CN117470104A - Semiconductor device surface dust removing method and system based on visual detection - Google Patents
Semiconductor device surface dust removing method and system based on visual detection Download PDFInfo
- Publication number
- CN117470104A CN117470104A CN202311781372.8A CN202311781372A CN117470104A CN 117470104 A CN117470104 A CN 117470104A CN 202311781372 A CN202311781372 A CN 202311781372A CN 117470104 A CN117470104 A CN 117470104A
- Authority
- CN
- China
- Prior art keywords
- dust
- semiconductor device
- image
- dust removing
- foreign matter
- 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.)
- Granted
Links
- 239000000428 dust Substances 0.000 title claims abstract description 379
- 239000004065 semiconductor Substances 0.000 title claims abstract description 95
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000001514 detection method Methods 0.000 title claims abstract description 27
- 230000000007 visual effect Effects 0.000 title claims abstract description 24
- 238000011179 visual inspection Methods 0.000 claims description 15
- 238000004891 communication Methods 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 5
- 238000005520 cutting process Methods 0.000 claims description 3
- 239000007787 solid Substances 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims 2
- 238000004519 manufacturing process Methods 0.000 description 5
- 230000003287 optical effect Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000007598 dipping method Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000001179 sorption measurement Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000007664 blowing Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B08—CLEANING
- B08B—CLEANING IN GENERAL; PREVENTION OF FOULING IN GENERAL
- B08B7/00—Cleaning by methods not provided for in a single other subclass or a single group in this subclass
- B08B7/0028—Cleaning by methods not provided for in a single other subclass or a single group in this subclass by adhesive surfaces
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/28—Measuring arrangements characterised by the use of optical techniques for measuring areas
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/94—Investigating contamination, e.g. dust
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/457—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by analysing connectivity, e.g. edge linking, connected component analysis or slices
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Abstract
The invention discloses a semiconductor device surface dust removing method and a system based on visual detection, wherein the method comprises the following steps: acquiring a real-time image of the semiconductor device, and matching and comparing the real-time image with a preset reference image to obtain position information of dust and foreign matters on the surface of the semiconductor device and a dust and foreign matter image; and the dust removing rod with corresponding specification is picked up according to the area size of the dust and foreign matter image, and is driven to contact and adhere with the dust and foreign matter according to the position information of the dust and foreign matter. The method has the advantages that the dust and foreign matters on the surface of the semiconductor device are rapidly identified by adopting an image identification mode, and the best matched dust removing rod is selected to contact and adhere and remove according to the dust and foreign matter image obtained in real time, so that the dust removing work can be rapidly completed, meanwhile, the consumption of the dust removing rod is minimum, and the contact times of the semiconductor device and the dust removing rod are minimized when the semiconductor device is used for removing dust, so that the foreign matters on the surface of the device are rapidly removed, and the damage or secondary pollution on the surface of the device is also furthest avoided.
Description
Technical Field
The invention relates to the technical field of semiconductor production and manufacturing, in particular to a semiconductor device surface dust removing method and a system based on visual detection.
Background
In the process of manufacturing and assembling electronic components and optical elements, surface appearance inspection is generally required for the electronic components and the optical elements. Surface appearance detection is generally to detect whether or not the surface of an electronic component and the surface of an optical element have surface appearance problems such as dust foreign matter, dirt, scratches, cracks, breakage, dust, fingerprints, watermarks, wire bond anomalies, die bond anomalies, and the like. When the surfaces of the electronic components and the surfaces of the optical elements are inspected for dust, the surfaces of the electronic components and the surfaces of the optical elements having dust need to be cleaned.
When dust and foreign matters appear on the surface of the existing device, the existing device is generally treated by the following methods: sticking, cleaning, erasing, blowing and adsorbing; in order to avoid damage to the surface of the device or secondary pollution, chinese patent publication No. CN113399377B discloses a dust removing head, a viscous dust removing device, a dust removing device and a method, in which the surface detecting vision device is used to detect the surface problem of the detected surface of the detected object, the viscous dust removing device is used to adhere the dust on the detected surface of the detected object, and the viscous dust is easy to be removed. The sticky dust removal device is not easy to generate new pollutants on the cleaned detected surface after dust is removed. The visual inspection device is used for shooting a size image of the viscous semisolid part connected to the dust removing mechanism, the size image is convenient for a control center or a user to know the real-time size of the viscous semisolid part corresponding to the size image, so that whether the viscous semisolid part is in a preset size range or not is judged according to the real-time size, if the real-time size is not in the preset size range, the control center or the user replaces the dust removing head corresponding to the size image, and if the real-time size is in the preset size range, the control center or the user continues to use the dust removing head corresponding to the size image. When the using times of the dust removing head connected to the dust removing mechanism reach the preset times and the real-time size is in the preset size range, the dust removing head connected to the dust removing mechanism needs to be subjected to gum dipping treatment. In production practice, the scheme solves the problem that the dust removing rod with single specification is not adaptive to the dust size, but the dust removing head is adopted to dip the adsorption solution and carry out image detection, so that the action is complex, the data processing amount is large, the improvement of the dust removing efficiency of the surface of the device is restricted, and the further optimization is needed.
