[go: up one dir, main page]

CN118520227B - A method, device and readable storage medium for detecting foreign matter defects in microelectronic products - Google Patents

A method, device and readable storage medium for detecting foreign matter defects in microelectronic products Download PDF

Info

Publication number
CN118520227B
CN118520227B CN202410986735.XA CN202410986735A CN118520227B CN 118520227 B CN118520227 B CN 118520227B CN 202410986735 A CN202410986735 A CN 202410986735A CN 118520227 B CN118520227 B CN 118520227B
Authority
CN
China
Prior art keywords
sparse
target
foreign object
coding matrix
object defect
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
CN202410986735.XA
Other languages
Chinese (zh)
Other versions
CN118520227A (en
Inventor
宿磊
李可
段昕芳
明雪飞
顾杰斐
赵新维
周秀峰
李杨
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.)
Jiangnan University
CETC 58 Research Institute
Original Assignee
Jiangnan University
CETC 58 Research Institute
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 Jiangnan University, CETC 58 Research Institute filed Critical Jiangnan University
Priority to CN202410986735.XA priority Critical patent/CN118520227B/en
Publication of CN118520227A publication Critical patent/CN118520227A/en
Application granted granted Critical
Publication of CN118520227B publication Critical patent/CN118520227B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/10—Pre-processing; Data cleansing
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/04—Analysing solids
    • G01N29/045—Analysing solids by imparting shocks to the workpiece and detecting the vibrations or the acoustic waves caused by the shocks
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/44—Processing the detected response signal, e.g. electronic circuits specially adapted therefor
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G06F17/141—Discrete Fourier transforms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/28—Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2291/00—Indexing codes associated with group G01N29/00
    • G01N2291/26—Scanned objects
    • G01N2291/269—Various geometry objects
    • G01N2291/2697—Wafer or (micro)electronic parts
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Mathematical Optimization (AREA)
  • Computational Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Pure & Applied Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Analysis (AREA)
  • Algebra (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Pathology (AREA)
  • Chemical & Material Sciences (AREA)
  • Immunology (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Discrete Mathematics (AREA)
  • Signal Processing (AREA)
  • Acoustics & Sound (AREA)
  • Computing Systems (AREA)
  • Testing Or Measuring Of Semiconductors Or The Like (AREA)

Abstract

The application belongs to the technical field of microelectronic detection, and relates to a method and a device for detecting foreign matter defects of a microelectronic product and a readable storage medium; acquiring a foreign object defect collision signal of a microelectronic product to be detected, and setting weights for each sparse coefficient in a penalty term of a GMC sparse denoising model of the foreign object defect collision signal to obtain a target penalty term; obtaining a target GMC sparse denoising model based on the fidelity term and the target penalty term; iteratively solving a target GMC sparse denoising model to obtain a sparse coding matrix of the foreign object defect collision signal; performing sparse representation on the foreign object defect collision signals by using a sparse coding matrix to obtain target foreign object defect collision signals, so as to perform foreign object defect detection on the microelectronic product to be detected; according to the application, collision signals generated by sliding and collision of the foreign object defects in the cavity are reserved as much as possible, and environmental noise in the collision signals of the foreign object defects is removed, so that the accuracy of the detection result of the foreign object defects is improved.

