阅读说明：本技术 一种回收电路板焊接缺陷快速检定系统及方法 (Rapid verification system and method for welding defects of recovered circuit board ) 是由 魏巍 汪姗姗 袁燕 谭孝东 朱光辉 李再宝 于 2021-09-14 设计创作，主要内容包括：本发明公开了一种回收电路板焊接缺陷快速检定方法,包括：步骤一、向回收电路板的正面照射反射光用的可见光,反面照射透射光用的非可见光,并获取回收电路板的正面的图像；步骤二、对图像进行解析,得到回收电路板的焊接缺陷区域；步骤三、通过对焊接缺陷区域进行特征识别,得到回收电路板的缺陷特征参数；步骤四、根据回收电路板的类型,通过预设的算法库,选择分类算法,基于缺陷特征参数,得到分类结果,并根据分类结果进行分类处理,通过可见光反射和非可见光透射能够快速检测到回收电路板的表面及内部缺陷,进一步通过特征识别和分类算法,得到回收电路板的分类结果,进行分类处理,能够提高电路板的回收效率和利用率,节能环保。(The invention discloses a rapid detection method for welding defects of a recovered circuit board, which comprises the following steps: irradiating visible light for reflecting light to the front side of the recovery circuit board, irradiating non-visible light for transmitting light to the back side of the recovery circuit board, and acquiring an image of the front side of the recovery circuit board; analyzing the image to obtain a welding defect area of the recovered circuit board; thirdly, performing characteristic identification on the welding defect area to obtain defect characteristic parameters of the recovered circuit board; and step four, selecting a classification algorithm according to the type of the recovered circuit board through a preset algorithm library, obtaining a classification result based on the defect characteristic parameters, performing classification processing according to the classification result, rapidly detecting the surface and internal defects of the recovered circuit board through visible light reflection and non-visible light transmission, further obtaining the classification result of the recovered circuit board through characteristic identification and classification algorithms, performing classification processing, improving the recovery efficiency and the utilization rate of the circuit board, saving energy and protecting environment.)

1. A method for rapidly detecting welding defects of a recovered circuit board is characterized by comprising the following steps:

irradiating visible light for reflecting light to the front side of the recovery circuit board, irradiating non-visible light for transmitting light to the back side of the recovery circuit board, and acquiring an image of the front side of the recovery circuit board;

analyzing the image to obtain a welding defect area of the recovered circuit board;

thirdly, performing characteristic identification on the welding defect area to obtain defect characteristic parameters of the recovered circuit board;

and step four, selecting a classification algorithm through a preset algorithm library according to the type of the recovered circuit board, obtaining a classification result based on the defect characteristic parameters, and performing classification processing according to the classification result.

2. The method for rapidly detecting the welding defects of the recycled circuit board as claimed in claim 1, wherein the visible light is a white light source, and the non-visible light is a near infrared light source.

3. The method for rapidly detecting the welding defects of the recycled circuit board as claimed in claim 2, wherein the second step comprises the following steps:

carrying out bilateral filtering by adopting a 3 x 3 filtering template to eliminate noise in the image;

performing threshold segmentation on the image by adopting a maximum inter-class variance method, and outputting a binary image;

and comparing the binarized image with a standard binarized image of the circuit board, and obtaining a defect area by adopting a difference method.

4. The method for rapidly detecting the welding defects of the recycled circuit board as claimed in claim 3, wherein the defect characteristic parameters comprise morphological characteristics, color characteristics and texture characteristics.

5. The method for rapidly detecting the welding defects of the recycled circuit board as claimed in claim 4, wherein the morphological characteristics comprise:

the number of the defective points can be obtained by labeling the areas which are not communicated with each other in the welding defective area;

carrying out statistical calculation on pixels of the defect points in the cross-sectional area of the horizontal plane projection to obtain a defect area A (pi mn), wherein m represents a major semi-axis of the defect point, and n represents a minor semi-axis of the defect point;

and calculating the pixel number of the ridge line of the defect point to obtain the perimeter C of the defect point as 2 pi n +4 (m-n).

