CN114820613B - Incoming material measuring and positioning method for SMT (surface mount technology) patch processing - Google Patents

Incoming material measuring and positioning method for SMT (surface mount technology) patch processing Download PDF

Info

Publication number
CN114820613B
CN114820613B CN202210745213.1A CN202210745213A CN114820613B CN 114820613 B CN114820613 B CN 114820613B CN 202210745213 A CN202210745213 A CN 202210745213A CN 114820613 B CN114820613 B CN 114820613B
Authority
CN
China
Prior art keywords
processing
image
smt
incoming
smt patch
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202210745213.1A
Other languages
Chinese (zh)
Other versions
CN114820613A (en
Inventor
苏平
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Ruiyi Electronic Science And Technology Co ltd
Original Assignee
Shenzhen Ruiyi Electronic Science And Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Ruiyi Electronic Science And Technology Co ltd filed Critical Shenzhen Ruiyi Electronic Science And Technology Co ltd
Priority to CN202210745213.1A priority Critical patent/CN114820613B/en
Publication of CN114820613A publication Critical patent/CN114820613A/en
Application granted granted Critical
Publication of CN114820613B publication Critical patent/CN114820613B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • H—ELECTRICITY
    • H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K13/00—Apparatus or processes specially adapted for manufacturing or adjusting assemblages of electric components
    • H05K13/08—Monitoring manufacture of assemblages
    • H—ELECTRICITY
    • H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K3/00—Apparatus or processes for manufacturing printed circuits
    • H05K3/30—Assembling printed circuits with electric components, e.g. with resistor
    • H05K3/32—Assembling printed circuits with electric components, e.g. with resistor electrically connecting electric components or wires to printed circuits
    • H05K3/34—Assembling printed circuits with electric components, e.g. with resistor electrically connecting electric components or wires to printed circuits by soldering
    • H05K3/341—Surface mounted components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30141—Printed circuit board [PCB]

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Manufacturing & Machinery (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a method for measuring and positioning incoming materials for SMT patch processing, relates to the technical field of measurement, and solves the technical problem that the existing method for measuring and positioning the incoming materials for SMT patch processing is lagged. The technical scheme is that the image information of incoming materials is obtained through an optical detection system and is transmitted to a computer for image processing; when image data information is obtained, positioning incoming materials for SMT patch processing by a three-level positioning method; the invention can realize the measurement and positioning capacity of the incoming material for SMT patch processing in an automatic measurement mode. The SMT paster processing is used for the material assembly line measuring work efficiency has been improved greatly.

