Multi-type circuit board intelligent detection system

Abstract

A multi-type circuit board intelligent detection system disclosed by the present invention comprises a hardware platform and a software module, the hardware platform comprises a circuit board grabbing mechanism, and the circuit board grabbing mechanism grabs a circuit board to a posture adjusting mechanism; then, an image acquisition assembly and a laser scanning assembly are adopted to carry out two-dimensional image acquisition and circuit board top contour scanning respectively; the software module comprises a three-dimensional reconstruction unit, a feature acquisition unit, a feature parameter processing unit, a state analysis and evaluation unit and a risk management and control unit. According to the method, a non-contact high-precision detection mode is adopted, point cloud data are collected through the laser scanner, rendering is carried out in combination with a two-dimensional image, a three-dimensional contour model is obtained, indexes can be accurately quantified, defects existing on the circuit board can be efficiently and rapidly detected through judgment of multiple defects, data are objective and real, and the detection precision is high. And a multi-dimensional judgment mode also ensures the defect detection rate, and meanwhile, the risk management and control unit is used for carrying out initial detection and redetection verification, so that the misjudgment rate is further reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of circuit board detection, and in particular to an intelligent detection system for multiple types of circuit boards. Background Art

[0002] As electronic devices become increasingly miniaturized and complex, the circuits and components arranged on circuit boards are becoming increasingly dense, and the requirements for their production processes are becoming increasingly stringent in order to meet the requirements of high performance, high reliability, and small size and lightness. During the production process of circuit boards, quality control mainly includes component status and welding quality inspection; Traditional quality inspection methods rely primarily on electrical probes (such as multimeter probes) and manual visual inspection. Visual inspection for PCB defects is not only inefficient, but also subject to long-term, intensive eye strain, which can affect the inspector's visual health. Furthermore, contact-based inspection requires physical contact between the probe and the components on the soldered PCB, applying a current. This method requires specific template production and debugging for different PCBs, and can only detect PCB defects for specific tasks. Furthermore, the equipment is expensive, leading to high testing costs. Furthermore, contact-based inspection requires direct contact with the PCB, which can easily cause secondary damage and increase testing costs. To address this, we provide an intelligent inspection system for multiple types of PCBs. Summary of the Invention

[0003] The object of the present invention is to provide an intelligent detection system for multiple types of circuit boards to solve the above problems.

[0004] The present invention can be implemented through the following technical solutions: A multi-type circuit board intelligent detection system, including a hardware platform and software modules: The hardware platform includes a circuit board grabbing mechanism, which grabs the circuit board to the posture adjustment mechanism for loading and posture adjustment. Then, an image acquisition component and a laser scanning component fixed directly above the posture adjustment mechanism are used to collect two-dimensional images and scan the top contour of the circuit board respectively. The software module includes a 3D reconstruction unit, a feature acquisition unit, a feature parameter processing unit, a state analysis and evaluation unit, and a risk management unit. The 3D reconstruction unit performs downsampling and supplementary scanning based on the point cloud data acquired by the laser scanning component to generate a 3D contour model. The feature acquisition unit extracts the projection coordinates, pin height, solder joint contour and stacking height features of the components based on the three-dimensional contour model. Then the feature parameter processing unit calculates the overlap ratio, pin coplanarity and solder joint roundness of the components. The state analysis and evaluation unit determines the short circuit, component offset, empty soldering risk and solder joint defects based on the feature parameters, and generates a detection report. At the same time, the risk control unit constructs a parameter timing curve based on historical detection edging, and issues early warnings for curve indicators with a trend of breaking the threshold.

[0005] A further technical improvement of the present invention is that the 3D reconstruction unit performs downsampling in the following manner: Divide the cube surrounded by the point cloud data into a number of voxels; Calculate the centroid coordinates of all points within each voxel; The point closest to the centroid within the voxel is selected as the feature point and the other points are deleted.

[0006] A further technical improvement of the present invention is that the method of the 3D reconstruction unit identifying the fault region and performing a supplementary scan on the fault region includes: Project the point cloud onto the xy plane, i.e. the horizontal plane, and compare the original depth differences of adjacent projected points; The points where the depth difference exceeds the threshold are marked as upper / lower fault points; Connect adjacent fault points to generate upper / lower fault baselines and enclose the fault area; Obtaining the normal vector of the center point of the fault area, and performing supplementary scanning with the direction of the normal vector as the target direction of the laser scanning; The point cloud data obtained by the supplementary scan is downsampled and the coordinates are changed, unified into the coordinate system of the initial scan, and the complete point cloud data is obtained, and the three-dimensional contour model is fitted.

