch, reducing the probability of false alarm / missed alarm in single-dimensional monitoring.

[0194] Normalized peak-to-peak amplitude value of the micro-voltage jitter sequence of the i-th operating unit within the programming cycle t and the normalized value of the standard deviation of the jitter sequence The image abnormality can be recorded for a long time to form a health curve; when the micro-contact impedance abnormality coefficient of the i-th operating unit within the burning cycle t If the impedance anomaly threshold Zth is less than or equal to the threshold, preventive maintenance is indicated when the threshold is not exceeded, extending the life of the needle bed and fixture. If the alarm is caused by voltage jitter but the image anomaly is low, priority is given to checking the power supply filter and signal integrity; otherwise, focus on probe wear or the positioning mechanism. This layered information improves fault location efficiency and repair accuracy.

[0195] Example 8

[0196] This embodiment is explained in Example 1. Specifically, S45 includes:

[0197] S451, identifying and obtaining the pinpoint compression integrity index of the i-th operation unit from the image sequence set of the i-th operation unit within the burning period t The specific steps are:

[0198] Assume that the compression depth of the mth probe identified in the crimping state frame is , the compression depth target is , and the acceptance deviation is preset to ,satisfy The crimping is judged to meet the standards, and the needle tip compression integrity index of the i-th operation unit is calculated using the following formula :

[0199]

[0200] in, represents the indicator function, which is 1 if the condition is met, otherwise 0, and M is the total number of needles; The closer it is to 1, the better the overall compression; lower values ​​indicate inconsistent or inadequate compression.

[0201] S452: Identify and obtain the needle arrangement regularity index of the i-th operating unit from the image sequence set of the i-th operating unit within the burning cycle t. The specific steps are:

[0202] The nominal center pixel coordinate of each probe head is , perform linear fitting on the set of points arranged horizontally or vertically on the probe to obtain the needle arrangement neatness index of the i-th operation unit :

[0203]

[0204] in, It represents the vertical deviation of the mth probe from the fitted line after the linear equation is obtained by the least squares method. a and b are the slope and intercept of the fitted line. D represents the normalized scale parameter, which is set to the maximum allowable array deviation; for example, 20px or the region height.

[0205] The needle arrangement neatness index of the i-th operating unit ∈[0,1]: reflects whether the probes in the same group are arranged straight, skewed or offset.

[0206] In this embodiment, a tip compression integrity index is generated by calculating whether the needle tip compression depth meets the target value and calculating the compliance rate. This index can be used to determine whether the probe faces risks such as overall insufficient compression or individual needle row failure, replacing traditional single-point visual or pressure-sensing judgments and improving assessment reliability. A least-squares method is used to linearly fit the needle row arrangement path to construct a needle row alignment index. This index effectively identifies transverse and longitudinal unevenness of the probe caused by mechanical offset, fatigue deformation, or loose assembly, which are alignment defects that are difficult to detect with traditional methods. The index format (∈[0,1]) ensures good interpretability and quantification of the assessment results: values ​​closer to 1 indicate a more ideal overall state. Low values ​​in either index can provide early warning of potential problems such as unstable signal contact, inconsistent contact time, and brief power outages. These two indices, forming the basis of the image penalty factor, not only reflect the degree of physical deviation but also further influence the accuracy of the micro-contact impedance anomaly coefficient judgment, contributing to the establishment of a comprehensive mechanical, electrical, and image monitoring system.

[0207] Example 9

[0208] This embodiment is an explanation of the embodiment 1. Specifically, it also includes step 5, collecting the first positioning risk record, the second risk record, the third timing abnormality risk record, and the fourth micro-contact abnormality risk record generated in steps 1 to 4; and binding them with the corresponding burning task number to form the abnormality risk time anchor point of the i-th operation unit in the burning cycle;

[0209] Count the total number of the first positioning risk record, the second risk record, and the third time series abnormal risk record to form the first abnormal risk anchor point number; count the total number of abnormal risk time anchor points of the fourth micro-contact abnormal risk record to form the second abnormal risk anchor point number;

[0210] And submit it through the blockchain node interface for chain operation. After the chain is successfully uploaded, if the number of first abnormal risk anchor points of the i-th work unit is greater than 3 times or the number of second abnormal risk anchor points is greater than 5 times, the work station lock, inspection mark or process review will be automatically triggered.

