Method, device and system for detecting defects of pin welding spots of plug-in component

Abstract

The invention discloses a defect detection method, device and system for a pin welding spot of a plug-in component. The defect detection method comprises the following steps: acquiring depth three-dimensional data of a detection material; performing high-precision template matching based on the brightness map of the depth stereo data, and determining the relative position of the detected material; performing plane correction by taking the PCB as a reference plane; according to the corrected depth data, the ROI of the detection area is set, and pin height detection, welding spot size detection, welding spot mass center height detection, welding spot tip pulling detection and the like are carried out respectively. The device not only can realize pin height detection and detection of defects such as non-uniform distribution of welding spots, too much or too little soldering tin amount, continuous welding, insufficient welding, too high welding spots, tip welding, welding slag around the welding spots and the like, but also has the characteristics of high efficiency, stability, high accuracy and the like; manual detection can be replaced, the labor cost is saved, and the welding spot defect detection efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of industrial product solder joint defect detection, and in particular to a method, device and system for detecting solder joint defects of plug-in component pins. Background Art

[0002] With the rapid development of electronic technology, the reliability and stability of electronic components are increasingly valued. Plug-in components (sockets: pins: connectors: terminals: LEDs, etc.) are widely used in electronic devices, and play an important role in connecting and transmitting signals in electronic devices. The quality of their pin solder joints directly affects the performance and life of electronic devices. In the processing and production of electronic products containing plug-in components, solder is usually used to solder the pins of plug-in components to the PCB circuit board, and in order to save labor, automatic machine welding is usually used for welding. However, due to problems such as welding process, automatic machine welding may also have abnormal soldering, such as: too much or too little solder, cold soldering, continuous soldering, and sharp soldering, which will affect the electrical connection and stability of components. Therefore, it is necessary to detect the pin solder joints to prevent bad equipment from entering the market; it is crucial to achieve effective detection of pin solder joint defects of plug-in components to improve product quality and reliability.

[0003] At present, the industry usually uses two methods to detect solder joint defects: manual inspection and machine vision inspection. The traditional welding inspection method uses the human eye to judge the quality of solder joints according to the welding qualification standards. However, there are many shortcomings in manual inspection, including low inspection efficiency and excessive reliance on workers' experience, resulting in unstable inspection accuracy. Based on the application of machine vision to the inspection of plug-ins, however, due to the small size and small spacing of the solder balls, the existing technology is difficult to check the welding quality by checking the height of the solder joints, which has low inspection efficiency and high missed detection rate and false detection rate. The traditional solder joint inspection method has problems such as low inspection efficiency and high false judgment rate, which is difficult to meet the high requirements of modern electronic manufacturing. Summary of the invention

[0004] In order to solve the deficiencies in the prior art, the present invention discloses a method, device and system for detecting defects in the solder joints of plug-in component pins, which are used to detect defects such as excessive or insufficient solder, cold solder joints, continuous solder joints, and pointed solder joints.

[0005] The present invention provides a method for defect detection of solder joints of plug-in components, comprising the following contents:

[0006] Collect deep stereo data of the inspected materials;

[0007] Perform high-precision template matching based on the brightness map of the depth stereo data to determine the relative position of the inspection material;

[0008] Use the PCB board as the reference surface for plane correction;

[0009] According to the corrected depth data, the ROI of the detection area is set to perform pin height detection, solder joint size detection, solder joint centroid height detection and solder joint tip detection respectively.

[0010] Result judgment.

[0011] Preferably, the specific implementation of collecting the depth stereoscopic data of the detected material is:

[0012] S11: Use a 3D line laser camera to obtain profile data;

[0013] S12: splicing the contour data to obtain the depth stereo data of the detected material, wherein the depth stereo data includes two types of data: a depth map and a brightness map. The depth stereo data needs to include at least two features common to the detected material to facilitate subsequent matching and positioning.