Description

Method and device for detecting foreign matter defects of microelectronic product and readable storage medium
Technical Field
The present invention relates to the field of microelectronic testing technology, and more particularly, to a method and apparatus for detecting defects of foreign matters in microelectronic products, and a computer readable storage medium.
Background
The integrated circuit is an important component of technological development and is the basis for realizing intellectualization and digitalization in various industries. Electronic packaging is an indispensable procedure in integrated circuit manufacturing processes, and has important effects on the cost, performance, reliability, and the like of integrated circuit products. However, more than 25% of failures in integrated circuit products result from packaging, because the electronic packaging process is complex, involves a variety of materials and processes, and tends to introduce uncleaned and thorough bonding wire tails, conductive paste particles, ceramic particles, silicon chips, and foreign particles such as dust from the environment. When the product is in the environment of cluster impact vibration or high-speed phase change motion, movable foreign particles are easily activated and dissociated, and irregularly move in the cavity, even collide with the core cavity wall, the inner chip and the circuit, so that the situation that the functions of the bonding wire such as deformation, short circuit and short circuit are invalid can be possibly caused, the reliability of the microelectronic product is seriously influenced, and particularly in the aerospace field, the low reliability of the microelectronic product can not only cause huge economic loss, but also possibly threaten the life safety of astronauts, so that how to effectively detect the foreign defects in the microelectronic product after encapsulation is a key for guaranteeing the reliability of the product.
Common foreign matter Detection methods include a pre-capping microscopic method, a post-capping X-ray Detection method, a matela test method, a particle impact Noise Detection method (PIND) and the like, wherein the microscopic method relies on manual appearance Detection of raw material chips by using a microscope, the Detection efficiency and the Detection precision are low, and foreign matter particles can be introduced when the method is capped after Detection is completed, so that foreign matter defects are brought; the X-ray inspection method can nondestructively inspect the packaged product, but the inspection accuracy is affected by the resolution of the X-ray apparatus, and the X-ray inspection apparatus is expensive and requires periodic maintenance and calibration, which makes the inspection cost high; the matela detection method involves complicated mechanical test conditions and data analysis processes, requires a long detection time, and results in low detection efficiency, so that the foreign matter defect detection is usually performed by the PIND method with high efficiency and low cost at present.
The PIND method utilizes a vibrating table to generate certain mechanical impact and vibration, the impact loosens foreign matter defects bound in the cavity of the tested device, the vibration enables the foreign matter defects to slide and collide in the cavity and generates collision signals, and the sensor acquires the collision signals and then converts the collision signals into voltage forms to output and display, so that whether the tested device has the foreign matter defects is judged. However, as the process level is continuously increased, the size and quality of the foreign object defect can be controlled to a smaller range, which may result in that the foreign object collision signal may be submerged in the environmental noise even under the optimal test conditions, greatly affecting the accuracy of the foreign object defect detection.
Disclosure of Invention
Therefore, the technical problem to be solved by the invention is to solve the problem of lower accuracy of the detection result of the method for detecting the foreign matter defects of the microelectronic products in the prior art.
In order to solve the technical problems, the invention provides a method for detecting foreign matter defects of a microelectronic product, which comprises the following steps:
acquiring a foreign object defect collision signal of a microelectronic product to be detected, setting weights for each sparse coefficient in a punishment item of a GMC sparse denoising model of the foreign object defect collision signal, and obtaining a target punishment item, wherein the method specifically comprises the following steps:
dividing the foreign matter defect collision signal to obtain A dimensional foreign object defect collision signal matrix;
Singular value decomposition is carried out on the foreign matter defect collision signal matrix to obtain Singular values; taking each column element in the foreign object defect collision signal matrix as a sub-signal to obtain m sub-signals, and respectively calculating the signal kurtosis of each sub-signal;
Calculate the first Sum of singular values ofThe square root of the sum of squares of the signal kurtosis of the sub-signals is taken as the inverse of the square root of the sum of squares as the first of the sparse coding matrices of the collision signals of the foreign object defectsLine 1Weighting of column sparse coefficients; wherein, ，；
Obtaining a target GMC sparse denoising model based on the fidelity term and the target penalty term; iteratively solving the target GMC sparse denoising model to obtain a sparse coding matrix of the foreign object defect collision signal;
and performing sparse representation on the foreign object defect collision signal by using the sparse coding matrix to obtain a target foreign object defect collision signal, so as to perform foreign object defect detection on the microelectronic product to be detected.
Preferably, the first of the sparse coding matrices of the foreign object defect collision signalsLine 1The calculation formula of the weight of the column sparse coefficient is as follows:
，
wherein, Sparse coding matrix representing foreign object defect collision signalLine 1Weighting of column sparse coefficients; Represent the first Signal kurtosis of the sub-signals; Represent the first Singular values;
，
wherein, Represent the firstThe first of the sub-signalsData points; Represent the first The number of data points in the sub-signal; Represent the first The mean of all data points in the sub-signal.
Preferably, the target GMC sparse denoising model is expressed as:
，
wherein, Representing a target GMC sparse denoising model; Is a fidelity term; A collision signal matrix representing a foreign object defect; Representing a Laplace wavelet analysis dictionary; sparse coding matrix for representing collision signal of foreign object defect, comprising Row of linesColumn sparseness factor composition; representing regularization parameters; punishment items for targets; Sparse coding matrix representing foreign object defect collision signal Line 1Weighting of column sparse coefficients; Sparse coding matrix representing foreign object defect collision signal Line 1Column sparseness factor; representing the hadamard product.
Preferably, the construction process of the fidelity term comprises the following steps:
Establishing a Laplace wavelet analysis dictionary based on the foreign object defect collision signals and the foreign object defect collision signal matrix;
obtaining a target defect collision signal matrix based on a sparse coding matrix of the foreign object defect collision signal and the Laplace wavelet analysis dictionary;
and constructing a fidelity term based on the foreign object defect collision signal matrix and the target defect collision signal matrix.
Preferably, based on the foreign object defect collision signal and the foreign object defect collision signal matrix, establishing the Laplace wavelet analysis dictionary includes:
setting a value range of oscillation frequency, a value range of time shift parameters and a value range of viscous damping ratio; respectively taking values of the oscillation frequency, the time shift parameter and the viscous damping ratio in the value range of the oscillation frequency, the value range of the time shift parameter and the value range of the viscous damping ratio, constructing a Laplace wavelet atom based on each value, and obtaining a Laplace wavelet original subset based on all value combinations;
respectively calculating the similarity of each Laplace wavelet atom in the Laplace wavelet atom set and the foreign object defect collision signal;
And constructing a Laplace wavelet analysis dictionary based on the oscillation frequency, the time shift parameter, the viscous damping ratio and the Laplace wavelet atomic basis function corresponding to the Laplace wavelet atom with the maximum similarity.
Preferably, the Laplace wavelet atoms are represented as:
，
wherein, Represents Laplace wavelet atoms; representing the oscillation frequency; represents the viscous damping ratio; Representing a time shift parameter; Representing time; Representing a wavelet support interval;
the calculation formula of the similarity between Laplace wavelet atoms and the foreign object defect collision signals is as follows:
，
wherein, Representation ofAnd (3) withSimilarity of (2); represents Laplace wavelet atoms; a collision signal indicating a foreign object defect; representing the euclidean norm.