6. The method for rapidly detecting the welding defects of the recycled circuit board as claimed in claim 5, wherein the texture features comprise:

setting pixel displacement parameters (a, b), wherein a represents an angle, a belongs to (0 degrees, 45 degrees, 90 degrees and 135 degrees), and b represents a distance;

generating a gray value pair by a single pixel point and an offset point thereof;

traversing each pixel point to generate a gray level co-occurrence matrix, and calculating a characteristic value of the gray level co-occurrence matrix;

wherein the characteristic values include energy, contrast, inverse difference, and entropy.

7. The method for rapidly verifying the welding defects of the recycled circuit boards as claimed in claim 6, wherein the classification algorithm comprises a Canopy clustering algorithm, a logistic regression algorithm and an Xmeans clustering algorithm.

8. The method for rapidly detecting the welding defects of the recycled circuit board as claimed in claim 7, wherein the selecting the classification algorithm comprises:

acquiring the defect characteristic parameters and manual classification results of a plurality of recovered circuit boards of the same type, and storing the defect characteristic parameters and the manual classification results in a recovered circuit board database;

normalizing the defect characteristic parameters, and respectively inputting the Canopy clustering algorithm, the logistic regression algorithm and the Xmeans clustering algorithm to obtain a first classification result, a second classification result and a third classification result;

and matching the first classification result, the second classification result and the third classification result with the manual classification result, and screening out the classification algorithm corresponding to the classification result with the highest matching degree.

9. The method for rapidly detecting the welding defects of the recycled circuit board as claimed in claim 8, wherein the classification result comprises:

the product is excellent in type, small in repair difficulty and capable of being reused after being repaired;

the circuit board is good in type and moderate in repair difficulty, and the circuit board is repaired or directly discarded according to the type of the recovered circuit board;

the inferior type has high repairing difficulty and is directly abandoned.

10. The utility model provides a retrieve quick verification system of circuit board welding defect which characterized in that includes:

an algorithm library for storing the classification algorithms;

the recovery circuit board database is used for storing the types, defect characteristic parameters and manual classification results of the recovery circuit boards;

a visible light illumination module that irradiates visible light for reflecting light to the front surface of the recovery circuit board;

a non-visible light illumination module that irradiates non-visible light for transmitted light to a reverse surface of the recovery circuit board;

a photographing device which photographs an image of a front surface of the recycle circuit board;

the image processing module is connected with the shooting device, can analyze the image to obtain a welding defect area of the recovered circuit board, and extracts defect characteristic parameters;

the classification processing module is connected with the algorithm library and the recovered circuit board database, can select a classification algorithm through a preset algorithm library according to the type of the recovered circuit board, and obtains a classification result based on the defect characteristic parameters;

and the result output module is connected with the classification processing module and can display the classification result.

Technical Field

The invention relates to the technical field of circuit board recovery, in particular to a system and a method for quickly detecting welding defects of a recovered circuit board.

Background

With the recent trend of drawing the demand for electronic devices in the background of "internet +", and toward miniaturization, weight reduction, high speed, high reliability, and economy of electronic devices, the scale of PCB usage will still be enlarged. Particularly, for aviation and aerospace electronic devices with high reliability requirements, the reliability requirements for internally matched PCBs are high due to the harsh requirements of the application environment of the electronic devices, and if a PCB in the electronic device has a welding defect, the whole product must be scrapped or the PCB must be disassembled again.

Along with the continuous updating and replacement, the number of waste circuit boards is gradually increased, and because the waste circuit boards contain a large amount of recyclable materials, in order to reduce the waste of resources, the waste circuit boards are usually effectively recycled, the circuit boards are crushed, valuable components such as precious metals in the circuit boards are extracted, the circuit boards are incinerated, the recycled circuit boards are rarely graded and evaluated, and different treatment means are adopted, so that the resource is greatly wasted, the recycling rate is low, and the environment is polluted, therefore, a system and a method for rapidly detecting the welding defects of the recycled circuit boards are needed to be provided.