Description

Incoming material measuring and positioning method for SMT (surface mount technology) patch processing
Technical Field
The invention relates to the technical field of measurement, in particular to a method for measuring and positioning incoming materials for SMT patch processing.
Background
The SMT sheet refers to a short name of a series of process flows processed on the basis of a PCB (Printed Circuit Board), which is a Printed Circuit Board. SMT is Surface mount Technology (Surface mount Technology), an acronym for Surface Mounted Technology, which is one of the most popular techniques and processes in the electronic assembly industry. Incoming material inspection before SMT paster processing is the primary condition of guaranteeing the paster quality, and the quality direct influence of components and parts, PCB board, SMT paster processing material is the paster quality of PCB board. Therefore, strict incoming material inspection and management systems are required for electrical performance parameters of components and parts, solderability of soldering terminals and pins, producibility design of PCB (printed circuit board) and solderability of bonding pads, quality of SMT (surface mount technology) chip processing materials such as soldering paste, chip paste, stick-shaped solder, welding flux, cleaning agent and the like.
In the prior art, a patch processing plant typically performs the following visual inspection,
1. and (4) visually or by using a magnifying glass to check whether the surfaces of the welding ends or the pins of the components are oxidized or have contaminants.
2. The nominal value, specification, model, precision, external dimension and the like of the components are in accordance with the technological requirements of the product.
The pins of SOT and SOIC can not be deformed, and the pin coplanarity of the multi-lead QFP device with the lead pitch of less than 0.65mm is less than 0.1mm (can be detected by a mounting machine optically).
4. The mark of the component does not fall off after being cleaned, and the performance and the reliability of the component are not influenced (visual inspection after cleaning).
Although the method improves the incoming material detection capability of SMT patch processing to a certain extent, in a large-scale production line, a large amount of incoming materials for SMT patch processing cannot be automatically detected, and the positioning capability is delayed when the SMT patches are detected.
Disclosure of Invention
Aiming at the defects of the technology, the invention discloses a measuring and positioning method for incoming materials for SMT patch processing, which can realize the measuring and positioning capacity of the incoming materials for SMT patch processing in an automatic measuring mode and improve the detection precision and the production capacity of a production line.
In order to realize the technical effects, the invention adopts the following technical scheme:
a method for measuring and positioning incoming materials for SMT patch processing comprises the following steps: placing an SMT (surface mount technology) paster processing incoming material in a production line, acquiring image information of the incoming material through an optical detection system, and transmitting the image information to a computer for image processing;
when image data information is obtained, positioning incoming materials for SMT patch processing by a three-level positioning method;
the first positioning method senses incoming material data information for SMT patch processing through a sensor, and realizes primary positioning of the incoming material information for SMT patch processing; the initial positioning reflects the type of incoming materials for SMT patch processing;
the second positioning method is that the incoming material data information screening for SMT patch processing is realized through a classifier screening method, incoming materials for SMT patch processing of the same type are released, and incoming materials for SMT patch processing of different types are blocked from flowing to the next station;
the third positioning method is to realize the screening of the incoming material single incoming material data information for SMT patch processing by a template matching method based on a metadata engine, and to block the incoming material data information with faults from flowing to the next station.
As a further technical scheme, the optical detection system comprises an image acquisition module, an image processing module, a DSP processing unit, an LED array lighting unit and an A/D conversion unit, wherein the DSP processing unit is respectively connected with the image acquisition module, the A/D conversion unit and the LED array lighting unit, the image acquisition module is connected with the image processing module, the DSP processing unit is a TMS320C6747 chip, and the image processing module comprises an improved SURF algorithm model.