Preferably, the iterative solution of the target GMC sparse denoising model to obtain a sparse coding matrix of the foreign object defect collision signal includes:
Step 1: introducing a proxy variable and a Lagrangian multiplier to update the target GMC sparse denoising model, and decomposing the updated target GMC sparse denoising model into a proxy variable constraint sub-problem and a sparse coding matrix constraint sub-problem;
Step 2: setting an initial sparse coding matrix, an initial Lagrangian multiplier and the weight of each sparse coefficient in the initial sparse coding matrix, and calculating the initial value of a target GMC sparse denoising model based on the initial sparse coding matrix and the weight of each sparse coefficient in the initial sparse coding matrix;
step 3: substituting the initial sparse coding matrix, the weight of each sparse coefficient in the initial sparse coding matrix and the initial Lagrangian multiplier into the agent variable constraint subproblem, and solving a target agent variable;
step 4: substituting the target proxy variable and the initial Lagrangian multiplier into the constraint sub-problem of the sparse coding matrix, and solving a target sparse coding matrix;
step 5: the target sparse coding matrix is utilized to sparse the foreign object defect collision signals, and singular value decomposition is carried out on the sparse foreign object defect collision signals to obtain Calculating the weight of each sparse coefficient in the target sparse coding matrix by using the target singular values;
step 6: solving a target lagrangian multiplier based on the target proxy variable, the target sparse coding matrix, and the initial lagrangian multiplier;
step 7: calculating a target value of the GMC sparse denoising model based on the target sparse coding matrix and the weight of each sparse coefficient in the target sparse coding matrix, and calculating a difference value between an initial value of the target GMC sparse denoising model and the target value of the target GMC sparse denoising model;
Step 8: judging the difference value and the preset convergence error, and if the difference value is smaller than or equal to the preset convergence error, taking the target sparse coding matrix as a sparse coding matrix of a foreign object defect collision signal;
Step 9: and if the difference is greater than the preset convergence error, taking the target value of the GMC sparse denoising model as an initial value of the GMC sparse denoising model, taking the target sparse coding matrix as an initial sparse coding matrix, taking the weight of each sparse coefficient in the target sparse coding matrix as the weight of each sparse coefficient in the initial sparse coding matrix, taking the target Lagrangian multiplier as an initial Lagrangian multiplier, and returning to the execution step 3 until the difference between the initial value of the GMC sparse denoising model and the target value of the GMC sparse denoising model is smaller than or equal to the preset convergence error.
Preferably, the updated target GMC sparse denoising model is expressed as:
，
wherein, Representing the updated target GMC sparse denoising model; Representing the proxy variable; Representing lagrangian multipliers; weight matrix representing weight composition of all sparse coefficients in sparse coding matrix of foreign object defect collision signal Is the first of (2)A column; Is a fidelity term; A collision signal matrix representing a foreign object defect; Representing a Laplace wavelet analysis dictionary; a sparse coding matrix representing a foreign object defect collision signal; representing regularization parameters; punishment items for targets; Sparse coding matrix representing foreign object defect collision signal Line 1Weighting of column sparse coefficients; Sparse coding matrix representing foreign object defect collision signal Line 1Column sparseness factor; Representing the Hadamard product; Representing a penalty parameter;
the proxy variable constraint sub-problem is expressed as:
，
wherein, A solution vector representing the proxy variable,Solution vector of proxy variableLine 1The value of the element of the column,Representing the th of Lagrange multiplierLine 1Element values of columns;
the sparse coding matrix constraint sub-problem is expressed as:
，
wherein, A solution vector representing the sparse coding matrix;
The calculation formula of the target Lagrangian multiplier is as follows:
，
wherein, Representing the target lagrangian multiplier,Representing the lagrangian multiplier and,The relaxation factor is indicated as such,The target sparse coding matrix is represented by a matrix,Representing the target proxy variable.
The invention also provides a device for detecting the foreign matter defects of the microelectronic product, which comprises:
The target punishment item construction module is used for acquiring a foreign object defect collision signal of the microelectronic product to be detected, setting weights for each sparse coefficient in a punishment item of a GMC sparse denoising model of the foreign object defect collision signal to obtain a target punishment item, and specifically comprises the following steps:
A signal segmentation sub-module for segmenting the foreign object defect collision signal to obtain A dimensional foreign object defect collision signal matrix;
a singular value and signal kurtosis calculation sub-module for performing singular value decomposition on the foreign object defect collision signal matrix to obtain Singular values; taking each column element in the foreign object defect collision signal matrix as a sub-signal to obtain m sub-signals, and respectively calculating the signal kurtosis of each sub-signal;
A sparse coefficient weight assignment sub-module for calculating the first Sum of singular values ofThe square root of the sum of squares of the signal kurtosis of the sub-signals is taken as the inverse of the square root of the sum of squares as the first of the sparse coding matrices of the collision signals of the foreign object defectsLine 1Weighting of column sparse coefficients; wherein, ，；
The sparse coding matrix acquisition module is used for acquiring a target GMC sparse denoising model based on the fidelity item and the target penalty item; iteratively solving the target GMC sparse denoising model to obtain a sparse coding matrix of the foreign object defect collision signal;
And the foreign object defect detection module is used for carrying out sparse representation on the foreign object defect collision signal by utilizing the sparse coding matrix to obtain a target foreign object defect collision signal, so as to carry out foreign object defect detection on the microelectronic product to be detected.
The invention also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program realizes the steps of the method for detecting the foreign body defects of the microelectronic product when being executed by a processor.
According to the method for detecting the foreign object defects of the microelectronic product, the GMC punishment function is used for sparse denoising of the foreign object defect collision signals of the microelectronic product, the sparse coding matrix is obtained based on the GMC sparse denoising model, so that the environmental noise in the foreign object defect collision signals is removed, the signal amplitude is enhanced, the foreign object particle collision situation submerged in the environmental noise can be effectively reflected by considering the kurtosis characteristic of the signals, the greater the kurtosis of the signals indicates that the probability of the signals generated by the foreign object defect collision is greater, and the foreign object particle collision information is mainly concentrated in the larger singular value, so that the method for detecting the foreign object defect collision signals by using the GMC sparse denoising model is obtained by dividing the foreign object defect collision signals when the punishment item of the GMC sparse denoising model is constructedThe foreign matter defect collision signal matrix is maintained to obtain m sub-signals, the signal kurtosis of each sub-signal is calculated respectively, and the foreign matter defect collision signal is subjected to singular value decomposition to obtainThe singular value is calculated, the weight of the j th row and the i th column sparse coefficient in the sparse coding matrix is constructed based on the square of the signal kurtosis of the j th singular value and the i th sub-signal and the inverse of the arithmetic square root, and the greater the signal kurtosis and the singular value, the smaller the weight of the sparse coefficient is, so that the sparse coding matrix obtained by solving can attenuate the sub-signal with larger kurtosis and the larger singular value component in a smaller way, the collision signal generated by sliding and collision of the foreign defect in the cavity is reserved as much as possible, the environmental noise in the foreign defect collision signal is removed, and the accuracy of the foreign defect detection result is improved.
Drawings
In order that the invention may be more readily understood, a more particular description of the invention will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings, in which:
FIG. 1 is a flowchart of a method for detecting a foreign object defect of a microelectronic product according to the present application;