Disclosure of Invention

The invention provides a method for rapidly detecting welding defects of a recovered circuit board, which is characterized in that images of the recovered circuit board are collected based on visible light reflection and non-visible light transmission, defects of the circuit board are analyzed, defect characteristic parameters of the circuit board are further identified, classification is carried out based on the defect characteristic parameters, surface and internal defects of the recovered circuit board can be rapidly detected, and the method is rapid in classification speed and high in precision.

A rapid detection method for welding defects of a recovered circuit board comprises the following steps:

irradiating visible light for reflecting light to the front side of the recovery circuit board, irradiating non-visible light for transmitting light to the back side of the recovery circuit board, and acquiring an image of the front side of the recovery circuit board;

analyzing the image to obtain a welding defect area of the recovered circuit board;

thirdly, performing characteristic identification on the welding defect area to obtain defect characteristic parameters of the recovered circuit board;

and step four, selecting a classification algorithm through a preset algorithm library according to the type of the recovered circuit board, obtaining a classification result based on the defect characteristic parameters, and performing classification processing according to the classification result.

Preferably, the visible light is a white light source and the non-visible light is a near infrared light source.

Preferably, the second step includes:

carrying out bilateral filtering by adopting a 3 x 3 filtering template to eliminate noise in an image;

performing threshold segmentation on the image by adopting a maximum inter-class variance method, and outputting a binary image;

and comparing the binary image with the standard binary image of the circuit board, and obtaining a defect area by adopting a difference method.

Preferably, the defect feature parameters include morphological features, color features, and texture features.

Preferably, the morphological characteristics include:

the number of defect points can be obtained by labeling areas which are not communicated with each other in the welding defect area;

carrying out statistical calculation on pixels of the defect points in the cross-sectional area of the horizontal plane projection to obtain the defect area A (pi mn), wherein m represents a major semi-axis of the defect point, and n represents a minor semi-axis of the defect point;

and calculating the pixel number of the ridge line of the defect point to obtain the perimeter C of the defect point as 2 pi n +4 (m-n).

Preferably, the texture features include:

setting pixel displacement parameters (a, b), wherein a represents an angle, a belongs to (0 degrees, 45 degrees, 90 degrees and 135 degrees), and b represents a distance;

generating a gray value pair by a single pixel point and an offset point thereof;

traversing each pixel point to generate a gray level co-occurrence matrix, and calculating a characteristic value of the gray level co-occurrence matrix;

wherein the characteristic values include energy, contrast, inverse difference, and entropy.

Preferably, the classification algorithms include the Canopy clustering algorithm, the logistic regression algorithm, and the Xmeans clustering algorithm.

Preferably, selecting the classification algorithm comprises:

acquiring defect characteristic parameters and manual classification results of a plurality of recovered circuit boards of the same type, and storing the defect characteristic parameters and the manual classification results in a recovered circuit board database;

normalizing the defect characteristic parameters, and respectively inputting the defect characteristic parameters into a Canopy clustering algorithm, a logistic regression algorithm and an Xmeans clustering algorithm to obtain a first classification result, a second classification result and a third classification result;

and matching the first classification result, the second classification result and the third classification result with the manual classification result, and screening out the classification algorithm corresponding to the classification result with the highest matching degree.

Preferably, the classification result includes:

the product is excellent in type, small in repair difficulty and capable of being reused after being repaired;

the circuit board is good in type and moderate in repair difficulty, and the circuit board is repaired or directly discarded according to the type of the recovered circuit board;

the inferior type has high repairing difficulty and is directly abandoned.

A quick verification system for welding defects of a recovered circuit board comprises:

an algorithm library for storing the classification algorithms;

the recovery circuit board database is used for storing the types, defect characteristic parameters and manual classification results of the recovery circuit boards;

a visible light illumination module that irradiates visible light for reflecting light to the front surface of the recovery circuit board;

a non-visible light illumination module that illuminates non-visible light for transmitted light to the reverse side of the recovered circuit board;

a photographing device which photographs an image of the front surface of the recovered circuit board;

the image processing module is connected with the shooting device, can analyze the image to obtain a welding defect area of the recovered circuit board, and extracts defect characteristic parameters;

the classification processing module is connected with the algorithm library and the recovery circuit board database, can select a classification algorithm through a preset algorithm library according to the type of the recovery circuit board, and obtains a classification result based on the defect characteristic parameters;

and the result output module is connected with the classification processing module and can display the classification result.