[0211] In this embodiment, by binding multiple types of abnormal risk records, such as positioning, contact consistency, timing offset, and micro-contact, to specific programming task and operation unit numbers, structured abnormal risk time anchor data is generated, facilitating subsequent retrieval, analysis, and accountability. Various abnormal events are classified and counted, specifically separating conventional structural risks (positioning, contact, timing) from microscopic electrical risks (micro-contact). Threshold criteria are set for each, supporting more refined abnormality identification strategies and avoiding misjudgments and underreporting. After classification and statistics, all generated abnormal risk anchors are uniformly submitted to the blockchain node, ensuring tamper-proof documentation of abnormal programming process records and enhancing the authority of the traceability mechanism. This approach is suitable for manufacturing scenarios with high reliability and strong compliance requirements.

[0212] If the statistical results trigger the set threshold (for example, the first type of abnormality is greater than 3 times or the second type of abnormality is greater than 5 times), the system can automatically lock the current workstation, mark it for inspection, or push it to the process review link, reducing human intervention and preventing the expansion of batch abnormalities in a timely manner.

[0213] Example 10

[0214] A PCB board dynamic burning status monitoring system, please refer to Figure 2 ,include:

[0215] The operation unit identification module is used to classify and identify multiple key operation units within the burning station, register them by number, and generate an operation unit index table; the operation unit includes: a burning needle bed assembly, a clamping mechanism, and a feeding robot arm or a handling module; each operation unit number is bound to the control system address and task scheduling interface, and recorded in the equipment console database;

[0216] The image acquisition module is used to capture dynamic image frames of the needle bed crimping process before each burning cycle t by using an industrial camera set above or on the side of the burning needle bed. The image acquisition frequency is not less than 30 frames per second, forming an image sequence set of the i-th operation unit within the burning cycle t;

[0217] Positioning or misconnection status analysis module, used to identify the micro-angle deviation of the probe pressing path based on image processing algorithm, and calculate the positioning deviation coefficient of the i-th operation unit within the burning cycle t and misconnection anomaly index and evaluating, and when there is a risk of abnormal positioning, generating a first positioning risk record and triggering a first alarm signal;

[0218] The contact consistency analysis module is based on the image sequence of the i-th operation unit within the burning cycle t, identifies the change in the probe compression height in the continuous image frames, and calculates the contact consistency index of the i-th operation unit And evaluate, if there is a risk of abnormal contact of the probes in the same group, generate a second risk record and trigger a second alarm signal;

[0219] The contact timing analysis module is used to extract the image frame at the time of contact from the image sequence set of the i-th operating unit within the burning cycle t, and identify and obtain the compression deformation change rate of the k-th probe in the f-th frame. , Brightness change of the needle tip reflection area , the edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame After determining the contact occurrence time, collect the contact action occurrence time Power-on control time , calculate the contact timing offset index of the i-th operating unit And evaluate, when it is determined that there is a risk of abnormal contact timing, generate a third abnormal timing risk record and trigger a third alarm signal;

[0220] The micro-contact impedance analysis module is used to collect the micro-voltage jitter value sequence after determining the contact occurrence moment in the contact timing analysis module, and calculate the micro-contact impedance anomaly coefficient of the i-th operating unit within the burning cycle t in combination with the image sequence set. and evaluating, and when it is determined that there is a micro-contact abnormality, generating a fourth micro-contact abnormality risk record and triggering a fourth alarm signal;

[0221] The risk time anchor point statistics module collects the generated first positioning risk record, second risk record, third time sequence abnormal risk record and fourth micro-contact abnormal risk record; and binds them to the corresponding burning task number to form the abnormal risk time anchor point of the i-th operation unit in the burning cycle, and submits them for chain operation through the blockchain node interface. The total number of the first positioning risk record, the second risk record, and the third time sequence abnormal risk record is counted to form the first abnormal risk anchor point number, and the total number of abnormal risk time anchor points of the fourth micro-contact abnormal risk record is counted to form the second abnormal risk anchor point number;

[0222] The operation unit locking module is used to automatically trigger the workstation locking, inspection mark or process review when the number of first abnormal risk anchor points of the i-th operation unit is greater than 3 times or the number of second abnormal risk anchor points is greater than 5 times.