FIG. 2 is a diagram showing a foreign object defect collision signal according to an embodiment of the present application;
FIG. 3 is a diagram illustrating a foreign object defect collision signal after denoising the signal shown in FIG. 2 using a signal processing method according to an embodiment of the present application; the method includes (a) in fig. 3, in which the signal shown in fig. 2 is denoised by using a wavelet packet denoising method, (b) in fig. 3, in which the signal shown in fig. 2 is denoised by using a VMD denoising method, (c) in fig. 3, in which the signal shown in fig. 2 is denoised by using an L1 sparse denoising method, and (d) in fig. 3, in which the signal shown in fig. 2 is denoised by using the method provided by the present application;
Fig. 4 is a schematic structural diagram of a device for detecting a defect of a foreign body in a microelectronic product according to the present application.
Detailed Description
The present invention will be further described with reference to the accompanying drawings and specific examples, which are not intended to be limiting, so that those skilled in the art will better understand the invention and practice it.
Referring to fig. 1, fig. 1 is a flowchart of a method for detecting a foreign object defect of a microelectronic product according to an embodiment of the application, where the method specifically includes:
S10: acquiring a foreign object defect collision signal of a microelectronic product to be detected, setting weights for each sparse coefficient in a punishment item of a GMC sparse denoising model of the foreign object defect collision signal, and obtaining a target punishment item, wherein the method specifically comprises the following steps:
s100: dividing the collision signal of the foreign object defect to obtain A dimensional foreign object defect collision signal matrix;
illustratively, the obtained foreign matter defect collision signal with the length of N is segmented to obtain The foreign matter defect collision signal matrix of dimension, the partitioning operator is:
，
wherein, A collision signal indicating a foreign object defect; Representing a window length of a partitioning operator; The window shift length is represented and is used for eliminating the end effect of the blocking operator; based on Obtaining a dimensional foreign matter defect collision signal matrixThe sub-signals of the sub-signals,；
S101: singular value decomposition is carried out on the collision signal matrix of the foreign object defect to obtainSingular values; taking each column element in the foreign matter defect collision signal matrix as a sub-signal to obtain m sub-signals, and respectively calculating the signal kurtosis of each sub-signal;
S102: calculate the first Sum of singular values ofThe square root of the sum of squares and the arithmetic square root of the signal kurtosis of the sub-signals is taken as the inverse of the square root of the sum of squares and the arithmetic square root in the sparse coding matrix of the foreign matter defect collision signalsLine 1Weighting of column sparse coefficients; wherein, ，；
S20: obtaining a target GMC sparse denoising model based on the fidelity term and the target penalty term; iteratively solving a target GMC sparse denoising model to obtain a sparse coding matrix of the foreign object defect collision signal;
S30: and performing sparse representation on the foreign object defect collision signal by using the sparse coding matrix to obtain a target foreign object defect collision signal, thereby performing foreign object defect detection on the microelectronic product to be detected.
The method for detecting the foreign object defects of the microelectronic product provided by the application uses the GMC punishment function for sparse denoising of the foreign object defect collision signals of the microelectronic product, thereby removing the environmental noise in the foreign object defect collision signals and enhancing the signal amplitude, meanwhile, the condition of collision of foreign object particles submerged in the environmental noise can be effectively reflected by considering the kurtosis characteristic of the signals, the greater the kurtosis of the signals is, the greater the probability that the signals are generated by collision of the foreign object defects is, and the collision information of the foreign object particles is mainly concentrated in the larger singular value, so that the method for detecting the foreign object defect collision signals in the application divides the foreign object defect collision signals to obtain when constructing punishment items of a GMC sparse denoising modelThe foreign matter defect collision signal matrix is maintained to obtain m sub-signals, the signal kurtosis of each sub-signal is calculated respectively, and the foreign matter defect collision signal is subjected to singular value decomposition to obtainThe singular value is calculated, the weight of the j th row and the i th column sparse coefficient in the sparse coding matrix is constructed based on the square of the signal kurtosis of the j th singular value and the i th sub-signal and the inverse of the arithmetic square root, and the greater the signal kurtosis and the singular value, the smaller the weight of the sparse coefficient is, so that the sparse coding matrix obtained by solving can attenuate the sub-signal with larger kurtosis and the larger singular value component in a smaller way, the collision signal generated by sliding and collision of the foreign defect in the cavity is reserved as much as possible, the environmental noise in the foreign defect collision signal is removed, and the accuracy of the foreign defect detection result is improved.
Specifically, in some embodiments of the present application, the foreign object defect collision signal in step S102 is in the sparse coding matrixLine 1The calculation formula of the weight of the column sparse coefficient is as follows:
，
wherein, Sparse coding matrix representing foreign object defect collision signalLine 1Weighting of column sparse coefficients; Represent the first Signal kurtosis of the sub-signals; Represent the first Singular values;
，
wherein, Represent the firstThe first of the sub-signalsData points; Represent the first The number of data points in the sub-signal; Represent the first The mean of all data points in the sub-signal.
Specifically, the construction process of the fidelity item in step S20 includes:
based on the foreign object defect collision signals and the foreign object defect collision signal matrix, a Laplace wavelet analysis dictionary is established, which specifically comprises:
Setting a value range of oscillation frequency, a value range of time shifting parameter and a value range of viscous damping ratio, respectively carrying out value shifting on the oscillation frequency, the time shifting parameter and the viscous damping ratio in the value range of the oscillation frequency, the value range of the time shifting parameter and the value range of the viscous damping ratio, constructing a Laplace wavelet atom based on each value, and obtaining a Laplace wavelet original subset based on all value combinations; specifically, laplace wavelet atoms are expressed as:
，
wherein, Represents Laplace wavelet atoms; representing the oscillation frequency; represents the viscous damping ratio; Representing a time shift parameter; Representing time; Representing a wavelet support interval;
Respectively calculating the similarity of collision signals of each Laplace wavelet atom and the foreign object defect in the Laplace wavelet atom set; specifically, the calculation formula of the similarity degree of the Laplace wavelet atom and the foreign matter defect collision signal is as follows:
，
wherein, Representation ofAnd (3) withSimilarity of (2); represents Laplace wavelet atoms; a collision signal indicating a foreign object defect; representing euclidean norms;
Constructing a Laplace wavelet analysis dictionary based on the oscillation frequency, time shift parameters, viscous damping ratio and Laplace wavelet atomic basis functions corresponding to Laplace wavelet atoms with the maximum similarity;
Obtaining a target defect signal matrix based on a sparse coding matrix of the foreign object defect collision signal and a Laplace wavelet analysis dictionary;
and constructing a fidelity term based on the foreign object defect collision signal matrix and the target defect signal matrix.
Further, based on the above-mentioned fidelity term and the target penalty term, the target GMC sparse denoising model in the embodiment of the present application is expressed as:
，
wherein, Representing a target GMC sparse denoising model; Is a fidelity term; A collision signal matrix representing a foreign object defect; Representing a Laplace wavelet analysis dictionary; sparse coding matrix for representing collision signal of foreign object defect, comprising Row of linesColumn sparseness factor composition; representing regularization parameters; punishment items for targets; Sparse coding matrix representing foreign object defect collision signal Line 1Weighting of column sparse coefficients; Sparse coding matrix representing foreign object defect collision signal Line 1Column sparseness factor; representing the hadamard product.
Further, in step S20, solving the target GMC sparse denoising model specifically includes:
Step 1: introducing a proxy variable and a Lagrangian multiplier to update the target GMC sparse denoising model, and decomposing the updated target GMC sparse denoising model into a proxy variable constraint sub-problem and a sparse coding matrix constraint sub-problem;