[0014] The preferred step is to perform high-precision template matching based on the brightness map of the depth stereo data to determine the relative position of the detected material, which is specifically implemented as follows:

[0015] S21: Selecting a brightness image of the depth stereo data of the initially detected material as a template image;

[0016] S22: manually draw an ROI in the common feature area of ​​the inspected material, and extract edge feature information in the ROI feature area based on a contour extraction operator;

[0017] S23: preprocess and optimize edge feature information to shield edge information with more noise;

[0018] S24: performing template matching on the brightness image of the material to be inspected according to the optimized edge feature information to obtain relative position information of the material to be inspected.

[0019] The preferred specific implementation of plane correction using the PCB board as the reference plane is:

[0020] S31: Based on the matched position information, position correction is performed to correct the deviation of the detected material in motion: assisting precise positioning, and establishing a reference for position deviation according to the matching points and angles in the template matching results;

[0021] S32: Then, according to the relative position offset of the preset calculation point in the feature matching result, the coordinate rotation offset of the ROI detection frame is realized, so that the ROI area can keep up with the changes in the image angle and pixels;

[0022] S33: selecting no less than three feature areas (which may be in the shape of rectangle, circle, point, etc.) on the area without components of the PCB board, and performing plane fitting on the depth data in the feature areas based on the least square method;

[0023] S34: Create a coordinate system based on the point-plane method. Select a point on the PCB board as the origin. Use the normal vector direction n = (a, b, -1) or n = (a, b, 1) (depending on the orientation of the fitting equation) of the fitted plane as the Z axis. Use two mutually perpendicular vectors on the plane as the X axis and Y axis to construct a transformation matrix from the original coordinate system (O-XoYoZo) to the new coordinate system (O-XYZ). The transformation matrix form is:

[0024]

[0025] Among them, [x, y, z] are the coordinates in the new coordinate system, [x o ,y o ,z o ] is the coordinate in the original coordinate system, R is the 3×3 rotation matrix, and T is the 3×1 translation vector.

[0026] Preferably, according to the corrected depth data, the ROI of the detection area is set, and the specific implementation of the pin height detection, solder joint size detection, solder joint centroid height detection and solder joint pull tip detection is as follows:

[0027] S41: Pin height detection, which is used to determine the pin height by comparing the threshold of the pin height pixel statistics;

[0028] S42: solder joint size detection, by comparing the solder joint area threshold to achieve uneven distribution of solder joints, detection of solder joints with too much or too little solder, continuous soldering and other defects;

[0029] S43: Solder point centroid height detection, by comparing the solder point centroid height threshold to detect defects such as cold solder joints and excessively high solder joints;

[0030] S44: Tip detection, by comparing the area threshold to detect defects such as solder tip and welding slag around solder joints.

[0031] The present invention also provides a system for defect detection of pin solder joints of plug-in components, comprising: a data acquisition module, a data preprocessing module, a defect detection module, and a result execution module;

[0032] The data acquisition module is used to collect the depth and stereoscopic data of the detected material;

[0033] The data preprocessing module is used to achieve high-precision template matching and plane correction of images;

[0034] The defect detection module is used to detect defects such as pin height, excessive or insufficient soldering of solder joints, cold solder joints, continuous solder joints, and pointed solder joints;

[0035] The result execution module is used to flow NG products and OK products into different processes respectively.

[0036] Preferably, the defect detection module includes: a pin height detection module, a solder joint size detection module, a solder joint centroid height detection module, and a pull tip detection module.

[0037] A device for detecting defects in solder joints of pins of plug-in components, and a method for detecting defects in solder joints of pins of plug-in components.

[0038] A preferred defect detection device for the solder joints of the pins of plug-in components also includes a counter chassis, an industrial computer, a programmable logic controller (PLC), a material conveying unit, a material tray, a detection material, a single-axis translation stage, a connecting sheet metal, a 3D line laser camera, a manipulator, and a temporary storage table for NG products. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of a method for defect detection of solder joints of plug-in component pins in an implementation case of the present invention.

[0040] Figure 2 The present invention is a schematic diagram of a device for detecting defects in solder joints of plug-in component pins according to an embodiment of the present invention.