[0223] In this embodiment, high-precision dynamic monitoring of key details of the burning process is achieved through the collaboration of multiple modules such as integrated image acquisition, path offset recognition, crimping consistency assessment, timing analysis and micro-contact resistance monitoring. The system can not only identify potential problems such as probe bias, inconsistent crimping depth, and contact timing misalignment, but also collect micro-voltage jitter at the moment of contact in real time to further identify micro-contact anomalies. Combined with the chain mechanism of blockchain nodes, a traceable "time anchor point" is formed for various types of abnormal risk records to improve data credibility. At the same time, a statistical mechanism is introduced to set a trigger threshold. Once the number of abnormalities exceeds the limit, the operation unit can be automatically locked to prevent bad boards from flowing into subsequent links, thereby improving the burning consistency, equipment stability and production line intelligence level, and effectively reducing maintenance costs and failure risks.

[0224] The threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technicians in this field for each set of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0225] The above formulas are obtained by collecting a large amount of data and performing software simulation, and a formula close to the actual value is selected. The coefficients in the formula are set by those skilled in the art according to actual conditions. The above is only a preferred specific implementation method of the present invention, but the protection scope of the present invention is not limited to this. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, can make equivalent replacements or changes based on the technical solution and inventive concept of the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for monitoring the dynamic burning status of a PCB board, characterized in that: The following steps are involved: Step 1: Identify and number multiple key operating units of the burning station; The image sequence set of the i-th operating unit in the burning cycle t is collected by the industrial camera, the micro-angle deviation of the probe pressing path is identified based on the image processing algorithm, and the positioning deviation coefficient D of the i-th operating unit in the burning cycle t is calculated. i，t and misconnection abnormality index CY i，t and evaluating, and when there is a risk of abnormal positioning, generating a first positioning risk record and triggering a first alarm signal; Positioning offset coefficient D i，t The specific formula is: Among them, θ j represents the offset angle between the probe trajectory and the standard path in the jth frame, δ j represents the two-dimensional pixel displacement of the needle tip relative to the standard crimping starting point in the jth frame; n represents the number of image frames, α and β represent weight coefficients; Misconnection abnormality index CY i，t The specific formula is: Where N represents the total number of probe tip and corresponding pad pairs measured; represents the actual three-dimensional coordinate of the k-th probe tip, represents the standard three-dimensional coordinates corresponding to the k-th pad; ||·|| represents the Euclidean distance; Step 2: Based on the image sequence of the i-th operation unit within the burning cycle t, identify the probe compression height change in the continuous image frames and calculate the contact consistency index C of the i-th operation unit i，t And evaluate, if there is a risk of abnormal contact of the probes in the same group, generate a second risk record and trigger a second alarm signal; Contact consistency index C i，t , which is used to measure the overall discreteness of the compression height. The formula is as follows: in, Represents the set of compressed height values ​​of the same group of probes in the i-th operation unit during the programming cycle t The standard deviation of Represents the set of compressed height values ​​of the same group of probes in the i-th operation unit during the programming cycle t The average value of Step 3: Extract the image frame at the moment of contact from the image sequence set of the i-th operation unit within the burning cycle t, and identify and obtain the compression deformation change rate of the k-th probe in the f-th frame. Brightness change of the needle tip reflective area Edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame After determining the moment of contact, collect the time when the contact action occurs Power-on control time Calculate the contact timing offset index of the i-th operation unit And evaluate, when it is determined that there is a risk of abnormal contact timing, generate a third abnormal timing