specifically, the updated GMC sparse denoising model is expressed as:
，
wherein, Representing the updated target GMC sparse denoising model; Representing the proxy variable; Representing lagrangian multipliers; weight matrix representing weight composition of all sparse coefficients in sparse coding matrix of foreign object defect collision signal Is the first of (2)A column; Is a fidelity term; A collision signal matrix representing a foreign object defect; Representing a Laplace wavelet analysis dictionary; a sparse coding matrix representing a foreign object defect collision signal; representing regularization parameters; punishment items for targets; Sparse coding matrix representing foreign object defect collision signal Line 1Weighting of column sparse coefficients; Sparse coding matrix representing foreign object defect collision signal Line 1Column sparseness factor; Representing the Hadamard product; Representing a penalty parameter;
the proxy variable constraint sub-problem is expressed as:
，
wherein, A solution vector representing the proxy variable,Solution vector of proxy variableLine 1The value of the element of the column,Representing the th of Lagrange multiplierLine 1Element values of columns; specifically, solving for the resulting target proxy variable，The closed-form solution of (2) is:
，
wherein, Representing the first of the target proxy variablesThe column vector is used to determine the position of the column,Representing the first of the sparse coding matricesColumn sparseness factor;
wherein, A threshold function representing a target penalty term, which is expressed in particular as:
，
wherein, Representing non-convexity parameters, satisfying；
The sparse coding matrix constraint sub-problem is:
，
wherein, A solution vector representing the sparse coding matrix; further, the method comprises the steps of,，Representing the identity matrix;
Step 2: setting an initial sparse coding matrix, an initial Lagrangian multiplier and the weight of each sparse coefficient in the initial sparse coding matrix, and calculating the initial value of a target GMC sparse denoising model based on the initial sparse coding matrix and the weight of each sparse coefficient in the initial sparse coding matrix;
step 3: substituting the initial sparse coding matrix, the weight of each sparse coefficient in the initial sparse coding matrix and the initial Lagrangian multiplier into the agent variable constraint subproblem, and solving a target agent variable;
step 4: substituting the target proxy variable and the initial Lagrangian multiplier into the constraint sub-problem of the sparse coding matrix, and solving a target sparse coding matrix;
step 5: the target sparse coding matrix is utilized to sparse the foreign object defect collision signals, and singular value decomposition is carried out on the sparse foreign object defect collision signals to obtain Calculating the weight of each sparse coefficient in the target sparse coding matrix by using the target singular values;
step 6: solving a target lagrangian multiplier based on the target proxy variable, the target sparse coding matrix, and the initial lagrangian multiplier;
specifically, the calculation formula of the target lagrangian multiplier is:
，
wherein, Representing the target lagrangian multiplier,Representing the lagrangian multiplier and,The relaxation factor is indicated as such,The target sparse coding matrix is represented by a matrix,Representing a target proxy variable;
step 7: calculating a target value of the GMC sparse denoising model based on the target sparse coding matrix and the weight of each sparse coefficient in the target sparse coding matrix, and calculating a difference value between an initial value of the target GMC sparse denoising model and the target value of the target GMC sparse denoising model;
Step 8: judging the difference value and the preset convergence error, and if the difference value is smaller than or equal to the preset convergence error, taking the target sparse coding matrix as a sparse coding matrix of a foreign object defect collision signal;
step 9: and if the difference is greater than the preset convergence error, taking the target value of the GMC sparse denoising model as an initial value of the GMC sparse denoising model, taking the target sparse coding matrix as an initial sparse coding matrix, taking the weight of each sparse coefficient in the target sparse coding matrix as the weight of each sparse coefficient in the initial sparse coding matrix, taking the target Lagrangian multiplier as an initial Lagrangian multiplier, and returning to the execution step 3 until the difference between the initial value of the GMC sparse denoising model and the target value of the GMC sparse denoising model is less than or equal to the preset convergence error.
In order to verify the effectiveness of the method for detecting the foreign object defects of the microelectronic product, the embodiment of the application also provides a comparative example for performing sparse denoising on the collision signals of the foreign object defects by using different sparse denoising methods:
As shown in fig. 2, the embodiment of the application provides a foreign matter defect collision signal which is detected and extracted by using SD4511 particle collision noise detection equipment and is provided by the embodiment of the application, the sampling frequency is 500KHz, the signal sampling length is 51200 sampling points, and it can be seen from the figure that the foreign matter defect collision signal with low amplitude is submerged in the environmental noise and is difficult to distinguish, and the missing detection or the false detection of micro foreign matter defects is easy to cause;
Fig. 3 shows a result of denoising the signal in fig. 2 by using different signal processing methods, specifically, (a) in fig. 3 is a foreign object defect collision signal schematic diagram obtained by denoising the signal in fig. 2 by using a wavelet packet denoising method, (b) in fig. 3 is a foreign object defect collision signal schematic diagram obtained by denoising the signal in fig. 2 by using a VMD method, (c) in fig. 3 is a foreign object defect collision signal schematic diagram obtained by sparsely denoising the signal in fig. 2 by using an L1 sparse denoising method, and (d) in fig. 3 is a foreign object defect collision signal schematic diagram obtained by denoising the signal in fig. 2 by using the method provided by the application.
Based on the method for detecting the foreign matter defect of the microelectronic product provided by the embodiment, the embodiment of the application also provides a device for detecting the foreign matter defect of the microelectronic product, as shown in fig. 4, which specifically comprises:
The target penalty term construction module 10 is configured to obtain a foreign object defect collision signal of a microelectronic product to be detected, set weights for each sparse coefficient in a penalty term of a GMC sparse denoising model of the foreign object defect collision signal, and obtain a target penalty term, and specifically includes:
A signal segmentation sub-module for segmenting the foreign object defect collision signal to obtain A dimensional foreign object defect collision signal matrix;
a singular value and signal kurtosis calculation sub-module for performing singular value decomposition on the foreign object defect collision signal matrix to obtain Singular values; taking each column element in the foreign object defect collision signal matrix as a sub-signal to obtain m sub-signals, and respectively calculating the signal kurtosis of each sub-signal;
A sparse coefficient weight assignment sub-module for calculating the first Sum of singular values ofThe square root of the sum of squares of the signal kurtosis of the sub-signals is taken as the inverse of the square root of the sum of squares as the first of the sparse coding matrices of the collision signals of the foreign object defectsLine 1Weighting of column sparse coefficients; wherein, ，；
The sparse coding matrix acquisition module 20 is configured to obtain a target GMC sparse denoising model based on a fidelity term and the target penalty term; iteratively solving the target GMC sparse denoising model to obtain a sparse coding matrix of the foreign object defect collision signal;
the foreign object defect detection module 30 is configured to perform sparse representation on the foreign object defect collision signal by using the sparse coding matrix to obtain a target foreign object defect collision signal, thereby performing foreign object defect detection on the microelectronic product to be detected.
The embodiment of the application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program which realizes the steps of the method for detecting the foreign matter defects of the microelectronic product when being executed by a processor.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It is apparent that the above examples are given by way of illustration only and are not limiting of the embodiments. Other variations and modifications of the present invention will be apparent to those of ordinary skill in the art in light of the foregoing description. It is not necessary here nor is it exhaustive of all embodiments. While still being apparent from variations or modifications that may be made by those skilled in the art are within the scope of the invention.