[0041] Figure 3 The present invention is a schematic diagram of a defect detection system for a plug-in component pin solder joint according to an embodiment of the present invention.

[0042] Figure 4 The present invention is a flowchart of a defect detection system for a plug-in component pin solder joint in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The following is combined with Figure 1-4 The present invention is described in further detail.

[0044] Figure 1 The present invention provides a method for detecting defects in solder joints of plug-in components. Figure 1 ,include:

[0045] S1: Collecting the depth stereo data of the inspection material;

[0046] Perform high-precision template matching based on the brightness map of the depth stereo data to determine the relative position of the inspection material;

[0047] Use the PCB board as the reference surface for plane correction;

[0048] S4: According to the corrected depth data, set the ROI of the detection area, and perform pin height detection, solder joint size detection, solder joint centroid height detection and solder joint tip detection respectively.

[0049] S5: Result judgment.

[0050] Specific:

[0051] The specific implementation of collecting and testing the depth and stereoscopic data of the material is:

[0052] S11: Use a 3D line laser camera to obtain profile data;

[0053] S12: splicing the contour data to obtain the depth stereo data of the detected material, wherein the depth stereo data includes two types of data: a depth map and a brightness map. The depth stereo data needs to include at least two features common to the detected material to facilitate subsequent matching and positioning.

[0054] Based on the brightness map of the depth stereo data, high-precision template matching is performed to determine the relative position of the detected material.

[0055] S21: Selecting a brightness image of the depth stereo data of the initially detected material as a template image;

[0056] S22: manually draw an ROI in the common feature area of ​​the inspected material, and extract edge feature information in the ROI feature area based on a contour extraction operator;

[0057] S23: preprocess and optimize edge feature information to shield edge information with more noise;

[0058] S24: performing template matching on the brightness image of the material to be inspected according to the optimized edge feature information to obtain relative position information of the material to be inspected.

[0059] The specific implementation of plane correction using the PCB board as the reference plane is as follows:

[0060] S31: Based on the matched position information, position correction is performed to correct the deviation of the detected material in motion: assisting precise positioning, and establishing a reference for position deviation according to the matching points and angles in the template matching results;

[0061] S32: Then, according to the relative position offset of the preset calculation point in the feature matching result, the coordinate rotation offset of the ROI detection frame is realized, so that the ROI area can keep up with the changes in the image angle and pixels;

[0062] S33: selecting no less than three feature areas (which may be in the shape of rectangle, circle, point, etc.) on the area without components of the PCB board, and performing plane fitting on the depth data in the feature areas based on the least square method;

[0063] S34: Create a coordinate system based on the point-plane method. Select a point on the PCB board as the origin. Use the normal vector direction n = (a, b, -1) or n = (a, b, 1) (depending on the orientation of the fitting equation) of the fitted plane as the Z axis. Use two mutually perpendicular vectors on the plane as the X axis and Y axis to construct a transformation matrix from the original coordinate system (O-XoYoZo) to the new coordinate system (O-XYZ). The transformation matrix form is:

[0064]

[0065] Among them, [x, y, z] are the coordinates in the new coordinate system, [x o ,y o ,z o ] is the coordinate in the original coordinate system, R is the 3×3 rotation matrix, and T is the 3×1 translation vector.

[0066] According to the corrected depth data, the ROI of the detection area is set to perform pin height detection, solder joint size detection, solder joint centroid height detection and solder joint tip detection. Specific detection methods include:

[0067] S41: Pin height detection, which is used to determine the pin height by comparing the threshold of the pin height pixel statistics. The specific detection method is as follows:

[0068] (1) Find the average height of multiple areas in the area without components on the reference surface after plane correction as the reference height h;

[0069] (2) Set the ROI area for pin detection and extract the depth data within the ROI;

[0070] (3) Traverse the depth data in the ROI to determine whether the depth data h' of the current pixel position is higher than the reference height by 4 mm, that is, h'-h>4, and count the number of pixels higher than the reference height by 4 mm;

[0071] (4) Set the pixel statistics threshold. If the number of pixels in the ROI area is greater than the threshold, the pin height meets the requirements and is detected as OK; otherwise, the pin height is too high and is detected as NG.