risk record and trigger a third alarm signal; The rate of change of the compression deformation of the kth probe at the fth frame Calculated using the following formula: in, and It means extracting the image frame at the moment of contact occurrence from the image sequence set of the i-th operation unit within the burning period t, and extracting the probe compression height of the f-th frame and the previous frame; Contact Timing Deviation Index The way to obtain is: According to the conditions for determining the contact occurrence time, the image frame from the contact occurrence time is extracted to obtain the contact action occurrence time Power-on control time Calculate the contact timing offset index of the i-th operation unit Step 4: After determining the contact occurrence time in step 3, collect the micro-voltage jitter value sequence, combine it with the image sequence set, and calculate the micro-contact impedance anomaly coefficient Z of the i-th operating unit within the burning cycle t i→t and evaluating, and when it is determined that there is a micro-contact abnormality, generating a fourth micro-contact abnormality risk record and triggering a fourth alarm signal; The micro-contact impedance anomaly coefficient Z of the i-th operating unit during the programming cycle t i→t The specific way to obtain it is: Combined with the peak-to-peak amplitude normalized value of the micro-voltage jitter value sequence of the i-th operating unit within the burning cycle t and the normalized value of the standard deviation of the jitter sequence Calculate the comprehensive voltage fluctuation index ΔV of the i-th operating unit during the programming cycle t i→t : In the formula, q1 represents the peak-to-peak amplitude normalized value of the micro-voltage jitter value sequence for the i-th operating unit within the programming cycle t Weight, q2 represents the normalized value of the jitter sequence standard deviation weight; From the image sequence set of the i-th operation unit within the burning cycle t, identify and obtain the pinpoint compression integrity index of the i-th operation unit and needle array regularity index The image penalty factor Pc of the i-th operation unit in the burning cycle t is calculated by the following formula: i→t ; Combined with the comprehensive voltage fluctuation index ΔV of the i-th operating unit during the programming cycle t i→t And the image penalty factor Pc of the i-th job unit in the burning cycle t i→t , calculate the micro-contact impedance anomaly coefficient Z of the i-th operating unit during the programming cycle t i→t :Z i→t =r1ΔV i→t +r2Pc i→t ; In the formula, r1 and r2 are weights; Tip Compression Integrity Index The specific steps are: Assume that the compression depth of the mth probe identified in the crimping state frame is h m , the compression depth target is h target , and the acceptance deviation is preset to δ h , satisfying |h m -h target |≤δ h The crimping is judged to meet the standards, and the needle tip compression integrity index of the i-th operation unit is calculated using the following formula Where I(·) represents the indicator function, which is 1 if the condition is met and 0 otherwise, and M is the total number of needles; The closer it is to 1, the better the overall compression; lower values ​​indicate inconsistent or inadequate compression. Identify and obtain the needle arrangement neatness index of the i-th operation unit The specific steps are: The nominal center pixel coordinate of each probe head is (x m ,y m ), perform linear fitting on the set of points arranged horizontally or vertically to obtain the needle arrangement neatness index of the i-th operation unit Among them, |y m -(ax m +b)| represents the vertical deviation of the mth probe to the fitted line after the linear equation is obtained by the least squares fitting method, a and b are the slope and intercept of the fitted line, and D represents the normalized scale parameter, which is set to the maximum allowable array deviation.

2. A PCB board dynamic burning state monitoring method according to claim 1, characterized in that, Step one includes: S11. Classify and identify multiple key operating units within the burning station, register them by number, and generate an operating unit index table; the operating units include: a burning needle bed assembly, a clamping mechanism, and a feeding robot arm or a handling module; each operating unit number is bound to the control system address and task scheduling interface, and recorded in the equipment main console database; S12. When receiving the burning task instruction, bind the PCB board number to be burned with the assigned operation unit, and generate a board burning operation binding record table, the record fields including the operation unit number, board number, task number and planned station coordinates.