As a further technical solution of the present invention, the improved SURF algorithm model includes an image feature extraction module and an accelerator connected to the image feature extraction module, wherein:
the working method for improving the SURF algorithm model comprises the following steps: the method comprises the steps of detecting incoming materials for SMT patch processing in an LED array lighting unit scale space, determining the scale and size of the incoming materials for SMT patch processing, utilizing the main inclined direction of an image neighborhood where the incoming materials for SMT patch processing are located as the directional characteristic of a target, adopting an integral image accelerator to achieve speed acceleration of convolution operation, improving image storage amount in a unit area through a mode of compressing an image format by the integral image accelerator, finally generating 12 x 12 dimensional feature point descriptors through extreme point neighborhood gradient information, and further achieving segmentation of the incoming materials for SMT patch processing through image information segmentation.
As a further technical scheme of the invention, the method for processing the incoming material for SMT patch processing by improving the SURF algorithm model comprises the following steps:
firstly, setting the scale space as G: (x，yσ) convolution to map the original SMT patch processing incoming material image to scale space G: (x，yσ) range, G: (x，ySigma) is that the SMT patch processing is used for displaying the image data information in a three-dimensional space unit, and the two-dimensional gray level processing of the obtained image is as follows:
Figure 536835DEST_PATH_IMAGE001
（1）
in the formula (1), (1)x，y) Represents the position of the pixel in the image, σ represents a scale space factor; i (a)x，y) The pixel of the grayscale image is processed by SMT, and the origin and point of the image are processed by SMTx，y) The sum expression of all pixels in the matrix area for the diagonal vertices is:
Figure 466745DEST_PATH_IMAGE002
（2）
in the formula (2), I 0 （x，y) Representing origin and point of images for SMT patch processingx，y) The sum of all pixels in the matrix area that are diagonal vertices; the difference function is the difference between two kernels at different scales, the difference function being:
Figure 834273DEST_PATH_IMAGE003
（3）
in the formula (3), the first and second groups,kis a constant between 3 and 4, L: (x，yσ) refers to the image dimension for SMT patch processing; then, screening and positioning characteristic points for SMT patch processing; selecting candidate feature points to eliminate unstable feature points; fitting a quadratic function according to data around the candidate feature points, and accurately estimating the positions of the feature points and a proportional function as follows:
Figure 810319DEST_PATH_IMAGE004
（4）
in formula (4), X =: (b)x，y，σ) T Representing the offset of the original sampling point of the incoming material image information for SMT patch processing;
Figure 40704DEST_PATH_IMAGE006
the position points are represented by a number of positions,
Figure 153016DEST_PATH_IMAGE007
the position points are subjected to a first-level offset calculation amount,
Figure 957024DEST_PATH_IMAGE008
the position points are subjected to secondary offset calculation quantity,
Figure 358050DEST_PATH_IMAGE009
a position function representing an accurately estimated feature point,
Figure 905706DEST_PATH_IMAGE010
representing a scaling function; ,
Figure 872525DEST_PATH_IMAGE011
representing three levels of offset computation for the location points,
Figure 378592DEST_PATH_IMAGE012
the amount of calculation that the position points are not biased.
As a further technical scheme of the invention, the sensor is a pressure sensor, a negative pressure sensor, an angle sensor, a position sensor, an image sensor, a photoelectric sensor or a material sensor.
As a further technical scheme of the invention, the classifier is a decision tree classifier.
As a further technical scheme of the invention, the template matching method based on the metadata engine comprises the following steps:
the method comprises the steps that a database engine is arranged in a BOM component database, data information of SMT patch processing incoming materials is searched for the database engine through an access engine, the access engine is a metadata access engine, the metadata access engine is used for processing data information of different SMT patches in the BOM component database, when the searched SMT patch processing incoming materials are consistent with template data information in the BOM component database, the SMT patch processing incoming materials are released, and otherwise, the SMT patch processing incoming materials are prevented from being released.