[0072] S42 solder joint size detection, by comparing the solder joint area threshold to achieve uneven distribution of solder joints, detection of solder joints with too much or too little solder, continuous welding and other defects detection, the specific detection method is:

[0073] (1) Find the average height of multiple areas in the area without components on the reference surface after plane correction as the reference height h;

[0074] (2) Set the ROI area for solder joint size detection and extract the depth data within the ROI;

[0075] (3) There will be flux around the PCB substrate around the effective area of ​​solder joint detection. In order to reduce the impact of flux, the depth starting threshold t of solder joint detection is set, and the depth data within the ROI is extracted. The depth greater than h+t is the valid depth data, and the rest is set as invalid depth data (in the embodiment, t is 0.15mm, and the invalid depth data is -9999mm).

[0076] (4) Set the high and low double thresholds of the depth data, binarize the depth data within the ROI, and then perform morphological closing operations on the binarized image to fill the empty parts.

[0077] (5) Then, BLOB connected domain analysis is performed to obtain the area of ​​BLOB partitions;

[0078] (6) Set the upper and lower thresholds of the solder joint BLOB area. Determine whether the BLOB area is between the upper and lower thresholds. If it is between the upper and lower thresholds, the solder joint size meets the detection criteria and is OK; otherwise, the solder joint is too large or too small and is detected as NG.

[0079] S43: Solder point centroid height detection, by comparing the solder point centroid height threshold to detect defects such as cold solder joints and excessive solder joint height. The specific detection method is as follows:

[0080] (1) Find the average height of multiple areas in the area without components on the reference surface after plane correction as the reference height h;

[0081] (2) Set the ROI area for solder joint centroid height detection and extract the depth data within the ROI;

[0082] (3) There will be flux around the PCB substrate around the effective area of ​​solder joint detection. In order to reduce the impact of flux, the depth starting threshold t of solder joint detection is set, and the depth data within the ROI is extracted. The depth greater than h+t is the valid depth data, and the rest is set as invalid depth data (in the embodiment, t is 0.15mm, and the invalid depth data is -9999mm).

[0083] (4) Set high and low double thresholds for depth data, perform binarization on the depth data within the ROI, then perform BLOB connected domain analysis and output the mask image of the BLOB partition;

[0084] (5) Multiply the depth data in the ROI by the corresponding pixel of the mask image. The depth in the middle of the threshold range in the ROI remains unchanged, and the rest is invalid depth data. In addition, the depth of the surrounding pixel positions is used for the hole part to perform bilinear interpolation to calculate the centroid height h' of the valid depth data in the ROI, that is, the depth average of all valid depth data;

[0085] (6) Set the upper and lower thresholds for solder joint height detection. Determine whether the centroid height h' of the depth data is between the upper and lower thresholds. If it is between the upper and lower thresholds, the solder joint height meets the detection criteria and is OK; otherwise, the solder joint height is too high or too low and is detected as NG.

[0086] S44: Pull tip detection, by comparing the area threshold to achieve the detection of defects such as weld tip and welding slag around the weld. The specific detection method is:

[0087] (1) Find the average height of multiple areas in the area without components on the reference surface after plane correction as the reference height h;

[0088] (2) Set the ROI area for solder joint tip detection and extract the depth data within the ROI;

[0089] (3) There will be solder flux around the PCB substrate around the effective area of ​​solder joint detection. In order to reduce the influence of solder flux, the depth starting threshold t of solder joint detection is set, and the depth data within the ROI greater than the depth of h+t is extracted as valid depth data, and the rest is set as invalid depth data (in the embodiment, t is 0.3 mm, and the invalid depth data is -9999 mm);