[0007] A further technical improvement of the present invention is that the step of rendering the three-dimensional contour model based on the two-dimensional image by the three-dimensional reconstruction unit includes: Scaling the 3D silhouette model so that its top-view projection matches the size of the 2D image; Map each pixel setting of the two-dimensional image to the coordinate area corresponding to the three-dimensional model to complete the rendering.

[0008] A further technical improvement of the present invention is that the overlap ratio of components is obtained by respectively taking the projection of the three-dimensional contour model and the projection of a component in the standard welding state on the horizontal plane, obtaining the overlapping area of ​​the two projections, and the ratio of the area of ​​the overlapping area to the area of ​​the projection area in the standard welding state is the overlap ratio; The coplanarity of the solder pins is the height variance of the same selected points of each solder pin on the component; The roundness of the solder joint is constructed based on the roundness of the edge profile of the solder joint projected on the horizontal plane and the curvature of the cross-sectional profile curve of the solder joint in the height direction.

[0009] A further technical improvement of the present invention is that the calculation formula for the roundness of the solder joint is: ; in, Represents the coverage area of ​​the edge contour projected by the point on the horizontal plane, Represents the perimeter of the edge contour projected on the horizontal plane, and the circularity is ; It represents the mean curvature of the points evenly selected on the cross-section profile curve. Indicates the calibrated ideal curvature value; and Respectively represent the circularity influence weight and curvature influence weight, .

[0010] A further technical improvement of the present invention is that the specific logic of the state analysis and evaluation unit in evaluating the welding quality includes: If the projection contours of two solder joints on the horizontal plane have at least one intersection, the solder joint is judged to be short-circuited; If the component overlap ratio is lower than the set threshold, the component is determined to be offset; If the pin coplanarity of all solder pins of a component exceeds the set threshold, a risk of empty soldering is marked; If the roundness of the solder joint is lower than the set threshold, it is marked as a solder joint defect.

[0011] A further technical improvement of the present invention is that when the state analysis and evaluation unit detects a corresponding defect, it generates a grabbing instruction to grab the circuit board to the upstream of the detection system for re-inspection; and when the re-inspection conclusions are the same, the detection report is output; when the re-inspection conclusions are different, it generates a grabbing instruction to grab the circuit board to the circuit board cache area for manual verification.

[0012] A further technical improvement of the present invention is that the risk control unit regularly obtains inspection reports of the same type of circuit boards, and constructs a timing change curve based on the overlap ratio of each component, the coplanarity of the welding pins, and the roundness of the corresponding solder joints. When the curve within a period of time from the current moment gradually shifts towards the set threshold, a risk control feedback alarm is generated to the alarm unit.

[0013] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention adopts a non-contact high-precision detection method to avoid physical damage. It collects point cloud data through a micron-level line laser scanner and combines it with the two-dimensional image obtained by the CCD camera for rendering to obtain a three-dimensional contour model, so that quantitative indicators such as overlap, pin coplanarity and solder joint roundness can be accurately extracted. Through short circuit judgment, offset judgment, empty solder risk judgment and solder joint defect judgment, defects on the circuit board can be detected efficiently and quickly. Moreover, the data is objective and true, and the multi-dimensional judgment method also ensures the defect detection rate. At the same time, the risk control unit is used for initial inspection and re-inspection verification, which further reduces the probability of false detection and missed detection.

[0014] 2. When constructing a three-dimensional contour model, the present invention adopts a Zigzag path for scanning and combines it with a supplementary scanning mechanism, which greatly improves the detection efficiency and model accuracy; the use of voxel downsampling and dynamic identification and correction algorithms of fault areas can improve model accuracy and make up for the low contour accuracy and tight computing power and storage resources caused by the existing three-dimensional model construction method.