The invention has the beneficial and positive effects that:
different from the conventional technology, the method has the advantages that the image information of the incoming material is obtained by the optical detection system, the incoming material for SMT patch processing is positioned by the three-level positioning method, the improved SURF algorithm model is utilized and is transmitted to the computer for image processing, the measurement and positioning capacity of the incoming material for SMT patch processing can be realized in an automatic measurement mode, and the detection precision and the production capacity of a production line are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without inventive labor, wherein:
FIG. 1 is a schematic diagram of the overall architecture of the present invention;
FIG. 2 is a schematic diagram of a template matching method of the metadata engine according to the present invention.
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings, and it should be understood that the embodiments described herein are merely for the purpose of illustrating and explaining the present invention and are not intended to limit the present invention.
As shown in fig. 1, a method for measuring and positioning incoming material for SMT patch processing includes:
placing an SMT (surface mount technology) paster processing incoming material in a production line, acquiring image information of the incoming material through an optical detection system, and transmitting the image information to a computer for image processing;
when image data information is obtained, positioning incoming materials for SMT patch processing by a three-level positioning method;
the first positioning method senses incoming material data information for SMT patch processing through a sensor, and realizes primary positioning of the incoming material information for SMT patch processing; the initial positioning reflects the type of incoming materials for SMT patch processing;
the second positioning method is that the incoming material data information screening for SMT patch processing is realized through a classifier screening method, incoming materials for SMT patch processing of the same type are released, and incoming materials for SMT patch processing of different types are blocked from flowing to the next station;
the third positioning method is to realize the screening of the incoming material single incoming material data information for SMT patch processing by a template matching method based on a metadata engine, and to block the incoming material data information with faults from flowing to the next station.
As shown in fig. 1, in the above embodiment, the optical detection system includes an image acquisition module, an image processing module, a DSP processing unit, an LED array illumination unit, and an a/D conversion unit, where the DSP processing unit is connected to the image acquisition module, the a/D conversion unit, and the LED array illumination unit, respectively, and the image acquisition module is connected to the image processing module, where the DSP processing unit is a TMS320C6747 chip, and the image processing module includes an improved SURF algorithm model.
In the above embodiment, the improved SURF algorithm model includes an image feature extraction module and an accelerator connected to the image feature extraction module, wherein:
the working method for improving the SURF algorithm model comprises the following steps: the method comprises the steps of detecting incoming materials for SMT patch processing in an LED array lighting unit scale space, determining the scale and size of the incoming materials for SMT patch processing, utilizing the main inclined direction of an image neighborhood where the incoming materials for SMT patch processing are located as the directional characteristic of a target, adopting an integral image accelerator to achieve speed acceleration of convolution operation, improving image storage amount in a unit area through a mode of compressing an image format by the integral image accelerator, finally generating 12 x 12 dimensional feature point descriptors through extreme point neighborhood gradient information, and further achieving segmentation of the incoming materials for SMT patch processing through image information segmentation.
The method for processing SMT paster processing incoming materials by the improved SURF algorithm model comprises the following steps:
firstly, setting the scale space as G: (x，yσ) convolution to map the original SMT patch processing incoming material image to a scale space G: (x，yσ) range, G: (x，yσ) is inDisplaying SMT paster processing used material image data information in a three-dimensional space unit, and processing the two-dimensional gray scale of the obtained image as follows:
Figure 798072DEST_PATH_IMAGE013
（1）
in the formula (1), the symbol 8727represents a convolution operation (1)x，y) The position of a pixel in an image is represented, sigma represents a scale space factor, the smaller the sigma value is, the smaller the corresponding scale is, the more image details are kept, and the image of the SMT patch processing incoming material becomes smoother and smoother as sigma gradually increases. I (a)x，y) The pixel of the grayscale image is processed by SMT, and the origin and point of the image are processed by SMTx，y) The sum expression of all pixels in the matrix area for the diagonal vertices is:
Figure 352682DEST_PATH_IMAGE014
（2）
in the formula (2), I 0 （x，y) Representing origin and point of image material processed by SMTx，y) The sum of all pixels in the matrix area that are diagonal vertices;
in order to effectively detect stable feature points in the scale space, the original SMT patch processing incoming material image needs to be projected into the differential scale space using a gaussian difference function. The difference function is the difference between two kernels at different scales, and is:
Figure 174007DEST_PATH_IMAGE015
（3）
in the formula (3), the first and second groups,kis a constant between 3 and 4, L: (x，yσ) refers to the image dimension for SMT patch processing;
in practical implementation, the subtraction of incoming material images for two adjacent scale SMT patch processing is used to obtain a differential scale space D: (x，y，σ），Since the subtraction operation greatly reduces the amount of computation.
Then, screening and positioning characteristic points of SMT patch processing materials; selecting candidate feature points to eliminate unstable feature points; deleting the candidate characteristic points of the incoming material image information for the SMT patch processing with low contrast, fitting a quadratic function according to data around the candidate characteristic points, and accurately estimating the positions and the proportional functions of the characteristic points as follows:
Figure 585397DEST_PATH_IMAGE016
（4）
in formula (4), X =: (b)x，y，σ) T Representing the offset of the original sampling point of the incoming material image information for SMT patch processing;
Figure 492173DEST_PATH_IMAGE017
the position points are represented by a number of positions,
Figure 381632DEST_PATH_IMAGE018
the position points are subjected to a first-level offset calculation amount,
Figure 588622DEST_PATH_IMAGE019
the position points are subjected to secondary offset calculation quantity,
Figure 170913DEST_PATH_IMAGE020
a position function representing the accurately estimated feature points,
Figure 768248DEST_PATH_IMAGE021
a scaling function is represented.
Figure 461397DEST_PATH_IMAGE022
Representing three levels of offset computation for the location points,
Figure 991736DEST_PATH_IMAGE023
the amount of calculation that the position points are not biased.
In a particular embodiment, the curvature in one direction along the side of the SMT patch process used to feed the image information is small and the curvature in the direction perpendicular to the side increases. Therefore, if the principal curvature ratio is too large, the edge points can be deleted because the eigenvalues of the Hessian matrix are proportional to the principal curvature. Therefore, in order to improve the image processing capability of the image information of the incoming material for SMT patch processing, the main curvature is calculated by a Hessian matrix:
Figure 744928DEST_PATH_IMAGE024
（5）
in the formula (5), H is Hessian matrix, ratio is curvature of the SMT patch processing used material image along the edge direction,rto formulate the simplified parameters. The ratio is only equal torIn connection with, followingrIs increased, the ratio is also increased whenrWhen =1, the ratio is the minimum value. Therefore, the edge point can be eliminated by the ratio calculation. However, when the Hessian matrix is obtained, gaussian smoothing is required first, and then a second derivative is obtained, so that discrete pixel points are formed by template convolution.