Claims (9)

1. A method for detecting a foreign object defect of a microelectronic product, comprising:
acquiring a foreign object defect collision signal of a microelectronic product to be detected, setting weights for each sparse coefficient in a punishment item of a GMC sparse denoising model of the foreign object defect collision signal, and obtaining a target punishment item, wherein the method specifically comprises the following steps:
Dividing the foreign object defect collision signal to obtain a foreign object defect collision signal matrix in n multiplied by m dimensions;
Performing singular value decomposition on the foreign object defect collision signal matrix to obtain n singular values; taking each column element in the foreign object defect collision signal matrix as a sub-signal to obtain m sub-signals, and respectively calculating the signal kurtosis of each sub-signal;
Calculating the square sum arithmetic square root of the signal kurtosis of the jth singular value and the ith sub-signal, and taking the reciprocal of the square sum arithmetic square root as the weight of the jth row and ith column sparse coefficient in the sparse coding matrix of the foreign object defect collision signal; wherein i epsilon [1, m ], j epsilon [1, n ];
Obtaining a target GMC sparse denoising model based on the fidelity term and the target penalty term; iteratively solving the target GMC sparse denoising model to obtain a sparse coding matrix of the foreign object defect collision signal; the method specifically comprises the following steps:
Step 1: introducing a proxy variable and a Lagrangian multiplier to update the target GMC sparse denoising model, and decomposing the updated target GMC sparse denoising model into a proxy variable constraint sub-problem and a sparse coding matrix constraint sub-problem;
Step 2: setting an initial sparse coding matrix, an initial Lagrangian multiplier and the weight of each sparse coefficient in the initial sparse coding matrix, and calculating the initial value of a target GMC sparse denoising model based on the initial sparse coding matrix and the weight of each sparse coefficient in the initial sparse coding matrix;
step 3: substituting the initial sparse coding matrix, the weight of each sparse coefficient in the initial sparse coding matrix and the initial Lagrangian multiplier into the agent variable constraint subproblem, and solving a target agent variable;
step 4: substituting the target proxy variable and the initial Lagrangian multiplier into the constraint sub-problem of the sparse coding matrix, and solving a target sparse coding matrix;
step 5: the target sparse coding matrix is utilized to carry out sparsity on the foreign object defect collision signals, singular value decomposition is carried out on the sparse foreign object defect collision signals, n target singular values are obtained, and therefore the weight of each sparse coefficient in the target sparse coding matrix is calculated;
step 6: solving a target lagrangian multiplier based on the target proxy variable, the target sparse coding matrix, and the initial lagrangian multiplier;
step 7: calculating a target value of the GMC sparse denoising model based on the target sparse coding matrix and the weight of each sparse coefficient in the target sparse coding matrix, and calculating a difference value between an initial value of the target GMC sparse denoising model and the target value of the target GMC sparse denoising model;
Step 8: judging the difference value and the preset convergence error, and if the difference value is smaller than or equal to the preset convergence error, taking the target sparse coding matrix as a sparse coding matrix of a foreign object defect collision signal;
Step 9: if the difference is greater than the preset convergence error, taking the target value of the GMC sparse denoising model as an initial value of the GMC sparse denoising model, taking the target sparse coding matrix as an initial sparse coding matrix, taking the weight of each sparse coefficient in the target sparse coding matrix as the weight of each sparse coefficient in the initial sparse coding matrix, taking the target Lagrangian multiplier as an initial Lagrangian multiplier, and returning to the execution step 3 until the difference between the initial value of the GMC sparse denoising model and the target value of the GMC sparse denoising model is smaller than or equal to the preset convergence error;
and performing sparse representation on the foreign object defect collision signal by using the sparse coding matrix to obtain a target foreign object defect collision signal, so as to perform foreign object defect detection on the microelectronic product to be detected.
2. The method for detecting a foreign object defect of a microelectronic product according to claim 1, wherein the weight of the j-th row and i-th column sparse coefficients in the sparse coding matrix of the foreign object defect collision signal is calculated by the following formula:
Wherein W j,i represents the weight of the ith row and ith column sparse coefficients in the sparse coding matrix of the foreign defect collision signal; k i denotes the signal kurtosis of the ith sub-signal; sigma j represents the jth singular value;
wherein Y il represents the first data point in the ith sub-signal; k represents the number of data points in the ith sub-signal; representing the average of all data points in the ith sub-signal.
3. The method for detecting a foreign object defect in a microelectronic product according to claim 1, wherein the target GMC sparse denoising model is expressed as:
Wherein J represents a target GMC sparse denoising model; Is a fidelity term; y represents a foreign matter defect collision signal matrix; d represents a Laplace wavelet analysis dictionary; s represents a sparse coding matrix of a foreign object defect collision signal, and consists of n rows and m columns of sparse coefficients; λ represents a regularization parameter; Σ j∑iPGMC(Wj,i⊙sj,i) is a target penalty term; w j,i represents the weight of the j-th row and i-th column sparse coefficient in the sparse coding matrix of the foreign defect collision signal; s j,i represents the ith row and ith column sparse coefficients in the sparse coding matrix of the foreign object defect collision signal; the ". Iy represents Hadamard product.
4. The method for detecting a foreign object defect of a microelectronic product according to claim 1, wherein the constructing process of the fidelity term comprises:
Establishing a Laplace wavelet analysis dictionary based on the foreign object defect collision signals and the foreign object defect collision signal matrix;
obtaining a target defect collision signal matrix based on a sparse coding matrix of the foreign object defect collision signal and the Laplace wavelet analysis dictionary;
and constructing a fidelity term based on the foreign object defect collision signal matrix and the target defect collision signal matrix.
5. The method of claim 4, wherein creating a Laplace wavelet analysis dictionary based on the foreign object defect collision signal and the foreign object defect collision signal matrix comprises:
setting a value range of oscillation frequency, a value range of time shift parameters and a value range of viscous damping ratio; respectively taking values of the oscillation frequency, the time shift parameter and the viscous damping ratio in the value range of the oscillation frequency, the value range of the time shift parameter and the value range of the viscous damping ratio, constructing a Laplace wavelet atom based on each value, and obtaining a Laplace wavelet original subset based on all value combinations;
respectively calculating the similarity of each Laplace wavelet atom in the Laplace wavelet atom set and the foreign object defect collision signal;
And constructing a Laplace wavelet analysis dictionary based on the oscillation frequency, the time shift parameter, the viscous damping ratio and the Laplace wavelet atomic basis function corresponding to the Laplace wavelet atom with the maximum similarity.