[0015] 3. The present invention uses a risk control unit to analyze historical data trends based on historical inspection forms, so as to pay attention to the data deviation direction and trend of the same type of products during the inspection process, so as to provide early warning of process deviations of the production line and control quality from the source. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0017] Figure 1 Schematic diagram of the external structure of the present invention; DETAILED DESCRIPTION

[0018] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0019] See also Figure 1 As shown, a multi-type circuit board intelligent detection system includes a hardware platform and software modules, wherein the hardware platform includes a posture adjustment mechanism, a circuit board grasping mechanism, an image acquisition component, and a laser scanning component; the software module includes a three-dimensional reconstruction unit, a feature acquisition unit, a data storage unit, a feature parameter processing unit, a state analysis and evaluation unit, a risk control unit, and an alarm unit; During operation, the circuit board to be inspected is placed on the carrier of the posture adjustment mechanism by the circuit board grabbing mechanism and fixed in position. Then, the image acquisition component captures the image of the circuit board from the upper position. After that, the laser scanning component is driven to scan the surface of the circuit board in a Zigzag scanning mode, and the point coordinate data obtained by the scan conversion is output to the data cache. It should be noted that the posture adjustment mechanism is specifically a multi-degree-of-freedom motion platform as shown in CN102862929A, the circuit board grasping mechanism is a multi-axis robotic arm equipped with a vacuum end pick-up that matches the circuit board, the image acquisition component is an industrial CCD camera fixed at a certain height directly above the posture adjustment mechanism, and the laser scanning component includes a micron-level line laser scanner fixed at the output end of the three-axis displacement platform.

[0020] During the scanning process of the laser scanning component, the 3D reconstruction unit obtains the point cloud coordinates from the data cache ,Since the amount of point cloud coordinate data obtained is very large, targeted downsampling is required to reduce computing power consumption; The specific methods of downsampling by the 3D reconstruction unit include: For all point cloud data, get the maximum and minimum values ​​of the three-dimensional coordinates, that is, and , then the length, width and height are cube; According to the set pixel size (such as ) Divide the above cube into a large number of voxels of small cube structures of the same size, traverse each voxel, and calculate the centroid coordinates of all points in each voxel ,Right now: ;in Indicates the number of points in the corresponding voxel; Then the distance between each point in the voxel and the centroid of the voxel is calculated, and the point closest to the centroid in the voxel is found. This point is used as the feature point representing the voxel, and other points in the voxel are deleted. Compared with directly using the centroid as the feature point, the original point cloud data can be retained to the greatest extent.

[0021] The 3D reconstruction unit obtains any three non-collinear points in all the point cloud data, and the depth data (ordinate) of these three points are consistent. Then, an xy plane is constructed based on these three points. Then, all the point clouds are projected onto the xy plane area. Within a certain radius of each projection point, if there are other points, the depth data between the original points of the two projection points is compared. When the depth difference between the two exceeds the set threshold, the two points with different depth data in the two original points are marked as upper and lower fault points. It should be noted that only one comparison is performed between any two points. After all points have been compared, the two nearest adjacent depth fault points are connected based on the marked upper or lower fault point. It should be noted that if the distance between the two nearest points exceeds the set threshold, the two points are not connected, thereby obtaining at least one pair of corresponding upper and lower fault baselines; the fault area is obtained by connecting the adjacent endpoints of the upper and lower fault baselines; Obtain the normal vector of the center point of the fault area as the direction for the laser scanner to perform supplementary scanning. Drive the posture adjustment mechanism to drive the circuit board to adjust its posture so that the laser incident direction of the laser scanner is parallel to the direction of the normal vector. Drive the laser scanning component to continue scanning the fault area using the Zigzag scanning method to obtain point cloud data of the fault area. The point cloud data of the area is downsampled in the same way, and then the coordinate system of this supplementary scan is converted to the coordinate system of the laser scanner during the first scan; the converted point cloud coordinates are added to the first downsampled point cloud data to obtain a 3D contour model; Furthermore, the 3D reconstruction unit combines the front 2D image acquired by the CCD camera and the 3D contour model to render the 3D contour. Specifically: The 3D contour model is scaled until the size of the top projection of the 3D contour matches the size of the 2D image. At this point, the 2D image is directly used as the top projection of the 3D contour model. Obtain the color value of each pixel on the two-dimensional image and use it as the color value of the corresponding horizontal and vertical coordinate position of the three-dimensional contour model, so as to use the color value to render the entity at the horizontal and vertical coordinate position, traverse all the pixels of the two-dimensional image, and complete the color rendering of the three-dimensional contour model, that is, the three-dimensional rendering model.