In one embodiment, the image acquisition module is usually a camera, and the image acquisition module acquires data information of incoming materials for SMT chip processing by using a high-definition camera to realize measurement of the incoming material information for SMT chip processing, wherein the image information acquired by the image acquisition module is PCB (printed circuit board) information, transistors, diodes, SMT chips, inductors, magnetic beads, connectors, crystal oscillators and other BOM (basic on module) components for SMT chip processing, and when the data information is acquired by the image acquisition module, the PCB (printed circuit board) information, the transistors, the diodes, the SMT chips, the inductors, the magnetic beads, the connectors, the crystal oscillators and the likeA/DWhen the conversion unit receives the incoming material data information for SMT patch processing, the conversion unit is startedA/DAnd (6) converting the action.A/DThe conversion unit converts the collected analog signal into digital signal and converts the digital signal into digital signalDSPThe processing unit sends an interrupt signal, then it can passDSPThe processing unit implements processing of data.DSPThe processing unit can carry out time marking on the phasor information output by calculation.DSPThe core processing unit in the processing unit mainly adoptsTIOf a companyTMS320C6747The chip is provided with a plurality of chips,MS320C6747the operation speed of the chip can satisfy300MHzIs provided with a serial port controlled by the measuring device,16A serializer andFIFObuffer-connected multi-channel serial terminal15(McASP). Then copying the processed information to double portsRAMData area, so as to realizeARMAnd identifying and reading the data management module. Double ports adopted by the deviceRAMThe chip adoptsIDT70V28L20PFIThereon is provided with32KThe data buffer capacity of (2).
In the above step, the sensor is a pressure sensor, a negative pressure sensor, an angle sensor, a position sensor, an image sensor, a photoelectric sensor, or a material sensor.
In the above steps, the classifier is a decision tree classifier.
In the above steps, the template matching method based on the metadata engine is as follows:
the method comprises the steps that a database engine is arranged in a BOM component database, data information of SMT paster processing incoming materials is searched for the database engine through an access engine, the access engine is a metadata access engine, the metadata access engine is used for processing the data information of different SMT paster processing incoming materials in the BOM component database, when the searched SMT paster processing incoming materials are consistent with template data information in the BOM component database, the SMT paster processing incoming materials are released, and otherwise, the SMT paster processing incoming materials are prevented from being released.
In the specific embodiment, a spatial metadata model is constructed, a metadata engine system is designed, a user can quickly search and query metadata from a metadata database through a metadata engine, the positioning and monitoring functions of the metadata are fully exerted, original data and data are integrated, and the quick management of the metadata is realized. When a data source is selected, a metadata access engine is adopted, and the access engine searches in a shared metadata database to output search records, such as which database the data source belongs to, which table the BOM component database belongs to, which attributes are contained, and the like. The database is constructed as a vertical distributed search, and data output by different node devices are stored in the database and are uniformly stored and managed. After the access engine enters the database, the data can be stored in a log database in a log record mode, and the log database can contain the time and the requirement of user access. Through log recording, a user can verify whether data is written correctly or not, if the data is not written correctly, the user can modify the access engine, and when the data is accessed again, the user selects a new data source. The modification information is then automatically updated in the database,
although specific embodiments of the present invention have been described above, it will be understood by those skilled in the art that these specific embodiments are merely illustrative and that various omissions, substitutions and changes in the form of the detail of the methods and systems described above may be made by those skilled in the art without departing from the spirit and scope of the invention. For example, it is within the scope of the present invention to combine the steps of the above-described methods to perform substantially the same function in substantially the same way to achieve substantially the same result. Accordingly, the scope of the invention is to be limited only by the following claims.