[0090] (4) Normalize the depth data in the ROI into a grayscale image, use the Sobel operator to calculate the gradients in the horizontal and vertical directions respectively, obtain the normal vector by calculating the cross product of the gradient vector, and then calculate the angle between the normal vector and the Z axis, and filter out the pixels whose angle is greater than the set parameter value (the intersection angle parameter is set to 30° in the embodiment);

[0091] (5) Set high and low double thresholds for depth data, perform binarization on the depth data after normal vector filtering, then perform BLOB connected domain analysis and output the area of ​​BLOB partitions;

[0092] (6) Set the upper and lower thresholds of the BLOB area. Determine whether the BLOB area is between the upper and lower thresholds. If it is between the upper and lower thresholds, the solder joint tip detection is NG; otherwise, the solder joint tip detection is OK.

[0093] like Figure 2, a system for defect detection of pin solder joints of plug-in components, comprising: a data acquisition module 301, a data preprocessing module 302, a defect detection module 303, and a result execution module 304;

[0094] Data acquisition module 301, used to collect the depth and stereoscopic data of the inspection material

[0095] Data preprocessing module 302, used to achieve high-precision template matching and plane correction of images;

[0096] The defect detection module 303 is used to detect defects such as pin height, excessive or insufficient solder volume on solder joints, cold solder joints, continuous solder joints, and pointed solder joints, and includes: a pin height detection module 3031, a solder joint size detection module 3032, a solder joint centroid height detection module 3033, and a pointed solder joint detection module 3034.

[0097] The pin height detection module 3031 is used to detect the pin height;

[0098] Solder joint size detection 3032, used to detect defects such as uneven distribution of solder joints, excessive or insufficient tin in solder joints, etc.

[0099] The solder joint centroid height detection module 3033 is used to detect defects such as cold solder joints and excessively high solder joints;

[0100] The pull tip detection module 3034 is used to detect defects such as pull tips of weld points and welding slag around weld points.

[0101] The result execution module 304 is used to flow NG products and OK products into different processes respectively.

[0102] like Figure 3 A defect detection device for the solder joints of the pins of plug-in components includes: a counter 401 chassis, an industrial computer 402, a programmable logic controller Programmable Logic Controller, PLC403, a material conveying part 404, a material tray 405, a detection material 406, a single-axis translation stage 407, a connecting sheet metal 408, a 3D line laser camera 409, a manipulator 410, and a NG product temporary storage table 411.

[0103] The above-mentioned device is installed on the counter chassis 401 as a whole. The lower end of the counter chassis 401 contains an industrial computer 402 and a PLC 403. The industrial computer 402 is installed with an executable program to realize the defect detection of the pin solder joints of the plug-in components; at the same time, the industrial computer 402 and the PLC 403 can realize corresponding communication interaction. The test material 406 is placed on the material tray 405. When the material tray 405 and the test material 406 are transported to the laser line scanning solder joint detection process by the material conveying part 404. The 3D camera 409 is fixed on the single-axis displacement stage 407 through the connecting sheet metal 408. When the laser line scanning solder joint detection process is started, the single-axis displacement stage 407 will be triggered to start displacement. After the displacement reaches the specified starting position, the 3D camera 409 will be triggered to collect depth stereo data. When the displacement reaches the specified predetermined position, the single-axis displacement stage 407 stops moving, and the 3D camera 409 stops collecting at the same time, and the acquisition of depth stereo data is completed. After data collection is completed, the defect detection program on the industrial computer 402 is executed. When the detection result is obtained, the industrial computer 402 will follow the communication interaction logic between the industrial computer 402 and the PLC 403 and send the detection result to the PLC 403. At the same time, the material tray 405 and the detection material 406 will be transmitted to the result sorting execution process by the material conveying unit 404. When the PLC 403 receives the result signal, it will judge the result. When the result is OK, the material tray 405 and the detection material 406 will be transmitted to the next process by the material conveying unit 404; when the result is NG, it will send an instruction to the robot 410 to grab the detection material 406 to the NG product temporary storage table 411.