The invention has the beneficial effects that:

1. the invention provides a method for rapidly detecting welding defects of a recycled circuit board, which comprises the steps of irradiating visible light on the front side of the circuit board, irradiating non-visible light on the back side of the circuit board, then carrying out image acquisition, further extracting a defect area of the circuit board, identifying defect characteristic parameters of the circuit board, selecting a classification algorithm according to the type of the circuit board, and classifying based on the defect characteristic parameters, so that the surface and internal defects of the recycled circuit board can be rapidly detected, the classification speed is high, the precision is high, the recycling efficiency can be effectively improved, the utilization rate is improved, and the method is energy-saving and environment-friendly.

2. The method establishes the algorithm library, can select the classification algorithm with high matching degree by taking the matching degree of the algorithm classification result and the manual classification result as the evaluation index according to the type of the recovered circuit board, is favorable for improving the classification precision, associates the algorithm classification with the manual classification, replaces the manual classification with the classification algorithm, improves the classification efficiency, reduces the labor cost and has high economic benefit.

3. The invention provides a rapid detection system for welding defects of a recovered circuit board, which can realize rapid identification of the defects of the recovered circuit board and classification of the recovered circuit board, and has the advantages of simple operation, stable performance and high classification precision.

Drawings

FIG. 1 is a flow chart of a method for rapidly detecting welding defects of a recycled circuit board provided by the invention.

FIG. 2 is a flow chart of image parsing in an embodiment of the invention.

Fig. 3 is a flowchart of defect feature parameter extraction according to an embodiment of the present invention.

Fig. 4 is a frame diagram of a rapid verification system for detecting a welding defect of a recycled circuit board according to the present invention.

Detailed Description

The present invention is described in terms of particular embodiments, other advantages and features of the invention will become apparent to those skilled in the art from the following disclosure, and it is to be understood that the described embodiments are merely exemplary of the invention and that it is not intended to limit the invention to the particular embodiments disclosed. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The terms "in" and the like refer to directions or positional relationships based on those shown in the drawings, which are for convenience of description only, and do not indicate or imply that a device or element must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.

Furthermore, it should be noted that, in the description of the present invention, unless otherwise explicitly specified or limited, the terms "disposed," "mounted," "connected," and "connected" are to be construed broadly and may be, for example, fixedly connected, detachably connected, or integrally connected; may be a mechanical connection; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.

After the waste circuit boards on the electronic equipment are recovered, generally, due to long-time loading, fatigue or corrosion, stains and the like generated in the disassembling process, a pretreatment is required for the verification, and the pretreatment process generally comprises the following steps: the circuit board is soaked in the cleaning agent to effectively remove surface impurities, and is dried after being washed by water, and low-temperature drying is adopted to prevent the performance of the circuit board from being damaged, reduce image acquisition noise and improve the efficiency of defect identification.

As shown in fig. 1, a method for rapidly detecting a welding defect of a recycled circuit board comprises:

s110, irradiating visible light for reflecting light to the front side of the recovered circuit board, irradiating invisible light for transmitting light to the back side of the recovered circuit board, and acquiring an image of the front side of the recovered circuit board.

The visible light is a white light source, and the invisible light is a near infrared light source.

The recovery circuit board image of single visible light irradiation collection is difficult to discern the little and inside invisible defect of surface image difference, utilizes non-visible light to transmit, can be according to the non-visible light volume that sees through the defect region and the difference of the non-visible light volume of normal region for the difference of defect region and normal region appears on gathering the image, is favorable to discerning circuit board surface and internal defect fast, and maneuverability is strong.

And S120, analyzing the image to obtain a welding defect area of the recovered circuit board.

S130, performing characteristic identification on the welding defect area to obtain defect characteristic parameters of the recycled circuit board.