Claims (4)

1. A method for measuring and positioning incoming materials for SMT patch processing is characterized in that: placing an SMT (surface mount technology) paster processing incoming material in a production line, acquiring image information of the incoming material through an optical detection system, and transmitting the image information to a computer for image processing; when image data information is obtained, positioning incoming materials for SMT patch processing by a three-level positioning method; the first positioning method senses incoming material data information for SMT patch processing through a sensor, and realizes primary positioning of the incoming material information for SMT patch processing; the initial positioning reflects the type of incoming materials for SMT patch processing; the second positioning method is to realize the data information screening of the incoming materials for SMT patch processing by a classifier screening method, the incoming materials for SMT patch processing of the same type are released, and the incoming materials for SMT patch processing of different types are blocked from flowing to the next station; the third positioning method is to realize the screening of the incoming material single incoming material data information for SMT patch processing by a template matching method based on a metadata engine and prevent the incoming material data information with faults from flowing to the next station; the template matching method based on the metadata engine comprises the following steps:
setting a database engine in a BOM component database, searching data information of incoming SMT patch processing materials for the database engine through an access engine, wherein the access engine is a metadata access engine which is used for processing the data information of different SMT patches in the BOM component database, and when the searched incoming SMT patch processing materials are consistent with template data information in the BOM component database, the incoming SMT patch processing materials are released, otherwise, the incoming SMT patch processing materials are prevented from being released;
the optical detection system comprises an image acquisition module, an image processing module, a DSP (digital signal processor) processing unit, an LED array illumination unit and an A/D (analog to digital) conversion unit, wherein the DSP processing unit is respectively connected with the image acquisition module, the A/D conversion unit and the LED array illumination unit, the image acquisition module is connected with the image processing module, the DSP processing unit is a TMS320C6747 chip, and the image processing module comprises an improved SURF (speeded Up robust feature) algorithm model; the improved SURF algorithm model comprises an image feature extraction module and an accelerator connected with the image feature extraction module, wherein: the working method for improving the SURF algorithm model comprises the following steps: detecting incoming materials for SMT (surface mount technology) patch processing in an LED array lighting unit scale space, determining the scale and size of the incoming materials for SMT patch processing, then using the main inclination direction of the image neighborhood where the incoming materials for SMT patch processing are located as the direction characteristic of a target, adopting an integral image accelerator to realize convolution operation speed acceleration, improving the image storage amount in a unit area by the integral image accelerator in a mode of compressing an image, finally generating 12 x 12-dimensional feature point descriptors by using extreme point neighborhood gradient information, and further realizing the incoming material segmentation for SMT patch processing through image information segmentation.
2. An SMT patch processing incoming material measuring and positioning method according to claim 1, wherein the method further comprises: the method for processing SMT paster processing incoming materials by the improved SURF algorithm model comprises the following steps:
first, setting the scale space as G: (x，yσ) convolution to map the original SMT patch processing incoming material image to scale space G: (x，yσ) range, G: (a)x，ySigma) is that the data information of SMT paster processing used material image is displayed in a three-dimensional space unit, and the obtained image is twoThe dimension gray level processing is as follows:
Figure 208375DEST_PATH_IMAGE001
（1）
in the formula (1), the symbol 8727represents a convolution operation (1)x，y) Represents the position of the pixel in the image, σ represents a scale space factor; i (a)x，y) The SMT patch processing is used for feeding gray image pixels, and the SMT patch processing is used for feeding image origin points and points: (x，y) The sum expression of all pixels in the matrix area for the diagonal vertices is:
Figure 466181DEST_PATH_IMAGE002
（2）
in the formula (2), I 0 （x，y) Representing origin and point of images for SMT patch processingx，y) The sum of all pixels in the matrix area that are diagonal vertices; the difference function is the difference between two kernels at different scales, the difference function being:
Figure 568129DEST_PATH_IMAGE003
（3）
in the formula (3), the reaction mixture is,kis a constant between 3 and 4, L: (x，yσ) refers to the image dimension for SMT patch processing; then, screening and positioning characteristic points for SMT patch processing; selecting candidate feature points to eliminate unstable feature points; fitting a quadratic function according to data around the candidate feature points, and accurately estimating the positions of the feature points and a proportional function as follows:
Figure 544175DEST_PATH_IMAGE004
（4）
in the formula (4), X =: (x，y，σ) T Representing the offset of original sampling points of the image information of incoming materials for SMT patch processing;
Figure 491403DEST_PATH_IMAGE005
the position points are represented by a number of positions,
Figure 869294DEST_PATH_IMAGE006
the position point is subjected to a first-level offset calculation amount,
Figure 378029DEST_PATH_IMAGE007
the position points are subjected to secondary offset calculation quantity,
Figure 575793DEST_PATH_IMAGE008
a position function representing an accurately estimated feature point,
Figure 389028DEST_PATH_IMAGE009
the function of the scale is expressed as,
Figure 559109DEST_PATH_IMAGE010
representing three levels of offset computation for the location points,
Figure 65177DEST_PATH_IMAGE011
the amount of calculation that the position points are not biased.
3. An SMT patch processing incoming material measuring and positioning method according to claim 1, wherein the method further comprises: the sensor is a pressure sensor, a negative pressure sensor, an angle sensor, a position sensor, an image sensor, a photoelectric sensor or a material sensor.
4. An SMT patch processing incoming material measuring and positioning method according to claim 1, wherein the method further comprises: the classifier is a decision tree classifier.
CN202210745213.1A 2022-06-29 2022-06-29 Incoming material measuring and positioning method for SMT (surface mount technology) patch processing Active CN114820613B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210745213.1A CN114820613B (en) 2022-06-29 2022-06-29 Incoming material measuring and positioning method for SMT (surface mount technology) patch processing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210745213.1A CN114820613B (en) 2022-06-29 2022-06-29 Incoming material measuring and positioning method for SMT (surface mount technology) patch processing