[0022] The feature acquisition unit acquires the quality feature parameters of components, pins, and solder joints based on the 3D rendering model. Specifically: Before testing, the data storage unit pre-stores the specification parameters of various types of circuit boards involved in the test. The specification parameters include component type, installation position, component pin number and corresponding soldering position, as well as the solder joint status requirements for matching pin size. The solder joint status includes solder joint roundness, coverage area and stacking height. Select components based on the shape of the 3D rendering model, number each component, and identify and extract the corresponding component's outer contour projection coordinates, solder pin numbers, the height of the same selected point on the solder pin, the stacking height of the solder joints, the edge contour of the solder joint projected on the horizontal plane, and any cross-sectional contour of the solder joint along the height direction; The feature acquisition unit binds the feature parameters of each component to the corresponding component and organizes them into a data form and outputs them to the feature parameter processing unit. The feature parameter processing unit calculates and processes the features of each component according to the data form to construct more representative status indicator parameters. The specific methods include: For a certain component, the corresponding component type and installation position are obtained from the data storage unit. According to the installation position and the size of the corresponding component type, the projection area formed by the component on the circuit board surface under the standard welding state can be simulated, and the overlap area is superimposed with the area enclosed by the outer contour projection coordinates of the actual component to obtain the overlap area. The ratio of the overlap area area to the projection area area under the standard welding state is calculated and defined as the overlap ratio. ; For the solder pins on the component, calculate the variance of the height of the solder pins with the same selection, and use the variance to define the coplanarity of the pins , the larger the variance value, the smaller the coplanarity.

[0023] For the solder joint of the pin, the coverage area is calculated based on the edge contour of the solder joint projected on the horizontal plane and perimeter , then the circularity is Based on the cross-sectional profile of the weld in the height direction, the corresponding contour curve is obtained. Points are evenly taken on both sides of the highest point of the contour curve, the curvature of these points is calculated, and the average curvature is calculated as the final curvature. .

[0024] The state analysis and evaluation unit analyzes and evaluates the soldering quality of components of the circuit board based on the state index parameters constructed by the characteristic parameter processing unit. Specifically: When the contour lines of the two solder joints projected on the horizontal plane have at least one intersection, it means that there is a bridging phenomenon between the two solder joints, causing a short circuit, so it is determined that there is a short circuit between the two solder joints; When the overlap ratio is less than the set threshold, it means that there is a displacement between the current component and the standard welding state, and the displacement amplitude exceeds the allowable displacement range, and the component offset is recorded; When the coplanarity of the solder pins of all pins in a component exceeds the set threshold, it indicates that the solder pins of the component are warped and there is a possibility of empty soldering, and the component empty soldering risk is recorded; Calculate the roundness of the corresponding solder joint based on the roundness and final curvature of the solder joint , the calculation formula is: ; in, and Respectively represent the circularity influence weight and curvature influence weight, , Indicates the calibrated ideal curvature value; When the calculated roundness is less than the set threshold, it indicates that the corresponding solder joint is in poor condition, and there may be problems with insufficient solder, redundancy, or wrapping. The corresponding solder joint status defect of the solder joint pin is recorded.

[0025] The status analysis and evaluation unit stores the evaluation results of each component in the data storage unit and compiles the detection report; after the status analysis and evaluation unit completes the detection of each circuit board, when there is at least one record of solder joint short circuit, component offset, empty solder risk, and poor solder joint status, a re-inspection instruction is generated to the circuit board grabbing mechanism, driving it to grab the circuit board and reflow it to the upstream of the detection system for re-inspection; if the conclusion of the re-inspection is consistent with that of the initial inspection, the detection report is output to the display unit and / or printer for display; if the conclusion of the re-inspection is inconsistent with that of the initial inspection, the status analysis and evaluation unit generates a grabbing instruction to grab the circuit board to the circuit board buffer area for manual inspection.

[0026] The risk control unit regularly obtains inspection reports of the same type of circuit boards in the data storage unit, and constructs a timing change curve based on the overlap ratio of each component, the coplanarity of the welding pins, and the roundness of the corresponding solder joints, and observes its changing trend. When the curve within a period of time from the current moment gradually shifts towards the set threshold, a risk control feedback alarm is generated to the alarm unit to prompt the production department to inspect and maintain the corresponding process.

[0027] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. An intelligent detection system for multiple types of circuit boards, characterized by: Including hardware platform and software modules: The hardware platform includes a circuit board grabbing mechanism, which grabs the circuit board to the posture adjustment mechanism for loading and posture adjustment. Then, an image acquisition component and a laser scanning component fixed directly above the posture adjustment mechanism are used to collect two-dimensional images and scan the top contour of the circuit board respectively. The software module includes a 3D reconstruction unit, a feature acquisition unit, a feature parameter processing unit, a state analysis and evaluation unit, and a risk management unit. The 3D reconstruction unit performs downsampling and supplementary scanning based on the point cloud data acquired by the laser scanning component to generate a 3D contour model. The feature acquisition unit extracts the projection coordinates, pin height, solder joint contour and stacking height features of the components based on the three-dimensional contour model. Then the feature parameter processing unit calculates the overlap ratio, pin coplanarity and solder joint roundness of the components. The state analysis and evaluation unit determines the short circuit, component offset, empty soldering risk and solder joint defects based on the feature parameters, and generates a detection report; at the same time, the risk control unit constructs a parameter timing curve based on historical detection edging, and issues early warning for curve indicators with a trend of breaking the threshold.