3. A PCB board dynamic burning state monitoring method according to claim 1, characterized in that, Step 1 also includes: S13, before each burning cycle t starts, the industrial camera set above or on the side of the burning needle bed is used to collect dynamic image frames of the needle bed crimping process, and the image collection frequency is not less than 30 frames / second to form an image sequence set of the i-th operation unit in the burning cycle t. S14, using the image detection algorithm to identify the probe arrangement and the board pad alignment area in the first image, using the standard crimping path vector As a benchmark, identify the path offset angle θ in each frame image j and position offset δ j , calculate the positioning offset coefficient D of the i-th operating unit in the burning cycle t i，t； S15, synchronously calculating the misconnection abnormality index CY of the i-th operating unit in the programming cycle t based on the relative position image of the probe needle tip and the board pad collected in the programming cycle t of the i-th operating unit i，t； S16, preset positioning offset threshold Dth and misconnection offset threshold CYth, if D i,t >Dth or CY i，t >CYth, if any one of the following is met, it is determined that there is a risk of abnormal probe crimping positioning in the i-th operation unit during the current programming cycle. At this time, the first positioning risk record is recorded, including: operation unit number, programming task number, offset image frame ID, offset angle and displacement value, and the first alarm signal is triggered immediately.

4. A PCB board dynamic burning state monitoring method according to claim 1, characterized in that, Step 2 includes: S21, in S12, the image sequence set of the i-th operation unit in the burning cycle t In the middle, the key image frame where the probe is about to be fully pressed and maintain a stable contact state is extracted, which is the middle and rear frame and is recorded as the stable frame interval. S22, in the stable frame interval Extract the compressed height value set of the same group of probes in the i-th operation unit within the burning cycle t Among them, h1 to h m represents the compression height identified by the 1st to the mth probes in the same group of probes in the i-th operation unit; S23, calculate the contact consistency index C of the probes in the same group of the i-th operation unit i，t , used to measure the overall dispersion of compression height; S24, preset consistency evaluation threshold Cth, if C i，t <Cth, it is determined that there is a risk of abnormal contact of the same group of probes in the current programming cycle of the i-th operation unit, a second risk record is generated, and a second alarm signal is immediately triggered.

5. A PCB board dynamic burning state monitoring method according to claim 4, characterized in that, Step three includes: S31, before each burning cycle starts, the control system outputs a power-on control signal and records the control output timestamp t ctrl And bind it with the operation unit number i and the burning task number, and write it into the task log table. The fields include: task number, operation unit number, control power-on time and power-on signal flag; S32, extracting the image frame at the time of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and identifying and obtaining the compression deformation change rate of the k-th probe in the f-th frame Brightness change of the needle tip reflective area Edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame And preset the compression variable mutation threshold Hth, reflection change threshold Lth, edge gradient mutation threshold Gth, and minimum stable contact time threshold Tmin; S33. When any two or more of the following four conditions are recognized, it is determined that the contact occurs, including: Condition 1 is a sudden change in the probe compression deformation, indicating a significant jump in the compression degree between the previous and next frames: Condition 2 is that the brightness or area of ​​the needle tip reflective area changes beyond the threshold, indicating that the optical reflection state has changed: Condition 3 is that the edge of the contact area between the probe and the PCB changes dramatically, and the gradient field in the contact area in the image changes suddenly: Condition 4: The duration of the crimping state reaches the stability threshold. Starting from frame f, the probe maintains a stable crimping state for longer than the set time: S34: Determine the contact occurrence time based on S33, extract the image frame from the contact occurrence time, and obtain the contact action occurrence time Power-on control time Calculate the contact timing offset index of the i-th operation unit S35, set the timing offset threshold Tth, when It is determined that the i-th operating unit in the current cycle has a contact timing abnormality risk, which is marked as the third timing abnormality risk record of the i-th operating unit in cycle t, and the third alarm signal is immediately triggered.