The defect characteristic parameters comprise morphological characteristics, color characteristics and texture characteristics.

S140, selecting a classification algorithm through a preset algorithm library according to the type of the recovered circuit board, obtaining a classification result based on the defect characteristic parameters, and performing classification processing according to the classification result.

The classification algorithm comprises a Canopy clustering algorithm, a logistic regression algorithm and an Xmeans clustering algorithm.

The front side of the circuit board is irradiated with visible light, the back side of the circuit board is irradiated with non-visible light, image acquisition is carried out, the defect area of the circuit board is washed out, the defect characteristic parameters of the circuit board are further identified, a classification algorithm is selected according to the type of the circuit board, classification is carried out based on the defect characteristic parameters, the surface and internal defects of the recovered circuit board can be rapidly detected, the classification speed is high, the precision is high, and the recovery efficiency and the utilization rate can be effectively improved.

Further, step S120 shown in fig. 2 includes:

and S121, performing bilateral filtering by adopting a 3 x 3 filtering template to eliminate noise in the image.

The bilateral filtering is also a nonlinear filtering algorithm, not only considers the spatial proximity, removes noise like Gaussian filtering, but also considers the pixel value similarity, has the edge-preserving characteristic which the Gaussian filtering does not have, and the invention or the acquired image acquires image information under the common irradiation of visible light and transmission light, the non-visible light transmittance is an important image characteristic, and the pixel value proximity of the image is higher, so the bilateral filtering denoising is preferred.

And S122, performing threshold segmentation on the image by adopting a maximum inter-class variance method, and outputting a binary image.

The feature image and the background image can be segmented by a maximum inter-class variance method, and the feature identification efficiency is improved.

And S122, comparing the binarized image with a standard binarized image of the circuit board, and obtaining a defect area by adopting a difference method.

The standard binary image is a complete circuit board and is output after bilateral filtering denoising and maximum inter-class variance method threshold segmentation.

Fig. 3 shows a defect feature parameter extraction process according to an embodiment of the present invention.

Wherein, the morphological characteristics include:

s131, labeling areas which are not communicated with each other in the welding defect area to obtain the number of welding spots;

s132, carrying out statistical calculation on pixels of the welding spot in the cross-sectional area of the horizontal plane projection to obtain the area A of the welding spot as pi mn, wherein m represents a welding spot major-semiaxis, and n represents a welding spot minor-semiaxis;

and S133, calculating the number of pixels of the ridge line of the welding spot to obtain the circumference C of the welding spot as 2 pi n +4 (m-n).

The texture features include:

s134, setting pixel displacement parameters (a, b), wherein a represents an angle, a belongs to (0 degrees, 45 degrees, 90 degrees and 135 degrees), and b represents a distance;

s135, generating a gray value pair by the single pixel point and the offset point thereof;

and S136, traversing each pixel point to generate a gray level co-occurrence matrix, and calculating the characteristic value of the gray level co-occurrence matrix.

We can usually use some scalar to characterize the gray level co-occurrence matrix, and the following are commonly used:

(1) energy is also called angular second moment. The formula is as follows:

where G denotes a gray level co-occurrence matrix having a size of κ × κ, and (x, y) denotes pixel coordinates. As can be seen from the equation, the energy is the sum of squares of the values of the elements of the gray level co-occurrence matrix, which reflects the uniformity of the gray level distribution of the image and the thickness of the texture. If all the values of the gray level co-occurrence matrix are equal, the gray level distribution of the image is relatively even, the texture is relatively fine, and the ASM value is small; conversely, if the values are different in size, it is indicated that the gray value of the image has a drastic change in a certain region and appears as a coarse texture on the image, and the ASM value is large.

(2) And (4) contrast. The formula is as follows:

the contrast reflects the sharpness of the image and the depth of the texture grooves. If the texture grooves are deep, the texture is obvious visually, and the contrast is high; on the contrary, the contrast is small, the gray level distribution of the image is even, the grooves are shallow, and the texture is difficult to distinguish visually.