6. The method for detecting a foreign object defect of a microelectronic product according to claim 5, wherein the Laplace wavelet atoms are expressed as:
Wherein ψ f,ζ,τ (t) represents a Laplace wavelet atom; f represents an oscillation frequency; ζ represents a viscous damping ratio; τ represents a time shift parameter; t represents time; w represents a wavelet support section;
the calculation formula of the similarity between Laplace wavelet atoms and the foreign object defect collision signals is as follows:
wherein C f,ζ,τ (t) represents the similarity of ψ f,ζ,τ (t) to y (t); psi f,ζ,τ (t) represents a Laplace wavelet atom; y (t) represents a foreign matter defect collision signal; II 2 denotes the Euclidean norm.
7. The method for detecting a foreign object defect of a microelectronic product according to claim 1, wherein the updated target GMC sparse denoising model is expressed as:
Wherein, L μ(s,z,β,Wi) represents the updated target GMC sparse denoising model; z represents a proxy variable; beta represents the Lagrangian multiplier; w i represents the ith column of a weight matrix W formed by weights of all sparse coefficients in the sparse coding matrix of the foreign object defect collision signal; is a fidelity term; y represents a foreign matter defect collision signal matrix; d represents a Laplace wavelet analysis dictionary; s represents a sparse coding matrix of the foreign object defect collision signal; λ represents a regularization parameter; Σ j∑iPGMC(Wj,i⊙sj,i) is a target penalty term; w j,i represents the weight of the j-th row and i-th column sparse coefficient in the sparse coding matrix of the foreign defect collision signal; s j,i represents the ith row and ith column sparse coefficients in the sparse coding matrix of the foreign object defect collision signal; the Hadamard product is indicated; μ represents a penalty parameter;
the proxy variable constraint sub-problem is expressed as:
wherein, Representing the solution vector of the proxy variable, z j,i representing the element value of the j-th row and i-th column in the solution vector of the proxy variable, and β j,i representing the element value of the j-th row and i-th column in the Lagrangian multiplier;
the sparse coding matrix constraint sub-problem is expressed as:
wherein, A solution vector representing the sparse coding matrix;
The calculation formula of the target Lagrangian multiplier is as follows:
β'＝β+γμ(s′-z')，
where β ' represents the target Lagrangian multiplier, β represents the Lagrangian multiplier, γ represents the relaxation factor, s ' represents the target sparse coding matrix, and z ' represents the target proxy variable.
8. A microelectronic product foreign object defect detection device, comprising:
The target punishment item construction module is used for acquiring a foreign object defect collision signal of the microelectronic product to be detected, setting weights for each sparse coefficient in a punishment item of a GMC sparse denoising model of the foreign object defect collision signal to obtain a target punishment item, and specifically comprises the following steps:
The signal segmentation submodule is used for segmenting the foreign object defect collision signal to obtain an n multiplied by m foreign object defect collision signal matrix;
The singular value and signal kurtosis calculation sub-module is used for carrying out singular value decomposition on the foreign object defect collision signal matrix to obtain n singular values; taking each column element in the foreign object defect collision signal matrix as a sub-signal to obtain m sub-signals, and respectively calculating the signal kurtosis of each sub-signal;
The sparse coefficient weight distribution sub-module is used for calculating the square sum arithmetic square root of the j singular value and the signal kurtosis of the i sub-signal, and taking the reciprocal of the square sum arithmetic square root as the weight of the j row and the i column sparse coefficient in the sparse coding matrix of the foreign defect collision signal; wherein i epsilon [1, m ], j epsilon [1, n ];
The sparse coding matrix acquisition module is used for acquiring a target GMC sparse denoising model based on the fidelity item and the target penalty item; iteratively solving the target GMC sparse denoising model to obtain a sparse coding matrix of the foreign object defect collision signal; the method specifically comprises the following steps:
Step 1: introducing a proxy variable and a Lagrangian multiplier to update the target GMC sparse denoising model, and decomposing the updated target GMC sparse denoising model into a proxy variable constraint sub-problem and a sparse coding matrix constraint sub-problem;
Step2: setting an initial sparse coding matrix, an initial Lagrangian multiplier and the weight of each sparse coefficient in the initial sparse coding matrix, and calculating the initial value of a target GMC sparse denoising model based on the initial sparse coding matrix and the weight of each sparse coefficient in the initial sparse coding matrix; step 3: substituting the initial sparse coding matrix, the weight of each sparse coefficient in the initial sparse coding matrix and the initial Lagrangian multiplier into the agent variable constraint subproblem, and solving a target agent variable;
step 4: substituting the target proxy variable and the initial Lagrangian multiplier into the constraint sub-problem of the sparse coding matrix, and solving a target sparse coding matrix;
step 5: the target sparse coding matrix is utilized to carry out sparsity on the foreign object defect collision signals, singular value decomposition is carried out on the sparse foreign object defect collision signals, n target singular values are obtained, and therefore the weight of each sparse coefficient in the target sparse coding matrix is calculated;
step 6: solving a target lagrangian multiplier based on the target proxy variable, the target sparse coding matrix, and the initial lagrangian multiplier;
step 7: calculating a target value of the GMC sparse denoising model based on the target sparse coding matrix and the weight of each sparse coefficient in the target sparse coding matrix, and calculating a difference value between an initial value of the target GMC sparse denoising model and the target value of the target GMC sparse denoising model;
Step 8: judging the difference value and the preset convergence error, and if the difference value is smaller than or equal to the preset convergence error, taking the target sparse coding matrix as a sparse coding matrix of a foreign object defect collision signal;
Step 9: if the difference is greater than the preset convergence error, taking the target value of the GMC sparse denoising model as an initial value of the GMC sparse denoising model, taking the target sparse coding matrix as an initial sparse coding matrix, taking the weight of each sparse coefficient in the target sparse coding matrix as the weight of each sparse coefficient in the initial sparse coding matrix, taking the target Lagrangian multiplier as an initial Lagrangian multiplier, and returning to the execution step 3 until the difference between the initial value of the GMC sparse denoising model and the target value of the GMC sparse denoising model is smaller than or equal to the preset convergence error;
And the foreign object defect detection module is used for carrying out sparse representation on the foreign object defect collision signal by utilizing the sparse coding matrix to obtain a target foreign object defect collision signal, so as to carry out foreign object defect detection on the microelectronic product to be detected.
9. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of the method for detecting a foreign object defect of a microelectronic product according to any of claims 1-7.
CN202410986735.XA 2024-07-23 2024-07-23 A method, device and readable storage medium for detecting foreign matter defects in microelectronic products Active CN118520227B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410986735.XA CN118520227B (en) 2024-07-23 2024-07-23 A method, device and readable storage medium for detecting foreign matter defects in microelectronic products