[0104] In the embodiment, a defect detection system for the pin solder joints of plug-in components further includes the need to complete the interaction through communication between the engineering machine PC and the PLC installed with the defect detection software, and the interaction logic is:

[0105] PC—>PLC sends a heartbeat signal: PC sends 0 and 1 to PLC at intervals of 0.5s, and PLC monitors in real time. If it hears the heartbeat signal, it means the communication is normal.

[0106] PLC—>PC sends trigger signal: The trigger signal is 0 before triggering, PLC sends 1 to PC, and when PC receives 1, it will execute the corresponding detection process 1.

[0107] PC—>PLC sends result signal: After the process is executed, the PC sends a result signal 1 / 2 to the PLC (1 represents OK, 2 represents NG).

[0108] PC—>PLC sends completion signal: the result signal is sent, and the PC sends a completion signal 1 to the PLC (indicating that the detection result is sent).

[0109] PC—>PLC sends result signal 0, completion signal 0: Set the result signal and the sending signal to 0, indicating that the execution is completed and the next trigger can be performed.

[0110] Figure 4 FIG. 1 is a flowchart of a defect detection device for a plug-in component pin solder joint provided by an embodiment of the present invention. Figure 4 ,include:

[0111] S210: After the entire production system is powered on, the 3D camera is powered on, and the industrial computer automatically starts and runs the executable program and software of the defect detection method for the pin solder joints of the plug-in components. The detection process starts automatically, automatically turns on the camera and configures the camera parameters. At the same time, the industrial computer detection software sends a heartbeat signal to the PLC in real time (sending 0 and 1 to the PLC at intervals of 0.5s);

[0112] S220: When the production line is fed with materials and the materials that have completed the welding process flow to the laser line scanning welding point detection process, the PLC will send a trigger signal to the industrial computer software;

[0113] S230: When the industrial computer software monitors the trigger signal, it will trigger the single-axis translation stage to move. The single-axis translation stage is in the initial position. When receiving the trigger signal sent by the PLC, the single-axis translation stage will start to move to the predetermined position.

[0114] S240: When the single-axis translation stage moves to the predetermined position, it will stop at the initial position for 0.5 seconds to eliminate the impact of vibration and then return to the initial position. When returning from the initial position to the predetermined position, the camera will be triggered to collect a frame of depth stereo data, and the movement of the translation stage will realize the camera scanning.

[0115] S250: When the single-axis positioning stage returns to the predetermined position, the camera stops collecting data and transmits the depth stereo data to the industrial computer. The industrial computer starts the detection process of the detection software.

[0116] S260: When the inspection process is completed, the inspection result OK-1 / NG-2 will be obtained, and the industrial computer software will send the result to the PLC; at the same time, the material will flow to the result sorting execution process.

[0117] S270: After receiving the result signal, the PLC will judge the result. If the result is OK, the material will be transferred to the next process; if the result is NG, a command will be sent to the robot to grab the material to the NG product temporary storage table.

[0118] When the new material flows to the laser line scanning welding point detection process, the above steps S220-S270 are continued to be performed.

[0119] In the workflow steps S220-S270, if a heartbeat interruption occurs, an alarm will be triggered and detection will be stopped.

[0120] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for detecting defects in solder joints of plug-in components, characterized in that: It includes the following steps: Collect deep stereo data of the inspected materials; Perform high-precision template matching based on the brightness map of the depth stereo data to determine the relative position of the inspection material; Use the PCB board as the reference surface for plane correction; According to the corrected depth data, the ROI of the inspection area is set to perform pin height inspection, solder joint size inspection, solder joint centroid height inspection and solder joint tip inspection respectively; Result judgment.

2. A method for defect detection of solder joints of plug-in components according to claim 1, characterized in that: The specific implementation of collecting the depth stereo data of the detected material is as follows: S11: Use a 3D line laser camera to obtain profile data; S12: splicing the contour data to obtain the depth stereo data of the detected material, wherein the depth stereo data includes two types of data: a depth map and a brightness map. The depth stereo data needs to include at least two features common to the detected material to facilitate subsequent matching and positioning.

