CN108231645B - High-precision positioning method and device in wafer-level flip vision system - Google Patents
High-precision positioning method and device in wafer-level flip vision system Download PDFInfo
- Publication number
- CN108231645B CN108231645B CN201711479192.9A CN201711479192A CN108231645B CN 108231645 B CN108231645 B CN 108231645B CN 201711479192 A CN201711479192 A CN 201711479192A CN 108231645 B CN108231645 B CN 108231645B
- Authority
- CN
- China
- Prior art keywords
- crystal grain
- super
- circle
- resolution
- sub
- 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
Links
Images
Classifications
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L21/00—Processes or apparatus adapted for the manufacture or treatment of semiconductor or solid state devices or of parts thereof
- H01L21/67—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere
- H01L21/68—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere for positioning, orientation or alignment
- H01L21/681—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere for positioning, orientation or alignment using optical controlling means
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- General Physics & Mathematics (AREA)
- Manufacturing & Machinery (AREA)
- Computer Hardware Design (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Power Engineering (AREA)
- Image Analysis (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a high-precision positioning method and a high-precision positioning device in a wafer-level flip vision system, wherein the method comprises the following steps: step 1, creating a crystal grain template, positioning the crystal grains to be positioned in the whole wafer by a template matching method, and obtaining coordinate information and a crystal grain graph of the crystal grains; step 2, cutting the grain pattern, and reconstructing the cut grain pattern by adopting a learning-based super-resolution reconstruction algorithm of a single image, so as to obtain a super-resolution grain image; and 3, preprocessing the super-resolution crystal grain image, and acquiring corner information and circle center coordinates of the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction. The crystal grains are positioned by combining a super-resolution reconstruction algorithm based on a single image and a circle center positioning algorithm based on sub-pixel edge extraction, so that the positioning precision is improved, the hardware cost increase caused by the adoption of a high-speed and high-resolution CCD camera is avoided, and the packaging cost is reduced.
Description
Technical Field
The invention relates to the field of semiconductor packaging crystals, in particular to a high-precision positioning method and device in a wafer-level flip vision system.
Background
At present, semiconductor integrated circuit design, fabrication and packaging test instruments are known as the three major pillars of the semiconductor industry. The semiconductor manufacturing process can be divided into a wafer processing process, a wafer probing process, a packaging process, a testing process, and so on. Generally, the wafer processing process and the wafer probing process are front-end processes, and the packaging and testing processes are back-end processes. Examples of microelectronic package engineering show that common process technologies for microelectronic packaging include capping (Sealing), Wire/ball bonding, Flip Chip, Die bonding, Chip (Chip), Substrate (Substrate), etc.
The semiconductor technology is gradually densified and highly integrated, the transmission amount of Chip signals is increasing day by day, the number of pins is gradually increased, and the packaging industry gradually moves from the traditional Dual In-line Package (Dual In-line Package), Small Outline Package (Small Outline Package), Pin Array Package (Pin Grid Array) to the novel packaging forms of Ball Array Package (Ball Grid Array), Chip size Package (Chip Scale Package), Flip Chip (Flip Chip), three-dimensional Package and the like
The trend in microelectronic products today is high density, low cost, and it is these factors that have prompted the development of flip chips, chip scale packages, and wafer level chip scale packages.
Wafer Level Package (WLP) has advantages in small size, excellent electrical performance, good heat dissipation, high cost performance, etc. it is based on solder ball array package, and is an improved and improved chip size package that has developed rapidly in recent years. WLP has become an important component of advanced packaging technology, where most or all of the packaging, testing procedures are performed directly on the wafer, followed by dicing.
Because of the wafer level flip chip, the dimensions are very small, mainly on the order of microns. The vision system needs to have extremely high positioning accuracy, and currently some world famous packaging equipment companies such as K & S, ASM and ESEC mainly use a high-resolution and high-speed CCD camera to improve the image positioning accuracy. Therefore, the hardware cost is increased sharply, the product price is high, and only large enterprises can use the method on the production line.
Disclosure of Invention
The invention provides a high-precision positioning method and device in a wafer-level flip vision system, which improve the positioning precision, avoid the use of a high-speed and high-resolution CCD camera and reduce the cost of packaging equipment.
In order to solve the above technical problem, an embodiment of the present invention provides a high precision positioning method in a wafer level flip-chip vision system, including: step 1, creating a crystal grain template, positioning the crystal grains to be positioned in the whole wafer by a template matching method, and obtaining coordinate information and a crystal grain graph of the crystal grains;
step 2, cutting the grain graph, and reconstructing the cut grain graph by adopting a learning-based super-resolution reconstruction algorithm of a single image so as to obtain a super-resolution grain image;
step 3, after preprocessing the super-resolution crystal grain image, obtaining the corner information of the crystal grain and the center coordinates of the salient points of the crystal grain by adopting a center positioning algorithm based on sub-pixel edge extraction,
preprocessing the super-resolution crystal grain image, including image segmentation, morphological operation and feature extraction on the super-resolution crystal grain image;
the method for obtaining the corner information of the crystal grain and the circle center coordinate of the salient point of the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction comprises the following steps:
after extracting the sub-pixel rounding outline and the coordinate information of the circle center of the salient point on the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction, fitting the circle centers of the salient points in four directions at the outermost periphery into a straight line to obtain the corner information of the crystal grain;
the step 3 comprises the following steps:
carrying out median and Gaussian filtering processing on the super-resolution crystal grain image;
performing threshold segmentation on the super-resolution crystal grain image;
performing sub-pixel edge extraction on the salient points of the super-resolution crystal grain image to obtain the circle centers and the radiuses of the salient points;
calculating the outermost chip salient points in four directions of the super-resolution crystal grain image, fitting a straight line for the outermost chip salient points in each direction by using a least square method, calculating the center and corner errors of the crystal grains, obtaining the coordinate information of the salient points on the crystal grains and the corner information of the crystal grains, and finishing the alignment of the crystal grains and the crystal grain template.
The method for obtaining the corner information of the crystal grain and the circle center coordinate of the salient point of the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction comprises the following steps:
firstly, sequentially carrying out projection processing and differential processing on the super-resolution crystal grain image, then carrying out sub-pixel processing after setting edge sensitivity and edge polarity on the super-resolution crystal grain image, and extracting the edge of the super-resolution crystal grain image;
processing the edge of the super-resolution grain image by adopting a Canny operator to obtain a pixel level edge of a circular profile, and performing least square circle fitting on the pixel level edge of the circular profile to obtain a rough position of the center of the circular profile;
calculating the accurate sub-pixel edge of the circle by using a caliper tool principle along the radius direction of the circle profile;
and performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circular contour.
Wherein, the step 3 further comprises:
judging whether the accuracy of the circle center coordinates of the salient points of the crystal grains meets the preset accuracy or not;
if not, repeatedly calculating the accurate sub-pixel edge of the circle by using a caliper tool principle along the radius direction of the circular contour, performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circular contour.
In addition, an embodiment of the present invention further provides a high precision positioning apparatus in a wafer level flip-chip vision system, including:
the shooting module is used for shooting a wafer to obtain a graph of the wafer;
the crystal grain template creating module is used for creating a corresponding crystal grain template according to the graph of the wafer, positioning the crystal grains needing to be positioned in the crystal grain template and obtaining the crystal grain coordinate information and the crystal grain graph;
the super-resolution reconstruction module based on the learned single image is connected with the crystal grain template creation module and is used for obtaining the crystal grain graph from the crystal grain template creation module, cutting and reconstructing the crystal grain graph to obtain a super-resolution crystal grain image;
a circle center positioning module based on sub-pixel edge extraction, connected with the super-resolution reconstruction module based on the learned single image, and used for obtaining the super-resolution crystal grain image from the super-resolution reconstruction module based on the learned single image, preprocessing the super-resolution crystal grain image, obtaining the sub-pixel edge circle outline and the coordinate information of the circle center of the salient point of the crystal grain, fitting the circle centers of the salient points in four directions at the outermost periphery to obtain the corner information of the crystal grain, and further comprising a circle center precision judging module connected with the circle center positioning module based on sub-pixel edge extraction, judging whether the coordinate information of the circle center of the sub-pixel edge circle outline of the salient point of the crystal grain, which is obtained by calculation of the circle center positioning module based on sub-pixel edge extraction, meets the preset precision, if not, controlling the circle center positioning module based on sub-pixel edge extraction to position the sub-pixel edge circle outline of the salient point of the crystal grain along the circle And repeatedly calculating the accurate sub-pixel edge of the circle in the radius direction of the circle profile by using a caliper tool principle, performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circle profile.
Compared with the prior art, the high-precision positioning method and device in the wafer-level flip-chip vision system provided by the embodiment of the invention have the following advantages:
the embodiment of the invention provides a high-precision positioning method in a wafer-level flip-chip vision system, which comprises the following steps: step 1, creating a crystal grain template, positioning the crystal grains to be positioned in the whole wafer by a template matching method, and obtaining coordinate information and a crystal grain graph of the crystal grains;
step 2, cutting the grain graph, and reconstructing the cut grain graph by adopting a learning-based super-resolution reconstruction algorithm of a single image so as to obtain a super-resolution grain image;
and 3, preprocessing the super-resolution crystal grain image, and then obtaining corner information of the crystal grain and the center coordinates of the salient points of the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction.
The embodiment of the invention also provides a high-precision positioning device in the wafer-level flip-chip vision system, which comprises:
the shooting module is used for shooting a wafer to obtain a graph of the wafer;
the crystal grain template creating module is used for creating a corresponding crystal grain template according to the graph of the wafer, positioning the crystal grains needing to be positioned in the crystal grain template and obtaining the crystal grain coordinate information and the crystal grain graph;
the super-resolution reconstruction module based on the learned single image is connected with the crystal grain template creation module and is used for obtaining the crystal grain graph from the crystal grain template creation module, cutting and reconstructing the crystal grain graph to obtain a super-resolution crystal grain image;
and the circle center positioning module is connected with the super-resolution reconstruction module based on the learned single image and used for acquiring the super-resolution crystal grain image from the super-resolution reconstruction module based on the learned single image, preprocessing the super-resolution crystal grain image, acquiring the sub-pixel edge circle outline and the coordinate information of the circle center of the salient point of the crystal grain, and fitting the circle centers of the salient points in four directions at the outermost periphery to acquire the corner information of the crystal grain.
According to the high-precision positioning method and device in the wafer-level flip vision system, the crystal grain template is firstly established, the crystal grains are positioned in the whole wafer through positioning prevention of template matching, coordinate information and crystal grain images are obtained, then the positioned crystal grain images are cut and reconstructed by adopting a super-resolution reconstruction algorithm based on a single learning image to obtain super-resolution crystal grain images, after a series of preprocessing is carried out on the super-resolution crystal grain images, the circle center is extracted by utilizing a circle center positioning algorithm based on sub-pixel edge extraction to obtain the corner information of the crystal grains, the positioning precision is improved, the hardware cost caused by the adoption of a high-speed and high-resolution CCD camera is avoided, and the packaging cost is reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flowchart illustrating a method for high precision positioning in a wafer level flip-chip vision system according to an embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating an embodiment of a high-precision positioning apparatus in a wafer-level flip-chip vision system according to the present invention;
fig. 3 is a schematic structural diagram of another embodiment of a high-precision positioning method in a wafer-level flip-chip vision system according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, fig. 1 is a schematic flowchart illustrating a process of an embodiment of a high precision positioning method in a wafer level flip-chip vision system according to the present invention; FIG. 2 is a schematic diagram illustrating an embodiment of a high-precision positioning apparatus in a wafer-level flip-chip vision system according to the present invention; fig. 3 is a schematic structural diagram of another embodiment of a high-precision positioning method in a wafer-level flip-chip vision system according to an embodiment of the present invention.
In one embodiment, the method for high precision positioning in a wafer level flip chip vision system comprises: step 1, creating a crystal grain template, positioning the crystal grains to be positioned in the whole wafer by a template matching method, and obtaining coordinate information and a crystal grain graph of the crystal grains; the positioning information of the crystal grains needing to be positioned can be roughly obtained by a positioning method of template matching.
Step 2, cutting the grain graph, and reconstructing the cut grain graph by adopting a learning-based super-resolution reconstruction algorithm of a single image so as to obtain a super-resolution grain image; the method comprises the steps of obtaining a graph of a crystal grain to be positioned, then cutting the graph, obtaining coordinate information of one of three information, namely three element angle postures and coordinate information of the crystal grain positioned on a wafer and the circle center of a salient point on the crystal grain, and finally reconstructing the cut crystal grain graph by adopting a super-resolution reconstruction algorithm based on a learned single image, thereby obtaining a super-resolution crystal grain image and providing a basis for subsequently obtaining corner information of the crystal grain and the circle center of the salient point.
And 3, preprocessing the super-resolution crystal grain image, and then obtaining corner information of the crystal grain and the center coordinates of the salient points of the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction. The super-resolution crystal grain image is preprocessed, so that the accuracy is improved and the positioning accuracy is improved when a circle center positioning algorithm based on sub-pixel edge extraction is adopted for corner positioning and bump calculation in the follow-up process.
The purpose of preprocessing the super-resolution crystal grain image is to improve contrast and facilitate calculation of center coordinates of salient points and corner information of crystal grains by using a center positioning algorithm based on sub-pixel edge extraction.
It should be noted that the present invention performs preprocessing on the super-resolution grain image, including but not limited to performing image segmentation, morphological operation and feature extraction on the super-resolution grain image, and may also include other preprocessing processes, which are smooth in the preprocessing process, and the present invention is not limited specifically, so as to save processing steps and improve resolution precision. The specific image segmentation, morphological operation and feature extraction are not limited in the present invention.
In the present invention, in the elements of the positioning of the crystal grain, the obtaining of the angular posture of the crystal grain and the center of the circle of the salient point on the crystal grain is a key point, and in an embodiment, the obtaining of the corner information of the crystal grain and the center coordinate of the salient point of the crystal grain by using the center positioning algorithm based on the extraction of the sub-pixel edge includes:
and after extracting the sub-pixel rounding outline and the coordinate information of the circle center of the salient point on the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction, fitting the circle centers of the salient points in four directions at the outermost periphery into a straight line to obtain the corner information of the crystal grain.
The corner information of the crystal grain is obtained by extracting the sub-pixel rounding outline of the salient point on the crystal grain, then calculating the circle center coordinate of the sub-pixel rounding outline, and fitting the circle centers of the salient points in four directions at the outermost periphery into a straight line. The algorithm for fitting the straight line by using the centers of the salient points is not particularly limited.
In order to further enhance the positioning effect, in an embodiment of the present invention, the step 3 includes:
carrying out median and Gaussian filtering processing on the super-resolution crystal grain image;
performing threshold segmentation on the super-resolution crystal grain image;
performing sub-pixel edge extraction on the salient points of the super-resolution crystal grain image to obtain the circle centers and the radiuses of the salient points;
calculating the outermost chip salient points in four directions of the super-resolution crystal grain image, fitting a straight line for the outermost chip salient points in each direction by using a least square method, calculating the center and corner errors of the crystal grains, obtaining the coordinate information of the salient points on the crystal grains and the corner information of the crystal grains, and finishing the alignment of the crystal grains and the crystal grain template.
Performing median and Gaussian filtering on the super-resolution grain image, performing threshold segmentation on the image, then performing sub-pixel edge extraction, distinguishing sensitive boundaries in the image, and finally obtaining the circle center and the radius of the salient point. In the invention, because a plurality of salient points generally exist on one crystal grain chip, the crystal grain can be positioned only by positioning the salient point at the outermost periphery.
The learning-based single image super-resolution reconstruction algorithm is mainly used for intensively extracting a required high-frequency information model from a training sample by a machine learning method, so that the required information of an unknown test sample is predicted, and the aim of high resolution is fulfilled.
And computing the low-dimensional and domain-preserving embedding of high-dimensional input data by adopting a Local Linear (LLE) popular learning method. The algorithm mainly comprises three steps:
(1) and calculating the number of the field points of each sample point xi in the high-dimensional RD space in the D-dimensional space calculation field.
(2) And calculating a reconstruction weight matrix.
The reconstruction weight Wij represents the contribution made by the reconstruction of the sample point xj, and is selected by minimizing the reconstruction error:
and requires that the following two constraints be satisfied:
(1) for all of the i's, the average,
；
(2) if it is
Is out of position
In the field set of
After adding the constraint condition, the sample point
Can be written as
Order to
Weight vector
. Solving for a regularized linear system.
(3) And calculating d-dimensional embedding.
And the original data point is reconstructed by keeping the field weight unchanged in the d-dimensional space, so that the reconstruction error is minimum. Computing matrices
Is smallest in
A feature vector
Then, then
Is the calculated embedding result.
The purpose of the circle center positioning algorithm based on sub-pixel edge extraction in the invention is to obtain the circle center coordinates and the corner information of the salient points in the crystal grain image, and the invention does not limit the specific processing process.
In an embodiment, the obtaining the corner information of the die and the coordinates of the center of the salient point of the die by using a circle center positioning algorithm based on sub-pixel edge extraction includes:
firstly, sequentially carrying out projection processing and differential processing on the super-resolution crystal grain image, then carrying out sub-pixel processing after setting edge sensitivity and edge polarity on the super-resolution crystal grain image, and extracting the edge of the super-resolution crystal grain image;
processing the edge of the super-resolution grain image by adopting a Canny operator to obtain a pixel level edge of a circular profile, and performing least square circle fitting on the pixel level edge of the circular profile to obtain a rough position of the center of the circular profile;
calculating the accurate sub-pixel edge of the circle by using a caliper tool principle along the radius direction of the circle profile;
and performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circular contour.
It should be pointed out that, in the present invention, projection processing and differential processing are sequentially performed on the super-resolution grain image, where the differential processing process is not specifically limited, and may be a specified differential processing mode, or a differential processing mode in the same type of randomly selected picture processing process. And then, obtaining a pixel-level edge of the circular contour by using a Canny operator, and fitting the obtained pixel-level edge, wherein a rough position of a circle is generally obtained by using least square fitting in the invention, and other fitting methods can be adopted, which are not specifically limited in the invention. Then, the original accurate pixel level edge is calculated by applying the pubic principle, and then the least square fitting is carried out on the edge again to calculate the accurate position of the circle center.
It should be noted that, in the present invention, the fitting method adopted in the process of obtaining the circle center is not limited to least square fitting, the fitting of the pixel-level edge of the circle profile is not limited to twice fitting, if the accuracy cannot meet the requirement, fitting can be further performed again, and the fitting algorithm adopted in each fitting may be different, and may not be limited to least square algorithm.
Because the required positioning accuracy of different manufacturers for crystal grains with different sizes is different in the process of calculating the coordinates of the circle center of the salient point, in the calculation process, along with the increase of the calculation times, the accuracy of the circle center is improved, but the calculation amount is greatly increased, the cost is increased, the accuracy of the circle center obtained by twice fitting of a common manufacturer can already meet the positioning requirement, and some manufacturers can not meet the requirement, therefore, in one embodiment of the invention, the step 3 further comprises the following steps:
judging whether the accuracy of the circle center coordinates of the salient points of the crystal grains meets the preset accuracy or not;
if not, repeatedly calculating the accurate sub-pixel edge of the circle by using a caliper tool principle along the radius direction of the circular contour, performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circular contour.
The method comprises the steps of presetting a certain circle center positioning accuracy, judging the accuracy after performing least square circle fitting to obtain a circular position each time, stopping fitting if the accuracy can meet the requirement, otherwise, repeatedly calculating the original accurate sub-pixel edge, performing least square circle fitting again, calculating the accurate position of the circle center, namely, taking the accurate position of the last circle center as the current rough position of the circle center, and repeatedly calculating until the requirement is met.
In addition, an embodiment of the present invention further provides a high precision positioning apparatus in a wafer level flip-chip vision system, including:
the shooting module 10 is used for shooting a wafer to obtain a graph of the wafer;
a die template creating module 20, configured to create a corresponding die template according to the wafer pattern, locate a die to be located in the die template, and obtain the die coordinate information and the die pattern;
the super-resolution reconstruction module 30 based on the learned single image is connected with the crystal grain template creation module 20 and is used for obtaining the crystal grain graph from the crystal grain template creation module 20, cutting and reconstructing the crystal grain graph to obtain a super-resolution crystal grain image;
and the circle center positioning module 40 is connected with the super-resolution reconstruction module 30 based on the learned single image, and is used for obtaining the super-resolution crystal grain image from the super-resolution reconstruction module 30 based on the learned single image, preprocessing the super-resolution crystal grain image, obtaining the sub-pixel edge circle outline and the coordinate information of the circle center of the salient point of the crystal grain, and fitting the circle centers of the salient points in four directions at the outermost periphery to obtain the corner information of the crystal grain.
The preprocessing of the super-resolution grain image obtained from the learning-based single image super-resolution reconstruction module 30 by the circle center positioning module 40 based on sub-pixel edge extraction generally includes image segmentation, morphological operation, and feature extraction, to obtain the salient points on the grains, and to distinguish the salient point regions from the non-salient point regions. And then extracting the sub-pixel edge outline of the salient point to obtain circle center coordinate information and radius, finally fitting to obtain the error between the center of the chip and a corner, obtaining the corner information of the crystal grain, realizing the correspondence between the crystal grain and the substrate, and realizing the determined positioning.
The calculation process of the circle center positioning module 40 based on sub-pixel edge extraction includes:
carrying out median and Gaussian filtering processing on the image;
performing threshold segmentation on the image;
extracting sub-pixel edges of the salient points of the chip to obtain the circle centers and the radiuses of the salient points;
calculating the outermost chip salient points of the crystal grain image in four directions;
for the outermost salient point in each direction, fitting a straight line by using a least square method, and calculating the error between the center of the chip and the corner, so as to obtain the coordinate information of the salient point on each crystal grain and the corner information of the crystal grain, thereby finishing the alignment of the crystal grain and the substrate.
The specific process of extracting the edge of the circular contour and calculating the center of the circle is as follows:
1. and (5) an edge extraction process. The edge is a boundary dividing a light area and a dark area in an image. The industrial edge detection caliper tool comprises the following main steps:
performing a projection process, performing a differential process, setting edge sensitivity, edge polarity, and performing a sub-pixel process.
2. And (5) extracting the circle center. After the precise edge profile of the circle is obtained, circle fitting can be performed on the edge points through a direct circle fitting method, so that the precise center position of the circle is determined. The specific process is as follows:
(1) obtaining the pixel-level edge of the circular contour by using a Canny operator;
(2) performing least square circle fitting on the pixel level edge of the circle profile obtained in the step (1) to obtain the rough position of the circle center;
(3) calculating the accurate sub-pixel edge of the circle by using a caliper tool principle along the radius direction of the circle;
(4) and (4) performing least square circle fitting on the edge obtained in the step (3) again, and calculating the accurate position of the circle center.
Because the positioning accuracy required by different manufacturers for crystal grains with different sizes is different in the process of calculating the coordinates of the circle center of the salient point, and the accuracy of the circle center is increased along with the increase of the calculation times in the calculation process, but the calculation amount is greatly increased, the cost is increased, the accuracy of the circle center obtained by two-time fitting by common manufacturers can already meet the positioning requirement, and some manufacturers can not meet the requirement, for this reason, in one embodiment of the invention, the high-accuracy positioning method and device in the wafer level flip-chip vision system further comprise a circle center accuracy judging module 50 connected with the circle center positioning module 40 based on sub-pixel edge extraction, and the method and device judge whether the coordinate information of the circle center of the sub-pixel edge circle outline of the salient point of the crystal grain, which is obtained by calculation by the circle center positioning module 40 based on sub-pixel edge extraction, meets the preset accuracy, if not, controlling the circle center positioning module 40 based on sub-pixel edge extraction to repeatedly calculate the accurate sub-pixel edge of the circle for the sub-pixel edge circle contour of the salient point of the crystal grain along the radius direction of the circle contour of the circle by using a caliper tool principle, performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circle contour.
In summary, according to the high-precision positioning method and apparatus in the wafer-level flip-chip vision system provided by the embodiments of the present invention, the die template is created first, the die is positioned in the entire wafer by the positioning prevention of the template matching, the coordinate information and the die image are obtained, then the positioned die image is cut, and then the super-resolution reconstruction algorithm based on the learned single image is used to reconstruct the super-resolution die image, so as to obtain the super-resolution die image, and after a series of pre-processing is performed on the super-resolution die image, the center of a circle is extracted by using the center positioning algorithm based on the sub-pixel edge extraction, so as to obtain the corner information of the die, thereby improving the positioning precision, avoiding the hardware cost increase caused by using the high-speed and high-resolution CCD camera, and reducing the packaging cost.
The method and apparatus for high precision positioning in a wafer level flip chip vision system provided by the present invention are described in detail above. The principles and embodiments of the present invention are explained herein using specific examples, which are presented only to assist in understanding the method and its core concepts. It should be noted that, for those skilled in the art, it is possible to make various improvements and modifications to the present invention without departing from the principle of the present invention, and those improvements and modifications also fall within the scope of the claims of the present invention.
Claims (4)
1. A high-precision positioning method in a wafer-level flip-chip vision system is characterized by comprising the following steps: step 1, creating a crystal grain template, positioning the crystal grains to be positioned in the whole wafer by a template matching method, and obtaining coordinate information and a crystal grain graph of the crystal grains;
step 2, cutting the grain graph, and reconstructing the cut grain graph by adopting a learning-based super-resolution reconstruction algorithm of a single image so as to obtain a super-resolution grain image;
step 3, after preprocessing the super-resolution crystal grain image, obtaining corner information of the crystal grain and circle center coordinates of salient points of the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction;
preprocessing the super-resolution crystal grain image, including image segmentation, morphological operation and feature extraction on the super-resolution crystal grain image;
the method for obtaining the corner information of the crystal grain and the circle center coordinate of the salient point of the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction comprises the following steps:
after extracting the sub-pixel rounding outline and the coordinate information of the circle center of the salient point on the crystal grain by adopting a circle center positioning algorithm based on sub-pixel edge extraction, fitting the circle centers of the salient points in four directions at the outermost periphery into a straight line to obtain the corner information of the crystal grain, wherein the step 3 comprises the following steps:
carrying out median and Gaussian filtering processing on the super-resolution crystal grain image;
performing threshold segmentation on the super-resolution crystal grain image;
performing sub-pixel edge extraction on the salient points of the super-resolution crystal grain image to obtain the circle centers and the radiuses of the salient points;
calculating the outermost chip salient points in four directions of the super-resolution crystal grain image, fitting a straight line for the outermost chip salient points in each direction by using a least square method, calculating the center and corner errors of the crystal grains, obtaining the coordinate information of the salient points on the crystal grains and the corner information of the crystal grains, and finishing the alignment of the crystal grains and the crystal grain template.
2. The method of claim 1, wherein obtaining corner information of the die and coordinates of a center of a bump of the die using a circle center location algorithm based on sub-pixel edge extraction comprises:
firstly, sequentially carrying out projection processing and differential processing on the super-resolution crystal grain image, then carrying out sub-pixel processing after setting edge sensitivity and edge polarity on the super-resolution crystal grain image, and extracting the edge of the super-resolution crystal grain image;
processing the edge of the super-resolution grain image by adopting a Canny operator to obtain a pixel level edge of a circular profile, and performing least square circle fitting on the pixel level edge of the circular profile to obtain a rough position of the center of the circular profile;
calculating the accurate sub-pixel edge of the circle by using a caliper tool principle along the radius direction of the circle profile;
and performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circular contour.
3. A method for high precision positioning in a wafer level flip chip vision system as claimed in claim 2 wherein said step 3 further comprises:
judging whether the accuracy of the circle center coordinates of the salient points of the crystal grains meets the preset accuracy or not;
if not, repeatedly calculating the accurate sub-pixel edge of the circle by using a caliper tool principle along the radius direction of the circular contour, performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circular contour.
4. A high precision positioning device in a wafer level flip chip vision system, comprising:
the shooting module is used for shooting a wafer to obtain a graph of the wafer;
the crystal grain template creating module is used for creating a corresponding crystal grain template according to the graph of the wafer, positioning the crystal grains needing to be positioned in the crystal grain template and obtaining the crystal grain coordinate information and the crystal grain graph;
the super-resolution reconstruction module based on the learned single image is connected with the crystal grain template creation module and is used for obtaining the crystal grain graph from the crystal grain template creation module, cutting and reconstructing the crystal grain graph to obtain a super-resolution crystal grain image;
a circle center positioning module based on sub-pixel edge extraction, connected with the super-resolution reconstruction module based on the learned single image, and used for obtaining the super-resolution crystal grain image from the super-resolution reconstruction module based on the learned single image, preprocessing the super-resolution crystal grain image, obtaining the sub-pixel edge circle outline and the coordinate information of the circle center of the salient point of the crystal grain, fitting the circle centers of the salient points in four directions at the outermost periphery to obtain the corner information of the crystal grain, and further comprising a circle center precision judging module connected with the circle center positioning module based on sub-pixel edge extraction, judging whether the coordinate information of the circle center of the sub-pixel edge circle outline of the salient point of the crystal grain, which is obtained by calculation of the circle center positioning module based on sub-pixel edge extraction, meets the preset precision, if not, controlling the circle center positioning module based on sub-pixel edge extraction to position the sub-pixel edge circle outline of the salient point of the crystal grain along the circle And repeatedly calculating the accurate sub-pixel edge of the circle in the radius direction of the circle profile by using a caliper tool principle, performing least square circle fitting on the accurate sub-pixel edge, and calculating the accurate position of the circle center of the circle profile.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711479192.9A CN108231645B (en) | 2017-12-29 | 2017-12-29 | High-precision positioning method and device in wafer-level flip vision system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711479192.9A CN108231645B (en) | 2017-12-29 | 2017-12-29 | High-precision positioning method and device in wafer-level flip vision system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108231645A CN108231645A (en) | 2018-06-29 |
| CN108231645B true CN108231645B (en) | 2020-10-20 |
Family
ID=62647036
