A PCB board testing system

Abstract

The present invention discloses a PCB board testing system, and relates to the field of printed circuit boards. The present invention comprises: obtaining an image to be detected, and preprocessing the image to be detected; performing image recognition on the image to be detected and comparing it with a sample image, and screening out a first qualified image; using an image sensor to collect an RGB image of the first qualified image, and dividing the RGB image into a wiring picture, a solder joint picture, and a background picture; comparing the wiring picture, the solder joint picture, and the background picture with the wiring picture, the solder joint picture, and the background picture of the sample image, respectively, one by one, and if the wiring picture, the solder joint picture, and the background picture all match, then the image to be detected is qualified. The present invention flexibly uses data information to test the PCB board, saving material and manpower.

Description

Technical Field

[0001] The present invention relates to the field of printed circuit boards, and more particularly to a PCB board testing system. Background Art

[0002] At present, there are mainly the following ways to test PCB boards:

[0003] 1. Manual visual inspection of PCB board

[0004] The most traditional inspection method is to use a magnifying glass or a calibrated microscope to visually inspect the circuit board to determine whether it is qualified and when to perform correction operations. Its main advantages are low upfront cost and no test fixtures, while its main disadvantages are human subjective errors, high long-term costs, discontinuous defect detection, and difficulty in data collection. Currently, due to the increase in PCB production and the reduction in wire spacing and component size on PCB, this method is becoming increasingly unfeasible.

[0005] 2. PCB board online testing

[0006] There are several testing methods such as bed of nails tester and flying probe tester to find manufacturing defects and test analog, digital and mixed signal components to ensure they meet specifications through electrical performance testing. The main advantages are low test cost per board, strong digital and functional test capabilities, fast and thorough short and open circuit testing, programming firmware, high defect coverage and easy programming. The main disadvantages are the need for test fixtures, programming and debugging time, high cost of making fixtures, and difficulty in use.

[0007] 3. PCB board function test

[0008] Functional system testing is to use special test equipment to conduct comprehensive testing on the functional modules of the circuit board in the middle stage and at the end of the production line to confirm the quality of the circuit board. Functional testing can be said to be the earliest automatic testing principle. It is based on a specific board or a specific unit and can be completed with various equipment. There are types such as final product testing, latest physical models and stacking tests. Functional testing usually does not provide in-depth data such as pin-level and component-level diagnosis for process improvement, and requires special equipment and specially designed test processes. Writing functional test programs is complicated, so it is not suitable for most circuit board production lines.

[0009] 4. Automatic optical inspection

[0010] Also known as automatic visual inspection, it is based on optical principles and uses a variety of technologies such as image analysis, computers and automatic control to detect and process defects encountered in production. It is a newer method for confirming manufacturing defects. AOI is usually used before and after reflow and before electrical testing to improve the pass rate in the electrical processing or functional testing stage. The cost of correcting defects at this time is much lower than the cost after the final test, often reaching more than ten times.

[0011] 5. Automatic X-ray Inspection

[0012] By taking advantage of the different absorption rates of X-rays by different materials, the parts that need to be inspected can be viewed to find defects. It is mainly used to detect ultra-fine pitch and ultra-high density circuit boards and defects such as bridging, chip loss, and poor alignment generated during the assembly process. Its tomography technology can also be used to detect internal defects of IC chips. It is currently the only way to test the welding quality of ball grid arrays and obscured solder balls. The main advantages are the ability to detect BGA welding quality and embedded components without fixture costs; the main disadvantages are slow speed, high failure rate, difficulty in detecting rework solder joints, high cost, and long program development time. This is a relatively new detection method and needs further research.

[0013] 6. Laser detection system

[0014] It is the latest development in PCB testing technology. It uses a laser beam to scan the printed board, collect all measurement data, and compare the actual measurement value with the preset qualified limit value. This technology has been proven on light boards and is being considered for assembly board testing. The speed is sufficient for mass production lines. Fast output, no fixture required, and visual non-covered access are its main advantages; high initial cost, maintenance and use problems are its main disadvantages.