3. The method for defect detection of solder joints of plug-in components according to claim 1, characterized in that: Based on the brightness map of the depth stereo data, high-precision template matching is performed to determine the relative position of the detected material. The specific implementation is as follows: S21: Selecting a brightness image of the depth stereo data of the initially detected material as a template image; S22: manually draw an ROI in the common feature area of ​​the inspected material, and extract edge feature information in the ROI feature area based on a contour extraction operator; S23: preprocess and optimize edge feature information to shield edge information with more noise; S24: performing template matching on the brightness image of the material to be inspected according to the optimized edge feature information to obtain relative position information of the material to be inspected.

4. The method for defect detection of solder joints of plug-in components according to claim 1, characterized in that: The specific implementation of plane correction using the PCB board as the reference plane is as follows: S31: Based on the matched position information, position correction is performed to correct the deviation of the detected material in motion: assisting precise positioning, and establishing a reference for position deviation according to the matching points and angles in the template matching results; S32: Then, according to the relative position offset of the preset calculation point in the feature matching result, the coordinate rotation offset of the ROI detection frame is realized, so that the ROI area can keep up with the changes in the image angle and pixels; S33: selecting no less than three feature areas (which may be in the shape of rectangle, circle, point, etc.) on the area without components of the PCB board, and performing plane fitting on the depth data in the feature areas based on the least square method; S34: Create a coordinate system based on the point-plane method. Select a point on the PCB board as the origin. Use the normal vector direction n = (a, b, -1) or n = (a, b, 1) (depending on the orientation of the fitting equation) of the fitted plane as the Z axis. Use two mutually perpendicular vectors on the plane as the X axis and Y axis to construct a transformation matrix from the original coordinate system (O-XoYoZo) to the new coordinate system (O-XYZ). The transformation matrix form is: Among them, [x, y, z] are the coordinates in the new coordinate system, [x o ,y o ,z o ] is the coordinate in the original coordinate system, R is the 3×3 rotation matrix, and T is the 3×1 translation vector.

5. The method for defect detection of solder joints of plug-in components according to claim 1, characterized in that: According to the corrected depth data, the ROI of the detection area is set, and the specific implementation of pin height detection, solder joint size detection, solder joint centroid height detection and solder joint tip detection is as follows: S41: Pin height detection, which is used to determine the pin height by comparing the threshold of the pin height pixel statistics; S42: solder joint size detection, by comparing the solder joint area threshold to achieve uneven distribution of solder joints, detection of solder joints with too much or too little solder, continuous soldering and other defects; S43: Solder point centroid height detection, by comparing the solder point centroid height threshold to detect defects such as cold solder joints and excessively high solder joints; S44: Tip detection, by comparing the area threshold to detect defects such as solder tip and welding slag around solder joints.

6. A system for defect detection of solder joints of plug-in components, characterized in that: The system includes: a data acquisition module, a data preprocessing module, a defect detection module, and a result execution module; The data acquisition module is used to collect the depth and stereoscopic data of the detected material; The data preprocessing module is used to achieve high-precision template matching and plane correction of images; The defect detection module is used to detect defects such as pin height, excessive or insufficient soldering of solder joints, cold solder joints, continuous solder joints, and pointed solder joints; The result execution module is used to flow NG products and OK products into different processes respectively.

7. The device for defect detection method of solder joints of plug-in components according to claim 6, characterized in that: The defect detection module includes: a pin height detection module, a solder joint size detection module, a solder joint centroid height detection module, and a pull tip detection module.

8. A defect detection device for solder joints of plug-in components, characterized in that: The device is used to implement a method for defect detection of solder joints of pins of plug-in components as claimed in any one of claims 1 to 5.

9. A defect detection device for solder joints of plug-in components according to claim 8, characterized in that: It also includes counter chassis, industrial computer, programmable logic controller, material conveying unit, material tray, detection material, single-axis translation table, connecting sheet metal, 3D line laser camera, robot, and 4NG product temporary storage table.

Images