Publications (2)

Publication Number Publication Date
CN114820613A CN114820613A (en) 2022-07-29
CN114820613B true CN114820613B (en) 2022-10-28

Family

ID=82522922

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210745213.1A Active CN114820613B (en) 2022-06-29 2022-06-29 Incoming material measuring and positioning method for SMT (surface mount technology) patch processing

Country Status (1)

Country Link
CN (1) CN114820613B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112488512A (en) * 2018-04-18 2021-03-12 费希尔-罗斯蒙特系统公司 Quality inspection system employing metadata collection
CN114596301A (en) * 2022-03-18 2022-06-07 上海宝振自动化工程有限公司 Coating roller surface defect detection system adopting 3D line laser profile technology

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103049770A (en) * 2012-09-24 2013-04-17 共青城赛龙通信技术有限责任公司 SMT (Surface Mount Technology) anti-false material detection system and method
CN109523505B (en) * 2018-09-18 2021-06-11 深圳市智信精密仪器有限公司 Method for detecting pattern defects on surface of ceramic tile based on machine vision
US11972589B2 (en) * 2019-01-09 2024-04-30 Fuji Corporation Image processing device, work robot, substrate inspection device, and specimen inspection device
CN110135438B (en) * 2019-05-09 2022-09-27 哈尔滨工程大学 Improved SURF algorithm based on gradient amplitude precomputation
CN211090190U (en) * 2019-11-13 2020-07-24 深圳市瑞亿科技电子有限公司 Paster assembly platform convenient to position and install
CN113409409A (en) * 2020-03-17 2021-09-17 腾讯科技（深圳）有限公司 Image processing method, image processing device, electronic equipment and computer readable storage medium
CN113177563B (en) * 2021-05-07 2022-10-14 安徽帅尔信息科技有限公司 Post-chip anomaly detection method integrating CMA-ES algorithm and sequential extreme learning machine
CN113470005B (en) * 2021-07-23 2022-03-18 广东奥普特科技股份有限公司 Welding spot detection device and welding spot detection method for cylindrical battery cap

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112488512A (en) * 2018-04-18 2021-03-12 费希尔-罗斯蒙特系统公司 Quality inspection system employing metadata collection
CN114596301A (en) * 2022-03-18 2022-06-07 上海宝振自动化工程有限公司 Coating roller surface defect detection system adopting 3D line laser profile technology

Also Published As

Publication number Publication date
CN114820613A (en) 2022-07-29

Similar Documents

Publication Publication Date Title
US6549683B1 (en) Method and apparatus for evaluating a scale factor and a rotation angle in image processing
EP1388738A1 (en) Inspection data producing method and board inspection apparatus using the method
CN107705293A (en) A kind of hardware dimension measurement method based on CCD area array cameras vision-based detections
CN110621150B (en) Printed circuit board assembling method and related device
CN108596829A (en) A kind of PCB bare boards picture method for registering and device
CN113269775B (en) Defect detection method and device based on multi-scale feature fusion SSD
CN115294114A (en) Quality detection method based on ECU circuit welding
CN114723706A (en) Welding spot detection and positioning method based on machine vision
CN114755243A (en) FPC connector surface crack detection method based on machine vision
KR20100039816A (en) Method of detecting mainboard mark, detecting apparatus, and method of placing mainboard
CN115775246A (en) Method for detecting defects of PCB (printed circuit board) components
CN115797359A (en) Detection method and device based on solder paste on circuit board and storage medium
CN112419274A (en) Solder paste detection method, system, electronic device and medium
CN113920075A (en) Simple defect detection method and system based on object identification
CN112614086B (en) Learning device, inspection device, learning method, and inspection method
CN114820613B (en) Incoming material measuring and positioning method for SMT (surface mount technology) patch processing
CN116205835A (en) Circuit board flaw detection method and device and electronic equipment
CN114549393A (en) Image annotation method, device, equipment and computer storage medium
Klco et al. Automated detection of soldering splashes using YOLOv5 algorithm
CN111899248A (en) Automatic PCB soldering paste defect detection method based on machine learning
CN114926417B (en) Microscopic imaging detection method and system for weak scratches on surface of high-density flexible FPC
You et al. A vision system for an automatic assembly machine of electronic components
CN115100393A (en) Deep learning-based PCB welding defect detection method
CN115372793A (en) Circuit detection method for integrated circuit
Chen et al. A Model For IC Defect Identification Based on Feature Extraction

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant