CN115100168A - System and method for detecting sub-surface defects under wafer back sealing film - Google Patents
System and method for detecting sub-surface defects under wafer back sealing film Download PDFInfo
- Publication number
- CN115100168A CN115100168A CN202210794689.4A CN202210794689A CN115100168A CN 115100168 A CN115100168 A CN 115100168A CN 202210794689 A CN202210794689 A CN 202210794689A CN 115100168 A CN115100168 A CN 115100168A
- Authority
- CN
- China
- Prior art keywords
- image
- wafer
- defect
- detection
- detected
- 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.)
- Pending
Links
- 230000007547 defect Effects 0.000 title claims abstract description 98
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000007789 sealing Methods 0.000 title claims abstract description 18
- 238000001514 detection method Methods 0.000 claims abstract description 84
- 230000000007 visual effect Effects 0.000 claims abstract description 17
- 238000007599 discharging Methods 0.000 claims abstract description 14
- 230000033001 locomotion Effects 0.000 claims abstract description 6
- 238000005286 illumination Methods 0.000 claims description 26
- 230000007246 mechanism Effects 0.000 claims description 20
- 238000003384 imaging method Methods 0.000 claims description 19
- 230000003287 optical effect Effects 0.000 claims description 19
- 238000007781 pre-processing Methods 0.000 claims description 18
- 239000013598 vector Substances 0.000 claims description 17
- 239000010408 film Substances 0.000 claims description 15
- 238000003709 image segmentation Methods 0.000 claims description 15
- 230000011218 segmentation Effects 0.000 claims description 15
- 238000000605 extraction Methods 0.000 claims description 14
- 239000013039 cover film Substances 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims description 10
- 230000009466 transformation Effects 0.000 claims description 10
- 238000006073 displacement reaction Methods 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000009826 distribution Methods 0.000 claims description 8
- 238000009499 grossing Methods 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000007797 corrosion Effects 0.000 claims description 3
- 238000005260 corrosion Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 230000000873 masking effect Effects 0.000 claims description 3
- 125000004122 cyclic group Chemical group 0.000 claims description 2
- 230000002708 enhancing effect Effects 0.000 claims description 2
- 238000012423 maintenance Methods 0.000 claims description 2
- 230000000877 morphologic effect Effects 0.000 claims description 2
- 238000004088 simulation Methods 0.000 claims description 2
- 238000013519 translation Methods 0.000 claims description 2
- 235000012431 wafers Nutrition 0.000 description 78
- 238000012706 support-vector machine Methods 0.000 description 7
- 238000012549 training Methods 0.000 description 7
- 238000013461 design Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 239000000523 sample Substances 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- XUIMIQQOPSSXEZ-UHFFFAOYSA-N Silicon Chemical compound [Si] XUIMIQQOPSSXEZ-UHFFFAOYSA-N 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 229910052710 silicon Inorganic materials 0.000 description 2
- 239000010703 silicon Substances 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 238000004626 scanning electron microscopy Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20036—Morphological image processing
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/30148—Semiconductor; IC; Wafer
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a system and a method for detecting the defect of the sub-surface under a wafer back sealing film, wherein the detection system comprises: the detection black box is used for placing the detection platform and the detection terminal and shielding the influence of external environment noise and stray light on a detection result; the detection platform is used for moving the wafer to be detected to a preset detection position and detecting the defect of the sub-surface under the back sealing film of the wafer; the detection terminal is used for controlling the feeding and discharging movement of the wafer to be detected and feeding back the detection result of the sub-surface defects under the back sealing film of the wafer in a visual form; the invention has the advantages of high detection speed and high defect identification rate.
Description
The technical field is as follows:
the invention relates to the technical field of wafer detection, in particular to a system and a method for detecting sub-surface defects under a wafer back sealing film.
Background art:
wafer defect detection is crucial in semiconductor design, production, packaging, testing and other processes, and is the key to ensure product yield. The wafer defects can be divided into defects on the upper surface of the back sealing film and defects on the lower sub-surface of the back sealing film according to the characteristics of the production process. Existing surface defect detection techniques are classified into contact and non-contact, and the non-contact is classified into an imaging type and a non-imaging type. The contact type surface detection transmits the topography information of the object surface to a computer for processing through a probe and a sensor, but the method is easy to damage the surface of a detected sample and is not suitable for precise wafer detection. The imaging inspection methods commonly used in the semiconductor industry at present are mainly two methods, Automatic Optical Inspection (AOI) and Scanning Electron Microscopy (SEM). The AOI detection system is used for illuminating a detected object by designing an illumination system, imaging the detected object by using an imaging system, transmitting collected optical information of the detected object by using a sensor, converting the collected optical information into digital signals and finally delivering the digital signals to a computer system for subsequent processing. The AOI surface detection technology has the advantages of high detection speed, high flexibility, low cost, large field of view and the like. But the scheme still cannot solve the problem of detecting the defect of the sub-surface under the back sealing film of the wafer.
The invention content is as follows:
the invention overcomes the defects and provides the system for detecting the defects of the lower sub-surface of the wafer back cover film, and the scheme has important significance for detecting the defects of the lower sub-surface of the wafer back cover film with complex imaging conditions and more interference factors and can meet the requirements of related applications. The invention designs a system for detecting the defects of the lower sub-surface of a wafer back seal film, which solves the problems of the background technology and makes up the vacancy of the related technology.
In order to solve the technical problem, the technical scheme adopted by the invention is as follows:
a system for detecting defects on a sub-surface under a wafer back cover film, the system comprising:
the detection black box is used for placing the detection platform and the detection terminal and shielding the influence of external environment noise and stray light on a detection result; the detection platform is used for moving the wafer to be detected to a preset detection position and detecting the defects of the sub-surface under the back sealing film of the wafer; the detection terminal is used for controlling the feeding and discharging movement of the wafer to be detected and feeding back a detection result of the defect of the sub-surface under the back sealing film of the wafer in a visual form; wherein:
the detection black box comprises a movable roller, a mechanical arm bracket and an operation door; the movable roller is used for detecting the portable movement of the black box, the mechanical arm support is used for fixing the display and can perform multi-degree-of-freedom translation and rotation, and the operation door can be freely opened and closed and maintains and replaces internal detection equipment;
the detection platform comprises a visual imaging mechanism, a feeding and discharging mechanism and an optical platform;
the detection terminal comprises an industrial control computer, a display, a stepping motor controller and an operation input device; wherein:
the vision imaging mechanism is used for acquiring an image of the wafer to be detected and enhancing the contrast between the subsurface defect under the back sealing film and the background by a machine vision technology; the loading and unloading mechanism is used for moving the wafer to be detected to a preset detection position and moving the wafer to a preset loading position after detection is finished; the optical platform is used for bearing a visual imaging mechanism and a feeding and discharging mechanism, and the visual imaging mechanism comprises an industrial camera, an optical lens, a high-focus parallel light source and an angular displacement table; the feeding and discharging mechanism comprises a stepping motor, a wafer bearing platform and a linear module;
the industrial control computer is used for receiving the wafer image input by the visual imaging mechanism and outputting a detection result through the operation of a pre-designed detection unit; the display is used for visually carrying out software operation and maintenance and feeding back a defect detection result; the stepping motor controller controls the rotating speed of the stepping motor and the displacement distance of the feeding and discharging mechanism according to preset parameters of software; the operation input equipment is used for identifying user operation and inputting the operation into the industrial control computer for operation; the detection unit comprises a preprocessing image module, a wafer defect vector extraction module and a wafer defect feature identification module.
In order to solve the problems of the prior art, the invention can also adopt the following technical scheme:
a method for detecting the defects of the lower sub-surface of the wafer back cover film by adopting a system for detecting the defects of the lower sub-surface of the wafer back cover film comprises the following steps:
acquiring a to-be-detected image of a to-be-detected wafer;
the image preprocessing module is used for preprocessing the image to be detected to obtain a preprocessed image;
the wafer defect vector extraction module is used for extracting defect features of the preprocessed image to obtain a defect feature vector;
the wafer defect feature identification module carries out defect classifier threshold discrimination on the defect feature vector to obtain a wafer defect identification result to be detected; wherein:
the wafer defect vector extraction module for extracting the defect characteristics of the preprocessed image comprises the following steps:
performing target area and wafer background segmentation on the preprocessed image based on an iterative threshold method, and calculating the maximum gray value Z max And the minimum gray value Z min And generates an initial threshold value
According to a threshold value M k Calculating the average gray level M of the target and the background 1 And M 2 Cyclic calculation of
Up to M k And M k+1 The difference is less than the parameter threshold or reaches the preset iteration number according to the optimal segmentation threshold M k+1 Dividing to obtain a defect target area;
performing morphological filtering based on corrosion and expansion on the defect target area, setting corresponding structural elements to extract n connected components from the defect target area, and reducing the n connected components to a color original image based on connected component contour coordinate information to obtain a defect to-be-identified area;
extracting position, area, roundness and second-order invariant moment M from the region to be identified of the defect 1 ＝η 20 +η 02 Second order invariant moment
Third-order moment-invariant M 3 ＝(η 30 -3η 12 ) 2 +(3η 21 -η 03 ) 2 And obtaining 8-dimensional characteristic vectors of the defects through the characteristic information of the chromaticity region growth and the color moment.
Further, the image preprocessing module preprocesses the image to be detected, and comprises the following steps:
performing illumination compensation on the image to be detected to obtain an illumination compensation image;
carrying out image segmentation and area division on the illumination compensation image to obtain a target detection area;
and performing smooth filtering and image enhancement on the target detection area to obtain a preprocessed image.
Further, the image preprocessing module performs illumination compensation on the image to be detected to obtain an illumination compensation image: the preprocessing image module carries out global gray maximum value detection on an input image to obtain n local gray maximum values, determines corresponding coordinates of a light source center in the input image according to a light source center self-adaptive discrimination algorithm, simultaneously records gray values of corner points and edges of the image, synthesizes the coordinates of the light source center, the gray values of the corner points and the gray values of the edges to generate an adaptive illumination compensation template, convolves the input image with the illumination compensation template and outputs the convolved image, and the illumination compensation function is realized.
Further, the preprocessing image module performs image segmentation and area division on the illumination compensation image to obtain a target detection area:
image segmentation is carried out on the acquired image: calculating a gray histogram of an input image according to gray distribution of a foreground object and a background, which obeys normal distribution, and searching a segmentation boundary of a normal group, so that an image segmentation threshold is generated in a self-adaptive manner, automatic segmentation of a wafer and a redundant background is realized, meanwhile, the edge of the wafer is extracted and fitted based on the least square principle, the boundary information of a wafer area is output, and the image segmentation function is realized;
the method comprises the steps of carrying out region division on an obtained image, predefining a wafer template according to the shape and the size of a wafer, calculating an angular point coordinate aiming at an input image, calculating an output transformation matrix and a simulation matrix according to the angular point coordinate, mapping the divided regions on the wafer image based on affine transformation, outputting boundary information of each region in the wafer, and realizing the function of region division.
Further, the image preprocessing module performs smooth filtering and image enhancement on the target detection area to obtain a preprocessed image:
performing Sobel gradient processing on an input image, performing further smoothing filtering by using a smoothing filter, multiplying a Laplacian image and the smoothed gradient image to realize template masking, and summing the masked template and the input image to obtain a sharpened image; and carrying out gray scale transformation on the sharpened image by utilizing power law transformation in order to expand the gray scale dynamic range of the image, outputting the enhanced wafer image and realizing the image enhancement function.
Has the advantages that:
the invention provides a system and a method for detecting the defect of the lower sub-surface of a wafer back sealing film, which have the advantages of high detection speed and high defect identification rate. The device can solve the problem that the defect detection of the sub-surface under the existing wafer back sealing film is difficult.
Description of the drawings:
FIG. 1 is a schematic diagram of a system for detecting defects on a sub-surface under a wafer back cover film based on machine vision;
FIG. 2 is a front view of an in-black box apparatus;
FIG. 3 is a top view of an optical bench surface unit;
FIG. 4 is a flow chart of a detection method;
FIG. 5 is a flowchart of an illumination compensation algorithm;
FIG. 6 is a flow chart of an image segmentation algorithm;
FIG. 7 is a flow chart of a region partitioning algorithm;
FIG. 8 is a flow chart of an image enhancement algorithm;
FIG. 9 is a flow chart of a feature extraction algorithm;
FIG. 10 is a flow diagram of a support vector machine based classifier design.
Detailed Description
The invention provides a system and a method for detecting defects of a sub-surface under a wafer back-sealing film based on machine vision, and the implementation process of the invention is further described in detail with reference to fig. 1 to 10.
FIG. 1 is a schematic view of the apparatus of the present invention. The construction of the device is described in detail below.
The device mainly comprises a detection black box 1, a feeding and discharging outlet 2, a linear module 3, a wafer bearing platform 4, a movable roller 5, a display 6, an operation door 7, a mechanical arm support 8, an industrial camera 9, an optical lens 10, a high-focus parallel light source 11, an angular displacement table 12, an optical platform 13, an industrial control computer 14, a stepping motor controller 15, a stepping motor 16 and the like.
The front surface of the detection black box 1 is provided with a feeding and discharging outlet 2, a linear module 3 extends out of the feeding and discharging outlet 2, the surface of the linear module 3 is fixedly connected with a wafer bearing platform 4, the bottom of the detection black box 1 is fixedly connected with a movable roller 5, the front surface of the detection black box 1 is fixedly connected with a mechanical arm support 8, the free end of the mechanical arm support 8 is fixedly connected with a display 6, the side surface of the detection black box 1 is provided with an operation door 7, the detection black box 1 is internally and fixedly connected with an industrial camera 9, an optical lens 10 is fixedly connected below the industrial camera 9, the detection black box 1 is internally provided with an optical platform 13, the surface of the optical platform 13 is fixedly connected with the linear module 3, the free end of the linear module 3 is fixedly connected with a stepping motor 16, the surface of the optical platform 13 is fixedly connected with 4 symmetrical angular displacement platforms 12, and the surface of each angular displacement platform 12 is fixedly connected with a high-focus parallel light source 11, an industrial control computer 14 is fixedly connected in the detection black box 1, and a stepping motor controller 15 is fixedly connected in the detection black box 1.
The specific implementation process of the device function in the present invention is discussed in detail below.
The construction of the device of the invention is completed first. Opening the operation door 7 on the side surface of the detection black box 1 to debug the internal device, as shown in fig. 2, placing the wafer bearing platform 4 at a preset detection position, adjusting the position of the industrial camera 9 and the optical lens 10 fixedly connected with the industrial camera until the wafer bearing platform 4 appears in the center of the visual field of the visual imaging system, adjusting the focusing ring and the zooming ring of the optical lens 10 until the wafer bearing platform 4 is positioned at the focal plane of the visual imaging system and occupies a field area of 3/4 above, setting 4 angles of the angular displacement platform 12 until the high-focusing parallel light source 11 fixedly connected with the surface of the angular displacement platform realizes uniform illumination on the wafer bearing platform 4, adjusting the sensitivity of the industrial camera 9 and the aperture of the optical lens 10 until the target to be detected in the visual field is imaged clearly and has a high contrast with the background, and completing the device debugging process.
The wafer to be detected is placed in a groove on the surface of the wafer bearing platform 4, as shown in fig. 3, the industrial control computer 14 receives a feeding instruction and then sends the feeding instruction to the stepping motor controller 15, the stepping motor controller 15 controls the stepping motor 16 to execute command feeding, and the stepping motor 16 drives the linear module 3 to perform single-axis motion, so that the wafer bearing platform 4 fixedly connected with the surface of the linear module 3 moves to a preset detection position from an initial feeding position through the feeding and discharging outlet 2, and the feeding function is realized.
The wafer to be detected moves to a preset detection position, the industrial control computer 14 sends an image acquisition command to the industrial camera 9, the visual imaging system finishes image acquisition and then sends an acquired image to the industrial control computer 14, as shown in fig. 4, the industrial control computer 14 performs illumination compensation, image segmentation and area division, smoothing filtering and image enhancement on the acquired image to finish an image preprocessing function, performs defect feature extraction on the preprocessed image to obtain a defect feature vector, performs defect classifier threshold discrimination on the defect feature vector to obtain a wafer defect identification result to be detected, and performs visual display on the wafer defect identification result to be detected by the display 6 to realize a defect detection function.
After the wafer defect identification result to be detected is obtained, the industrial control computer 14 receives the blanking instruction and then sends the blanking instruction to the stepper motor controller 15, the wafer bearing platform 4 returns to the initial feeding position, and an operator replaces the wafer to be detected, so that the blanking function is realized.
The above industrial control computer 14 performs illumination compensation on the acquired image, including: as shown in fig. 5, global maximum gray value detection is performed on an input image to obtain n local maximum gray values, corresponding coordinates of a light source center in the input image are determined according to a light source center adaptive discrimination algorithm, gray values of corner points and edges of the image are recorded at the same time, an adaptive illumination compensation template is generated by integrating the light source center coordinates, the gray values of the corner points and the gray values of the edges, the input image and the illumination compensation template are convolved, and the convolved image is output to realize an illumination compensation function.
The industrial control computer 14 performs image segmentation on the acquired image, and includes: as shown in fig. 6, according to the principle that the gray distribution of the foreground object and the background obeys normal distribution, the gray histogram of the input image is calculated, and the segmentation limit of the normal population is found, so that the image segmentation threshold is adaptively generated, automatic segmentation of the wafer and the redundant background is realized, in order to improve the segmentation effect, the wafer edge is extracted and fitted based on the least square principle, and the boundary information of the wafer region is output, so that the image segmentation function is realized.
The industrial control computer 14 performs area division on the acquired image, including: as shown in fig. 7, a wafer template is predefined according to the shape and size of a wafer, corner coordinates are calculated for an input image, an output transformation matrix and a pseudo matrix are calculated according to the corner coordinates, a divided region is mapped on the wafer image based on affine transformation, and boundary information of each region in the wafer is output, so that a region division function is realized.
The above industrial control computer 14 performs image enhancement on the acquired image, including: as shown in fig. 8, to remove the noise enhancement introduced by the laplacian method, first, Sobel gradient processing is performed on the input image, smoothing filtering is performed by using a smoothing filter, the laplacian image is multiplied by the smoothed gradient image to realize template masking, a sharpened image is obtained by summing the masked template and the input image, gray scale conversion is performed on the sharpened image by using power law conversion to expand the gray scale dynamic range of the image, and the enhanced wafer image is output to realize the image enhancement function.
The industrial control computer 14 extracts the defect features of the preprocessed image to obtain a defect feature vector, as shown in fig. 9, and includes the following steps:
and (3) segmenting a target area and a silicon wafer background: an iterative threshold method is adopted, the principle is to approximate probability density functions of two or more normal distributions of a gray level histogram, the threshold is selected as a gray level value corresponding to the minimum probability position between the maximum values of the two or more normal distributions, and the result is segmentation with the minimum error.
Calculating the maximum gray value Z of the image area max And the minimum gray value Z min An initial threshold is defined as
According to the threshold value M k Dividing the image into target and background parts and calculating average gray scale M 1 And M 2 According to
Calculating new threshold, and circulating operation until M k And M k+1 The difference is less than the predefined parameter or reaches a certain number of iterations to obtain the optimal segmentation threshold M k+1 Completing the segmentation of the target area and the wafer background according to the optimal segmentation threshold;
extracting a region to be identified: removing residual components of a wafer background by adopting corrosion and expansion operations, setting appropriate structural elements to extract n connected components from a target area, recording element coordinates in the connected components, and reducing the n connected components to n areas to be identified in a color original image according to the coordinates;
characteristic extraction: aiming at n regions to be identified, position, area and roundness information are extracted according to a geometric feature extraction principle, and a second-order moment-invariant M is extracted according to a moment feature extraction principle 1 Second-order moment-invariant M 2 Third-order moment-invariant M 3 Extracting information of color area growth and color moment according to a color feature extraction principle, storing n pieces of feature information of the area to be identified in n 8-dimensional vectors, and finally realizing the feature of the wafer defectsAnd (5) sign extraction.
The industrial control computer 14 performs defect classifier threshold discrimination on the defect feature vector to obtain a wafer defect identification result to be detected, and includes the following steps:
the classifier is based on: the classifier design is based on the support vector machine principle, the support vector machine belongs to a linear classifier, and the basic principle is to construct a classification hyperplane and divide a classification space into two parts; the support vector machine calculates an optimized classification hyperplane through input training sample data to enable the distance between the training data and the hyperplane to be maximum; for nonlinear inputs, classification can be done by constructing a mapping to a dimensional space;
designing a classifier: as shown in fig. 10, n classifiers are designed for n types of silicon wafer defects based on the support vector machine technology, and the input image is subjected to feature extraction and then determined by the support vector machine to give a classification result; selecting Gaussian radial basis functions
As a support vector machine kernel function, the feature space of the Gaussian radial basis function is an infinite-dimension Hilbert space, the problem of integer programming of the maximum interval can be solved in the infinite-dimension space, and meanwhile, the calculation performance is good; the SVM classifier parameters are a penalty constant C and a kernel function intrinsic parameter sigma, the optimization selection of the parameters is respectively carried out on each classifier by adopting a grid search method, firstly, rough grid division is carried out, the parameters are selected from grids, the classification effect of the optimal parameter combination and the suboptimal parameter combination in the step is compared, if the optimal parameter combination is smaller than the set precision, the optimal parameter combination is selected, the parameter optimization is stopped, and otherwise, the optimal grids are further subdivided;
and (3) operation and test: selecting a certain number of defect samples with typical defect characteristics as training samples, setting the system into an off-line training mode, respectively training various types of defects, reading in a defect wafer and a defect area by the system, analyzing and recording the characteristic value of the defect area according to the selected defect type, performing parameter training after the training samples of the type are read, and converting the characteristic value of the defect into classifier parameters by a corresponding defect classifier for online detection; randomly selecting a certain number of wafers on a production line for testing, loading the wafers, automatically operating the system, reading wafer information, giving out a final classification result, comparing the final classification result with a result given manually, calculating the final accuracy according to the system accuracy evaluation index, and counting the discrimination performance by software to finally realize the wafer defect detection and classification functions.
In summary, the system and the method for detecting the defect of the sub-surface under the wafer back cover film based on the machine vision can solve the problem that the defect of the sub-surface under the wafer back cover film is difficult to detect in the prior art, and have the advantages of high detection speed and high defect identification rate.
The present invention is not limited to the above-described embodiments. The foregoing description of the specific embodiments is intended to describe and illustrate the technical solutions of the present invention, and the above specific embodiments are merely illustrative and not restrictive. Those skilled in the art can make many changes and modifications to the invention without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (6)
1. A system for detecting defects on a subsurface of a wafer back cover film, the system comprising:
the detection black box is used for placing the detection platform and the detection terminal and shielding the influence of external environment noise and stray light on a detection result; the detection platform is used for moving the wafer to be detected to a preset detection position and detecting the defect of the sub-surface under the back sealing film of the wafer; the detection terminal is used for controlling the feeding and discharging movement of the wafer to be detected and feeding back the detection result of the sub-surface defects under the back sealing film of the wafer in a visual form; wherein:
the detection black box comprises a movable roller, a mechanical arm bracket and an operation door; the movable roller is used for detecting the portable movement of the black box, the mechanical arm support is used for fixing the display and can perform multi-degree-of-freedom translation and rotation, and the operation door can be freely opened and closed and maintains and replaces internal detection equipment;
the detection platform comprises a visual imaging mechanism, a feeding and discharging mechanism and an optical platform;
the detection terminal comprises an industrial control computer, a display, a stepping motor controller and an operation input device; wherein:
the vision imaging mechanism is used for acquiring an image of the wafer to be detected and enhancing the contrast between the sub-surface defect under the back sealing film and the background by a machine vision technology; the loading and unloading mechanism is used for moving the wafer to be detected to a preset detection position and moving the wafer to a preset loading position after detection is finished; the optical platform is used for bearing a visual imaging mechanism and a loading and unloading mechanism, and the visual imaging mechanism comprises an industrial camera, an optical lens, a high-focus parallel light source and an angular displacement table; the feeding and discharging mechanism comprises a stepping motor, a wafer bearing platform and a linear module;
the industrial control computer is used for receiving the wafer image input by the visual imaging mechanism and outputting a detection result through the operation of a pre-designed detection unit; the display is used for visually carrying out software operation and maintenance and feeding back a defect detection result; the stepping motor controller controls the rotating speed of the stepping motor and the displacement distance of the feeding and discharging mechanism according to preset parameters of software; the operation input equipment is used for identifying user operation and inputting the operation into the industrial control computer for operation; the detection unit comprises a preprocessing image module, a wafer defect vector extraction module and a wafer defect feature identification module.
2. The method for detecting the defects of the sub-surface under the wafer back cover film by using the system as claimed in claim 1, which comprises the following steps:
acquiring a to-be-detected image of a to-be-detected wafer;
the image preprocessing module is used for preprocessing the image to be detected to obtain a preprocessed image;
the wafer defect vector extraction module is used for extracting defect features of the preprocessed image to obtain a defect feature vector;
the wafer defect feature identification module carries out defect classifier threshold discrimination on the defect feature vector to obtain a wafer defect identification result to be detected; wherein:
the wafer defect vector extraction module for extracting the defect characteristics of the preprocessed image comprises the following steps:
performing target area and wafer background segmentation on the preprocessed image based on an iterative threshold method, and calculating the maximum gray value Z max And the minimum gray value Z min And generating an initial threshold value
According to a threshold value M k Calculating the average gray level M of the target and the background 1 And M 2 Cyclic calculation of
Up to M k And M k+1 The difference is less than the parameter threshold or reaches the preset iteration number, and the optimal segmentation threshold M is used k+1 Dividing to obtain a defect target area;
performing morphological filtering based on corrosion and expansion on the defect target area, setting corresponding structural elements to extract n connected components from the defect target area, and reducing the n connected components to a color original image based on connected component contour coordinate information to obtain a defect to-be-identified area;
extracting position, area, roundness and second-order invariant moment M from the region to be identified of the defect 1 ＝η 20 +η 02 Second order invariant moment
Third-order moment-invariant M 3 ＝(η 30 -3η 12 ) 2 +(3η 21 -η 03 ) 2 And obtaining 8-dimensional characteristic vectors of the defects through the characteristic information of the chromaticity region growth and the color moment.
3. The method for detecting the defect on the sub-surface under the wafer back cover film according to claim 2, wherein the pre-processing image module pre-processes the image to be detected, and comprises the following steps:
performing illumination compensation on the image to be detected to obtain an illumination compensation image;
performing image segmentation and area division on the illumination compensation image to obtain a target detection area;
and performing smooth filtering and image enhancement on the target detection area to obtain a preprocessed image.
4. The method for detecting the defect of the sub-surface under the wafer back cover film according to claim 3, wherein the pre-processing image module performs illumination compensation on the image to be detected to obtain an illumination compensation image: the preprocessing image module detects global gray maximum values of an input image to obtain n local gray maximum values, determines corresponding coordinates of a light source center in the input image according to a light source center self-adaptive discrimination algorithm, records gray values of corner points and edges of the image, integrates the light source center coordinates, the gray values of the corner points and the gray values of the edges to generate a self-adaptive illumination compensation template, convolves the input image and the illumination compensation template and outputs the convolved image, and the illumination compensation function is achieved.
5. The method as claimed in claim 3, wherein the pre-processing image module performs image segmentation and area division on the illumination compensation image to obtain a target detection area:
image segmentation is carried out on the acquired image: calculating a gray level histogram of an input image according to gray level distribution of a foreground object and a background and obeying normal distribution, and searching a segmentation boundary of a normal group, so that an image segmentation threshold is generated in a self-adaptive manner, automatic segmentation of a wafer and a redundant background is realized, meanwhile, the edge of the wafer is extracted and fitted based on the least square principle, the boundary information of a wafer area is output, and the image segmentation function is realized;
the method comprises the steps of carrying out region division on an obtained image, predefining a wafer template according to the shape and the size of a wafer, calculating an angular point coordinate aiming at an input image, calculating an output transformation matrix and a simulation matrix according to the angular point coordinate, mapping the divided regions on the wafer image based on affine transformation, outputting boundary information of each region in the wafer, and realizing the function of region division.
6. The method as claimed in claim 3, wherein the pre-processing image module performs smoothing filtering and image enhancement on the target detection region to obtain a pre-processing image:
performing Sobel gradient processing on an input image, performing further smoothing filtering by using a smoothing filter, multiplying a Laplacian image and the smoothed gradient image to realize template masking, and summing the masked template and the input image to obtain a sharpened image; and carrying out gray scale transformation on the sharpened image by utilizing power law transformation in order to expand the gray scale dynamic range of the image, outputting the enhanced wafer image and realizing the image enhancement function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210794689.4A CN115100168A (en) | 2022-07-07 | 2022-07-07 | System and method for detecting sub-surface defects under wafer back sealing film |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210794689.4A CN115100168A (en) | 2022-07-07 | 2022-07-07 | System and method for detecting sub-surface defects under wafer back sealing film |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115100168A true CN115100168A (en) | 2022-09-23 |
Family
ID=83296286
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210794689.4A Pending CN115100168A (en) | 2022-07-07 | 2022-07-07 | System and method for detecting sub-surface defects under wafer back sealing film |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115100168A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115642103A (en) * | 2022-12-22 | 2023-01-24 | 昂坤视觉(北京)科技有限公司 | Optical detection equipment and system |
| CN116045827A (en) * | 2023-02-22 | 2023-05-02 | 无锡星微科技有限公司 | System and method for detecting thickness and bending degree of large-size wafer |
| CN117252876A (en) * | 2023-11-17 | 2023-12-19 | 江西斯迈得半导体有限公司 | LED support defect detection method and system |
| CN117576092A (en) * | 2024-01-15 | 2024-02-20 | 成都瑞迪威科技有限公司 | Wafer component counting method based on image processing |
-
2022
- 2022-07-07 CN CN202210794689.4A patent/CN115100168A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115642103A (en) * | 2022-12-22 | 2023-01-24 | 昂坤视觉(北京)科技有限公司 | Optical detection equipment and system |
| CN116045827A (en) * | 2023-02-22 | 2023-05-02 | 无锡星微科技有限公司 | System and method for detecting thickness and bending degree of large-size wafer |
| CN116045827B (en) * | 2023-02-22 | 2023-11-10 | 无锡星微科技有限公司 | System and method for detecting thickness and bending degree of large-size wafer |
| CN117252876A (en) * | 2023-11-17 | 2023-12-19 | 江西斯迈得半导体有限公司 | LED support defect detection method and system |
| CN117252876B (en) * | 2023-11-17 | 2024-02-09 | 江西斯迈得半导体有限公司 | LED support defect detection method and system |
| CN117576092A (en) * | 2024-01-15 | 2024-02-20 | 成都瑞迪威科技有限公司 | Wafer component counting method based on image processing |
| CN117576092B (en) * | 2024-01-15 | 2024-03-29 | 成都瑞迪威科技有限公司 | Wafer component counting method based on image processing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115100168A (en) | System and method for detecting sub-surface defects under wafer back sealing film | |
| Ren et al. | State of the art in defect detection based on machine vision | |
| CN108520274B (en) | High-reflectivity surface defect detection method based on image processing and neural network classification | |
| CN109978839B (en) | Method for detecting wafer low-texture defects | |
| CN110766684B (en) | Stator surface defect detection system and detection method based on machine vision | |
| US20170357895A1 (en) | Method of deep learining-based examination of a semiconductor specimen and system thereof | |
| CN111292305A (en) | Improved YOLO-V3 metal processing surface defect detection method | |
| KR20210069736A (en) | Systems and methods for determining the type and size of defects on a blank reticle | |
| US9846929B2 (en) | Fast density estimation method for defect inspection application | |
| CN110443791B (en) | Workpiece detection method and device based on deep learning network | |
| KR20220014805A (en) | Generating training data usable for examination of a semiconductor specimen | |
| CN115661110B (en) | Transparent workpiece identification and positioning method | |
| CN116664565A (en) | Hidden crack detection method and system for photovoltaic solar cell | |
| CN110426395B (en) | Method and device for detecting surface of solar EL battery silicon wafer | |
| CN113705564B (en) | Pointer type instrument identification reading method | |
| Gu et al. | A detection and identification method based on machine vision for bearing surface defects | |
| Kamani et al. | Car body paint defect inspection using rotation invariant measure of the local variance and one-against-all support vector machine | |
| CN117351314A (en) | Glass bottle defect identification method and system | |
| CN114120318B (en) | Dark field image target point accurate extraction method based on integrated decision tree | |
| Zhang et al. | Research on chips’ defect extraction based on image-matching | |
| US11961221B2 (en) | Defect examination on a semiconductor specimen | |
| Ding et al. | An effective on-line surface particles inspection instrument for large aperture optical element | |
| Wang et al. | An accurate method for the extraction of line structured light stripe. | |
| CN117934453B (en) | Method and system for diagnosing defects of backlight foreign matters of mobile phone screen | |
| CN113947561B (en) | Image segmentation of semiconductor samples |
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 |