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410986735.XA CN118520227B (en) 2024-07-23 2024-07-23 A method, device and readable storage medium for detecting foreign matter defects in microelectronic products

Publications (2)

Publication Number Publication Date
CN118520227A CN118520227A (en) 2024-08-20
CN118520227B true CN118520227B (en) 2024-11-15

Family

ID=92277114

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410986735.XA Active CN118520227B (en) 2024-07-23 2024-07-23 A method, device and readable storage medium for detecting foreign matter defects in microelectronic products

Country Status (1)

Country Link
CN (1) CN118520227B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110889458A (en) * 2019-12-03 2020-03-17 广东工业大学 Injection molding product defect classification method based on machine learning
CN117237314A (en) * 2023-09-27 2023-12-15 江苏恒冉新材料有限公司 PVC floor defect detection method, device and system based on machine vision

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116226627B (en) * 2023-03-08 2023-09-22 安徽大学 Lorenz constraint angle estimation method and system in non-Gaussian environment
CN117786322A (en) * 2023-12-28 2024-03-29 江南大学 A flip-chip vibration signal denoising method and system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110889458A (en) * 2019-12-03 2020-03-17 广东工业大学 Injection molding product defect classification method based on machine learning
CN117237314A (en) * 2023-09-27 2023-12-15 江苏恒冉新材料有限公司 PVC floor defect detection method, device and system based on machine vision

Also Published As

Publication number Publication date
CN118520227A (en) 2024-08-20

Similar Documents

Publication Publication Date Title
Zhan et al. Bridge surface roughness identification based on vehicle–bridge interaction
Alamdari et al. FRF-based damage localization method with noise suppression approach
CN109086555B (en) Simulation method for simulating energy spectrum curve by adopting Monte Carlo method
JP7646458B2 (en) Identification of arrays in semiconductor samples
CN117274533A (en) A three-dimensional geometric model construction method and performance prediction method for porous rocks
CN118520227B (en) A method, device and readable storage medium for detecting foreign matter defects in microelectronic products
CN117786322A (en) A flip-chip vibration signal denoising method and system
Favata et al. Realistic non-destructive testing of integrated circuit bond wiring using 3-D X-ray tomography, reverse engineering, and finite element analysis
Wang et al. A methodology for the multi-resolution simulation of grinding wheel surface
Dong et al. Incipient bearing fault feature extraction based on minimum entropy deconvolution and K-singular value decomposition
CN110427918A (en) A kind of Fault Classification and readable storage medium storing program for executing of electronic type electric power mutual-inductor
CN114035021A (en) Circuit fault prediction method based on EEMD-Prophet
CN115169423B (en) Stamping signal processing method, device, equipment and readable storage medium
US20250200719A1 (en) Method and system for reducing charging artifact in inspection image
Kashiwagi et al. Digital systems design based on DSP algorithms in FPGA for fault identification in rotary machines
Altay et al. Signal Processing for Acoustic Emission Signature Analysis and Defect Detection
CN117171510A (en) A method and system for denoising white noise in partial discharge detection
Makieła Problems of Selecting the Wavelet Transform Parameters in the Aspect of Surface Texture Analysis
Prins et al. End-member modelling of grain-size distributions of sediment mixtures
CN110716228B (en) Method for extracting characteristic signal for rotary type redundancy detection
CN111967799A (en) Method for identifying materials and process problems of integrated parameter table-attached circulator
Oneida et al. Flaw characterization using inversion of eddy current response and the effect of filters and scan resolution
Valpiani et al. Numerical modeling of Fluid-Structure Interactione of a 3D wedge during water impact with variation of velocity and pitch angle
Lunin Phenomenological and algorithmic methods for the solution of inverse problems of electromagnetic testing
CN114239294B (en) K distribution clutter parameter estimation method and device based on origin moment partial derivative

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