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711479192.9A Active CN108231645B (en) | 2017-12-29 | 2017-12-29 | High-precision positioning method and device in wafer-level flip vision system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108231645B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110927549B (en) * | 2019-11-21 | 2021-11-16 | 广西天微电子有限公司 | Wafer repositioning method and system |
| CN112710632A (en) * | 2021-03-23 | 2021-04-27 | 四川京炜交通工程技术有限公司 | Method and system for detecting high and low refractive indexes of glass beads |
| CN113538384B (en) * | 2021-07-19 | 2024-03-26 | 凌云光技术股份有限公司 | Feature positioning method and device |
| CN113822883B (en) * | 2021-11-22 | 2022-03-11 | 中导光电设备股份有限公司 | Wafer alignment method and system based on searching of straight edge of wafer |
| CN117058005B (en) * | 2023-10-13 | 2024-01-16 | 珠海埃克斯智能科技有限公司 | Crystal grain image reconstruction method and device of wafer, electronic equipment and storage medium |
| CN117393485B (en) * | 2023-11-01 | 2024-05-03 | 东莞触点智能装备有限公司 | Chip high-precision die bonder visual positioning system based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103107121A (en) * | 2013-01-30 | 2013-05-15 | 福建省威诺数控有限公司 | Wafer angular deviation automatic method based on vision |
| CN103235939A (en) * | 2013-05-08 | 2013-08-07 | 哈尔滨工业大学 | Datum point positioning method based on machine vision |
| CN103681427A (en) * | 2013-12-09 | 2014-03-26 | 深圳市大族激光科技股份有限公司 | Vision-based wafer rotation correction and centralized positioning method |
| US9748128B1 (en) * | 2016-06-01 | 2017-08-29 | Micron Technology, Inc. | Systems and methods for wafer alignment |
-
2017
- 2017-12-29 CN CN201711479192.9A patent/CN108231645B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103107121A (en) * | 2013-01-30 | 2013-05-15 | 福建省威诺数控有限公司 | Wafer angular deviation automatic method based on vision |
| CN103235939A (en) * | 2013-05-08 | 2013-08-07 | 哈尔滨工业大学 | Datum point positioning method based on machine vision |
| CN103681427A (en) * | 2013-12-09 | 2014-03-26 | 深圳市大族激光科技股份有限公司 | Vision-based wafer rotation correction and centralized positioning method |
| US9748128B1 (en) * | 2016-06-01 | 2017-08-29 | Micron Technology, Inc. | Systems and methods for wafer alignment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108231645A (en) | 2018-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108231645B (en) | High-precision positioning method and device in wafer-level flip vision system | |
| Romero-Ramirez et al. | Speeded up detection of squared fiducial markers | |
| US9401313B2 (en) | Automated optical inspection of unit specific patterning | |
| WO2022170844A1 (en) | Video annotation method, apparatus and device, and computer readable storage medium | |
| EP3537375B1 (en) | Image segmentation methods, image segmentation system and device comprising same, and storage medium | |
| US7324684B2 (en) | Bonding apparatus | |
| CN105046684B (en) | A kind of image matching method based on polygon generalised Hough transform | |
| JP6483168B2 (en) | System and method for efficiently scoring a probe in an image with a vision system | |
| US7321681B2 (en) | Bonding pattern discrimination program | |
| US20190147580A1 (en) | Positioning Method of Rectangular Pin Element Based on Vertex Points | |
| CN105205459B (en) | A kind of recognition methods of characteristics of image vertex type and device | |
| CN115690670A (en) | Intelligent identification method and system for wafer defects | |
| CN113221953A (en) | Target attitude identification system and method based on example segmentation and binocular depth estimation | |
| CN107507130A (en) | A kind of quickly QFN chip pins image obtains and amplification method | |
| US20110174865A1 (en) | Method of teaching eyepoints for wire bonding and related semiconductor processing operations | |
| Liu et al. | A novel subpixel industrial chip detection method based on the dual-edge model for surface mount equipment | |
| Blanz et al. | Image analysis methods for solderball inspection in integrated circuit manufacturing | |
| CN114202578A (en) | Wafer alignment method and device | |
| US20180082109A1 (en) | Curvature-based face detector | |
| CN115082552A (en) | Marking hole positioning method and device, assembly equipment and storage medium | |
| CN104009144A (en) | Matching method for substrate and chip in high-power LED eutectic welding | |
| Pan et al. | Improved image processing algorithms for microprobe final test | |
| TWI840217B (en) | Wafer positioning system and method using Hough transform | |
| CN117191815A (en) | Mini/Micro LED small target defect detection method and classification system | |
| CN117314906B (en) | Wafer grain detection method based on array distribution information guidance |
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 |