[0015] 7. Dimension detection

[0016] Use the two-dimensional image measuring instrument to measure the hole position, length, width, position and other dimensions. Since PCB is a small, thin and soft product, contact measurement is easy to cause deformation and inaccurate measurement. The two-dimensional image measuring instrument has become the best high-precision dimension measuring instrument. After programming, the image measuring instrument of Sirui Measurement can realize fully automatic measurement, which not only has high measurement accuracy, but also greatly shortens the measurement time and improves the measurement efficiency.

[0017] The above-mentioned various PCB board detection methods have many defects, are time-consuming and labor-intensive, have high labor costs, and the accuracy of the inspection results is not high. Summary of the invention

[0018] In view of this, the present invention provides a PCB board testing system, which flexibly utilizes data information to test the PCB board, saving material and manpower.

[0019] In order to achieve the above object, the present invention adopts the following technical solution:

[0020] A PCB board testing system, characterized by comprising the following modules:

[0021] An image preprocessing module is used to obtain an image to be detected and preprocess the image to be detected;

[0022] An image screening module, used for performing image recognition on the image to be detected and comparing it with a sample image to screen out a first qualified image;

[0023] An image segmentation module, used for collecting an RGB image of the first qualified image using an image sensor, and segmenting the RGB image into a wiring image, a solder joint image and a background image;

[0024] The image testing module is used to compare the wiring picture, the solder point picture, and the background picture with the wiring picture, the solder point picture, and the background picture of the sample image respectively. If the wiring picture, the solder point picture, and the background picture all match, the image to be tested is qualified.

[0025] Preferably, the image segmentation module also includes performing feature extraction on the first qualified image, extracting routing features, solder point features and background features, wherein the feature extraction adopts an LBP feature algorithm; the routing features, the solder point features and the background features are compared one by one with the routing features, the solder point features and the background features of the sample image, and if they all match, the image to be inspected is qualified.

[0026] Preferably, it also includes a fault marking module for marking the fault position of the PCB board.

[0027] Preferably, the image screening module adopts a semi-supervised learning algorithm and performs incremental update processing and hierarchical classification processing simultaneously.

[0028] Preferably, the image screening module adopts enhanced display technology, and utilizes multimedia, three-dimensional modeling, real-time image display and control, and scene fusion to screen images.

[0029] Preferably, the image preprocessing module includes image denoising, image enhancement, and image correction.

[0030] Preferably, the image segmentation module uses a clustering algorithm to perform segmentation, performs clustering multiple times, and selects the best clustering result.

[0031] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a PCB board testing system, which has the following beneficial effects:

[0032] The present invention can efficiently detect the fault location of the PCB board and effectively mark the fault location. At the same time, the present invention achieves a flexible detection effect through a semi-supervised learning algorithm, and the segmentation algorithm and feature extraction algorithm effectively improve the accuracy of the test. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0034] Figure 1 The accompanying drawing is a schematic diagram of a module of the present invention;

[0035] Figure 2 The accompanying drawing is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0037] The embodiment of the present invention discloses a PCB board testing system, such as Figure 1 As shown, it includes the following modules:

[0038] An image preprocessing module is used to obtain the image to be detected and preprocess the image to be detected;

[0039] An image screening module is used to perform image recognition on the image to be detected and compare it with the sample image to screen out a first qualified image;

[0040] An image segmentation module, used to collect an RGB image of the first qualified image using an image sensor, and segment the RGB image into a wiring image, a solder joint image and a background image;

[0041] The image testing module is used to compare the wiring image, solder point image, and background image with the wiring image, solder point image, and background image of the sample image one by one. If the wiring image, solder point image, and background image all match, the image to be tested is qualified.

[0042] Among them, the image segmentation module also includes feature extraction of the first qualified image, extracting the routing features, solder point features and background features, and the feature extraction adopts the LBP feature algorithm; the routing features, solder point features, and background features are compared one by one with the routing features, solder point features, and background features of the sample image. If they all match, the image to be inspected is qualified.

[0043] It also includes a fault marking module for marking the fault location of the PCB board.

[0044] Among them, the image screening module adopts a semi-supervised learning algorithm, and performs incremental update processing and hierarchical classification processing at the same time.

[0045] Among them, the image screening module adopts enhanced display technology, and uses multimedia, three-dimensional modeling, real-time image display and control, and scene fusion to screen images.

[0046] Among them, the image preprocessing module includes image denoising, image enhancement and image correction.

[0047] Among them, the image segmentation module uses a clustering algorithm for segmentation, clustering multiple times, and selects the best clustering result.

[0048] A PCB board testing method, such as Figure 2 The specific steps shown include:

[0049] Acquire the image to be detected and preprocess the image to be detected;

[0050] Performing image recognition on the image to be detected and comparing it with the sample image to select a first qualified image;

[0051] Using an image sensor to collect an RGB image of the first qualified image, and segmenting the RGB image into a wiring image, a solder joint image, and a background image;

[0052] The wiring image, solder point image, and background image are compared one by one with the wiring image, solder point image, and background image of the sample image. If the wiring image, solder point image, and background image all match, the image to be tested is qualified.

[0053] In addition, feature extraction algorithms, such as the LBP algorithm, can be used to extract routing features, solder joint features, and background features, and the extracted features can be compared with the sample features. If they all match, it means that the PCB board is qualified.

[0054] Among them, if there are defects in the PCB board, the location of the corresponding defect can also be marked.

[0055] A semi-supervised learning algorithm is used to screen the first qualified image, and incremental update processing and hierarchical classification processing are performed simultaneously. By using labels for training in both the generator and the discriminator, not only data with specific labels can be generated, but also the quality of the generated data can be improved, thereby effectively screening the images.

[0056] In addition, the first qualified image is screened by using enhanced display technology, using multimedia, three-dimensional modeling, real-time image display and control, and scene fusion to screen the image. This AR technology is more intuitive and easier to observe than the semi-supervised learning algorithm.

[0057] In this embodiment, the image to be detected is preprocessed, including image denoising, image enhancement, and image correction.

[0058] The RGB image segmentation adopts a clustering algorithm for segmentation, and clustering is performed multiple times. According to the test of the present invention, the clustering effect is best at the fifth time.

[0059] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A PCB board testing system, characterized in that: Includes the following modules: An image preprocessing module is used to obtain an image to be detected and preprocess the image to be detected; An image screening module, used for performing image recognition on the image to be detected and comparing it with a sample image to screen out a first qualified image; An image segmentation module, used for collecting an RGB image of the first qualified image using an image sensor, and segmenting the RGB image into a wiring image, a solder joint image and a background image; The image testing module is used to compare the wiring picture, the solder point picture, and the background picture with the wiring picture, the solder point picture, and the background picture of the sample image respectively. If the wiring picture, the solder point picture, and the background picture all match, the image to be tested is qualified.

2. A PCB board testing system according to claim 1, characterized in that: The image segmentation module also includes performing feature extraction on the first qualified image, extracting routing features, solder joint features and background features, wherein the feature extraction adopts an LBP feature algorithm; comparing the routing features, solder joint features and background features with the routing features, solder joint features and background features of the sample image one by one, and if they all match, the image to be inspected is qualified.

3. A PCB board testing system according to claim 1, characterized in that: It also includes a fault marking module for marking the fault location of the PCB board.

4. A PCB board testing system according to claim 1, characterized in that: The image screening module adopts a semi-supervised learning algorithm and performs incremental update processing and hierarchical classification processing at the same time.

5. A PCB board testing system according to claim 1, characterized in that: The image screening module adopts enhanced display technology, and uses multimedia, three-dimensional modeling, real-time image display and control, and scene fusion to screen images.

6. A PCB board testing system according to claim 1, characterized in that: Image preprocessing module, including image denoising, image enhancement, and image correction.

7. A PCB board testing system according to claim 1, characterized in that: The image segmentation module uses clustering algorithm for segmentation, multiple clustering, and selects the best clustering result.

Images