(3) The inverse difference is also called homogeneity. The formula is as follows:

homogeneity is used to measure how much the image texture changes locally. The large value of the gray scale indicates that the image has no variation among different areas of image textures, and the gray scale distribution of the image is very even. Otherwise, the image texture is indicated to be changed drastically in different areas.

(4) Entropy. The formula is as follows:

the entropy is a measure of the information content of the image, the texture information also belongs to the information of the image, and when all elements in the co-occurrence matrix have the maximum randomness, the entropy is larger, which indicates that more obvious textures appear in the original image. On the contrary, when the image distribution is more uniform, no obvious texture exists, and the entropy is smaller.

Further, selecting a classification algorithm includes:

acquiring defect characteristic parameters and manual classification results of a plurality of recovered circuit boards of the same type, and storing the defect characteristic parameters and the manual classification results in a recovered circuit board database;

normalizing the defect characteristic parameters, and respectively inputting the defect characteristic parameters into a Canopy clustering algorithm, a logistic regression algorithm and an Xmeans clustering algorithm to obtain a first classification result, a second classification result and a third classification result;

and matching the first classification result, the second classification result and the third classification result with the manual classification result, and screening out the classification algorithm corresponding to the classification result with the highest matching degree.

Further, the classification result includes:

the product is excellent in type, small in repair difficulty and capable of being reused after being repaired;

the circuit board is good in type and moderate in repair difficulty, and the circuit board is repaired or directly discarded according to the type of the recovered circuit board;

the inferior type has high repairing difficulty and is directly abandoned.

The algorithm library is established, the classification algorithm with high matching degree can be selected by taking the matching degree of the algorithm classification result and the manual classification result as an evaluation index according to the type of the recovered circuit board, the classification accuracy is favorably improved, the algorithm classification and the manual classification are associated, the classification algorithm is used for replacing the manual classification, the recovered circuit board is classified into an excellent type, a good type and a poor type, and then classification processing is carried out, so that the classification efficiency is improved, the labor cost is reduced, and the economic benefit is high.

In practical application, the poor circuit board can be crushed to extract materials such as heavy metals and the like which can be recycled from the circuit board, the excellent circuit board can be repaired and reused, and the good circuit board can be selectively crushed, recycled or repaired and reused according to the type of the circuit board, the use environment and the like.

As shown in fig. 4, a system for rapidly detecting a welding defect of a recycled circuit board includes an algorithm library 110, a recycled circuit board database 120, a visible light illumination module 210, a non-visible light illumination module 220, a camera 230, an image processing module 310, a classification processing module 320, and a result output module 330.

The algorithm library 110 is used for storing a classification algorithm, the recycled circuit board database 120 is used for storing the type, defect characteristic parameters and manual classification results of the recycled circuit board, the visible light illumination module 210 illuminates visible light for reflecting light to the front side of the recycled circuit board, the non-visible light illumination module 220 illuminates non-visible light for transmitting light to the back side of the recycled circuit board, the shooting device 230 shoots images of the front side of the recycled circuit board, the image processing module 310 is connected with the shooting device 230 and can analyze the images to obtain the welding defect area of the recycled circuit board and extract the defect characteristic parameters, the classification processing module 320 is connected with the algorithm library 110 and the recycled circuit board database 120 and can select a classification algorithm through the preset algorithm library 110 according to the type of the recycled circuit board and obtain a classification result based on the defect characteristic parameters, and the result output module 330 is connected with the classification processing module 320, the classification result can be displayed.

The rapid detection system for the welding defects of the recovered circuit board can realize rapid identification of the defects of the recovered circuit board and classification of the recovered circuit board, and is easy to realize, simple to operate, stable in performance and high in classification precision.

The above descriptions are only examples of the present invention, and common general knowledge of known specific structures, characteristics, and the like in the schemes is not described herein too much, and it is easily understood by those skilled in the art that the scope of the present invention is obviously not limited to these specific embodiments. Without departing from the invention, several changes and modifications can be made, which should also be regarded as the protection scope of the invention, and these will not affect the effect of the invention and the practicality of the patent.