6. A PCB board dynamic burning state monitoring method according to claim 5, characterized in that, S32 specifically includes: S321, identifying and obtaining the rate of change of the compression deformation of the k-th probe in the f-th frame The specific method is: S3201, extract the image frame at the moment of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and extract the probe compression height of the f-th frame and the previous frame and S3202, calculate the probe compression height of the f-th frame and the previous frame and The absolute value of the compression deformation change rate of the kth probe in the fth frame is obtained S322, identifying and obtaining the brightness change of the tip reflective area of ​​the k-th probe in the f-th frame The specific method is: S3221, extracting the image frame at the moment of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and presetting the high reflection grayscale threshold Fs th , to identify the high reflective area of ​​the needle tip Where G(x, y) represents the pixel grayscale in frame f; S3222, calculate the average brightness of the k-th probe in the f-th frame S3223, calculate the brightness change of the tip reflection area of ​​the k-th probe in the f-th frame S323, identifying and obtaining the edge gradient amplitude of the image of the contact area between the kth probe and the PCB S3231, extracting the image frame at the time of contact occurrence from the image sequence set of the i-th operation unit within the burning cycle t, and identifying the image coordinates of the k-th probe in the f-th frame Construct a rectangular region of interest centered at the coordinates The size is set to W×H, where W is the width and H is the height, covering the probe contact PCB area; S3232, in the rectangular region of interest In the example, the Sobel operator is used to calculate the grayscale gradient amplitude of all pixel points (x, y). S3233, in the rectangular region of interest The average grayscale gradient amplitude of all pixels is taken to obtain the average gradient energy of the kth probe contact area in the frame S3234, combined with the average gradient energy of the k-th probe contact area Calculate the edge gradient amplitude of the image of the contact area between the kth probe and the PCB 7. A PCB board dynamic burning state monitoring method according to claim 1, characterized in that, Step 4 includes: S41: When S33 determines the conditions of the contact occurrence moment, extract the image frame from the contact occurrence moment and obtain the contact action occurrence time Power-on control time During the process, the contact voltage micro-jitter signal is collected synchronously. From the moment the probe group contacts the PCB board in the i-th operation unit to the burning data transmission stage, the instantaneous voltage waveform is collected at a frequency of ≥10kHz, and the micro-voltage jitter value sequence is recorded: v1 to v B Indicates the voltage value from the 1st sampling time to the Bth sampling time; S42, calculating the peak-to-peak amplitude normalized value of the micro-voltage jitter value sequence of the i-th operation unit within the programming cycle t according to the micro-voltage jitter value sequence S43. Calculate the normalized value of the standard deviation of the jitter sequence based on the micro-voltage jitter value sequence. S44, combining the peak-to-peak amplitude normalized value of the micro-voltage jitter value sequence of the i-th operation unit within the programming cycle t and the normalized value of the standard deviation of the jitter sequence Calculate the comprehensive voltage fluctuation index ΔV of the i-th operating unit during the programming cycle t i→t ; S45, identifying and obtaining the pinpoint compression integrity index of the i-th operation unit from the image sequence set of the i-th operation unit within the burning period t and needle array regularity index And calculate the image penalty factor Pc of the i-th operation unit in the burning cycle t i→t ; S46: Combine S44 and S45 to obtain the comprehensive voltage fluctuation index ΔV of the i-th operating unit during the programming cycle t i→t And the image penalty factor Pc of the i-th job unit in the burning cycle t i→t , calculate the micro-contact impedance anomaly coefficient Z of the i-th operating unit during the programming cycle t i→t ; S47, preset impedance abnormality threshold Zth, if Z i→t >Zth, it is determined that the i-th operating unit has a micro-contact abnormality risk in the current programming cycle, and the fourth alarm signal is immediately triggered.

8. A PCB board dynamic burning state monitoring method according to claim 7, characterized in that, Specific S45 includes: S451, identifying and obtaining the pinpoint compression integrity index of the i-th operation unit from the image sequence set of the i-th operation unit within the burning period t The specific steps are: Assume that the compression depth of the mth probe identified in the crimping state frame is h m , the compression depth target is h target , and the acceptance deviation is preset to δ h , satisfying |h m -h target |≤δ h The crimping is judged to be up to standard, and the needle tip compression integrity index of the i-th operation unit is calculated. S452: Identify and obtain the needle arrangement regularity index of the i-th operating unit from the image sequence set of the i-th operating unit within the burning cycle t. The specific steps are: The nominal center pixel coordinate of each probe head is (x m ,y m ), perform linear fitting on the set of points arranged horizontally or vertically on the probes, and calculate the needle arrangement neatness index of the i-th operation unit 9. A PCB board dynamic burning state monitoring method according to claim 1, characterized in that, The process also includes: Step 5, collecting the first positioning risk record, the second risk record, the third timing abnormality risk record, and the fourth micro-contact abnormality risk record generated in Steps 1 to 4; and binding the records to the corresponding programming task numbers to form an abnormality risk time anchor point for the i-th operation unit in the programming cycle; Count the total number of the first positioning risk record, the second risk record, and the third time series abnormal risk record to form the first abnormal risk anchor point number; count the total number of abnormal risk time anchor points of the fourth micro-contact abnormal risk record to form the second abnormal risk anchor point number; And submit it through the blockchain node interface for chain operation. After the chain is successfully uploaded, if the number of first abnormal risk anchor points of the i-th work unit is greater than 3 times or the number of second abnormal risk anchor points is greater than 5 times, the work station lock, inspection mark or process review will be automatically triggered.