2. The intelligent detection system for multiple types of circuit boards according to claim 1, characterized in that: The way in which the 3D reconstruction unit performs downsampling includes: Divide the cube surrounded by the point cloud data into a number of voxels; Calculate the centroid coordinates of all points within each voxel; The point closest to the centroid within the voxel is selected as the feature point and the other points are deleted.

3. The intelligent detection system for multiple types of circuit boards according to claim 1, characterized in that: The method of the three-dimensional reconstruction unit identifying the fault region and performing a supplementary scan on the fault region includes: Project the point cloud onto the xy plane, i.e. the horizontal plane, and compare the original depth differences of adjacent projected points; The points where the depth difference exceeds the threshold are marked as upper / lower fault points; Connect adjacent fault points to generate upper / lower fault baselines and enclose the fault area; Obtaining the normal vector of the center point of the fault area, and performing supplementary scanning with the direction of the normal vector as the target direction of the laser scanning; The point cloud data obtained by the supplementary scan is downsampled and the coordinates are changed, unified into the coordinate system of the initial scan, and the complete point cloud data is obtained, and the three-dimensional contour model is fitted.

4. The intelligent detection system for multiple types of circuit boards according to claim 1, characterized in that: The step of rendering the three-dimensional contour model based on the two-dimensional image by the three-dimensional reconstruction unit includes: Scaling the 3D silhouette model so that its top-view projection matches the size of the 2D image; Map each pixel setting of the two-dimensional image to the coordinate area corresponding to the three-dimensional model to complete the rendering.

5. The intelligent detection system for multiple types of circuit boards according to claim 1, characterized in that: The overlap ratio of the components is obtained by respectively taking the projection of the three-dimensional contour model and the projection of a component in the standard welding state on the horizontal plane, obtaining the overlapping area of ​​the two projections, and the ratio of the area of ​​the overlapping area to the area of ​​the projection area in the standard welding state is the overlap ratio; The coplanarity of the solder pins is the height variance of the same selected points of each solder pin on the component; The roundness of the solder joint is constructed based on the roundness of the edge profile of the solder joint projected on the horizontal plane and the curvature of the cross-sectional profile curve of the solder joint in the height direction.

6. The intelligent detection system for multiple types of circuit boards according to claim 5, characterized in that: The calculation formula for the roundness of the solder joint is: ； in, Represents the coverage area of ​​the edge contour projected by the point on the horizontal plane, Represents the perimeter of the edge contour projected on the horizontal plane, and the circularity is ; It represents the mean curvature of the points evenly selected on the cross-section profile curve. Indicates the calibrated ideal curvature value; and Respectively represent the circularity influence weight and curvature influence weight, .

7. The intelligent detection system for multiple types of circuit boards according to claim 1, characterized in that: The specific logic of the state analysis and evaluation unit in evaluating welding quality includes: If the projection contours of two solder joints on the horizontal plane have at least one intersection, the solder joint is judged to be short-circuited; If the component overlap ratio is lower than the set threshold, the component is determined to be offset; If the pin coplanarity of all solder pins of a component exceeds the set threshold, a risk of empty soldering is marked; If the roundness of the solder joint is lower than the set threshold, it is marked as a solder joint defect.

8. The intelligent detection system for multiple types of circuit boards according to claim 7, characterized in that: When the state analysis and evaluation unit detects a corresponding defect, it generates a grabbing instruction to grab the circuit board to the upstream of the detection system for re-inspection; and if the re-inspection conclusions are the same, the detection report is output; if the re-inspection conclusions are different, it generates a grabbing instruction to grab the circuit board to the circuit board buffer area for manual verification.

9. The intelligent detection system for multiple types of circuit boards according to claim 1, characterized in that: The risk control unit regularly obtains inspection reports of the same type of circuit boards, and constructs a timing change curve based on the overlap ratio of each component, the coplanarity of the welding pins, and the roundness of the corresponding solder joints. When the curve within a period of time from the current moment gradually shifts towards the set threshold, a risk control feedback alarm is generated to the alarm unit.

Images