10. A PCB board dynamic burning status monitoring system, applied to a PCB board dynamic burning status monitoring method according to any one of claims 1 to 9, characterized in that: include: The operation unit identification module is used to classify and identify multiple key operation units in the burning station, register them by number, and generate an operation unit index table; The operation unit includes: a burning needle bed assembly, a clamping mechanism and a feeding robot arm or a handling module; each operation unit number is bound to the control system address and task scheduling interface, and recorded in the equipment console database; The image acquisition module is used to capture dynamic image frames of the needle bed crimping process before each burning cycle t by using an industrial camera set above or on the side of the burning needle bed. The image acquisition frequency is not less than 30 frames per second, forming an image sequence set of the i-th operation unit within the burning cycle t; Positioning or misconnection status analysis module, used to identify the micro-angle deviation of the probe pressing path based on image processing algorithm, and calculate the positioning deviation coefficient D of the i-th operation unit within the burning cycle t i，t and misconnection abnormality index CY i，t and evaluating, and when there is a risk of abnormal positioning, generating a first positioning risk record and triggering a first alarm signal; The contact consistency analysis module is based on the image sequence of the i-th operation unit within the burning cycle t, identifies the change in the probe compression height in the continuous image frames, and calculates the contact consistency index C of the i-th operation unit i，t And evaluate, if there is a risk of abnormal contact of the probes in the same group, generate a second risk record and trigger a second alarm signal; The contact timing analysis module is used to extract the image frame at the time of contact from the image sequence set of the i-th operating unit within the burning cycle t, and identify and obtain the compression deformation change rate of the k-th probe in the f-th frame. Brightness change of the needle tip reflective area Edge gradient amplitude of the image of the contact area between the probe and the PCB The stable time of the probe pressing state after the fth frame After determining the moment of contact, collect the time when the contact action occurs Power-on control time Calculate the contact timing offset index of the i-th operation unit And evaluate, when it is determined that there is a risk of abnormal contact timing, generate a third abnormal timing risk record and trigger a third alarm signal; The micro-contact impedance analysis module is used to collect the micro-voltage jitter value sequence after determining the contact occurrence moment in the contact timing analysis module, and calculate the micro-contact impedance anomaly coefficient Z of the i-th operating unit within the burning cycle t in combination with the image sequence set. i→t and evaluating, and when it is determined that there is a micro-contact abnormality, generating a fourth micro-contact abnormality risk record and triggering a fourth alarm signal; The risk time anchor point statistics module collects the generated first positioning risk record, second risk record, third time sequence abnormal risk record and fourth micro-contact abnormal risk record; and binds them to the corresponding burning task number to form the abnormal risk time anchor point of the i-th operation unit in the burning cycle, and submits them for chain operation through the blockchain node interface. The total number of the first positioning risk record, the second risk record, and the third time sequence abnormal risk record is counted to form the first abnormal risk anchor point number, and the total number of abnormal risk time anchor points of the fourth micro-contact abnormal risk record is counted to form the second abnormal risk anchor point number; The operation unit locking module is used to automatically trigger the workstation locking, inspection mark or process review when the number of first abnormal risk anchor points of the i-th operation unit is greater than 3 times or the number of second abnormal risk anchor points is greater than 5 times.

Images