Disclosure of Invention
The invention provides a semiconductor device surface dust removing method and a system based on visual detection, and aims to solve the technical problem that the existing dust removing head is low in efficiency when dipping adsorption solution to remove dust on the device surface.
In order to solve the technical problem, a first aspect of the present invention provides a method for removing dust on a surface of a semiconductor device based on visual inspection, the method comprising:
acquiring a real-time image of a semiconductor device, and matching and comparing the real-time image with a preset reference image to obtain position information of dust and foreign matters on the surface of the semiconductor device and a dust and foreign matter image;
and picking up a dust removing rod with corresponding specification according to the area of the dust and foreign matter image, and driving the dust removing rod to be contacted with the dust and foreign matter according to the position information of the dust and foreign matter, and sticking and removing the dust and the foreign matter.
Further, the method further comprises: and detecting the position information and the dust and foreign matter images of the dust and foreign matters on the surface of the semiconductor device again, and replacing the dust removing rod to contact and adhere the dust and foreign matters until all the dust and foreign matters are removed.
Further, before the step of picking up the dust removing rods with corresponding specifications according to the area size of the dust and foreign matter image, the method further includes:
calculating and counting the area sizes of all the dust and foreign matter images in the historical data, dividing the dust and foreign matter images into a plurality of numerical intervals according to the area sizes of the dust and foreign matter images, and setting the outer diameter value of the viscous semisolid part in the dust removing rod according to the threshold value of the numerical intervals;
and preparing the dust removing rod with corresponding specification in advance according to the external diameter value of the viscous semi-solid part.
Further, the step of driving the dust removing rod to contact with and adhere to the dust and foreign matters according to the position information of the dust and foreign matters includes:
and driving the dust removing rod to reach the position above the dust and foreign matters according to the position information of the dust and foreign matters, and downwards moving the dust removing rod according to the initial height value of the dust removing rod and the surface of the semiconductor device and the maximum depth value of the viscous semisolid part in the dust removing rod so that the viscous semisolid part is covered and contacted with the dust and foreign matters and is adhered and removed.
Further, the step of picking up the dust removing rod with the corresponding specification according to the area size of the dust and foreign matter image comprises the following steps:
and picking up the dust removing rod according to the area of the dust and foreign matter image, so that the dust and foreign matter is covered on the section of the dust removing rod where the outer diameter value of the viscous semisolid part is located.
Further, the dust removing rod for picking up corresponding specifications according to the area size of the dust and foreign matter image further comprises:
when the viscous semisolid part of the dust removing rod with the largest specification cannot cover the dust and foreign matters, the dust removing rod with the largest specification is picked up to contact and adhere the dust and foreign matters from the dust and foreign matters center position;
and detecting the position information and the dust and foreign matter images of the dust and foreign matters on the surface of the semiconductor device again, and replacing the dust removing rod to contact and adhere the dust and foreign matters until all the dust and foreign matters are removed.
Further, the step of matching and comparing the real-time image with a preset reference image to obtain position information of dust and foreign matters on the surface of the semiconductor device and a dust and foreign matters image includes:
carrying out gray level processing and binarization processing on the real-time image of the semiconductor device to obtain a corresponding black-and-white image, and obtaining a device effective area in the real-time image from the black-and-white image by using a connected area extraction algorithm;
calculating and obtaining the maximum value Mmax in a set M formed by the gray values of all pixels in the real-time image and the maximum value Nmax in a set N formed by the gray values of all pixels in a preset reference image, wherein the difference D=Mmax-Nmax between the maximum value Mmax and the maximum value Nmax; adding D to the gray value Mi of each pixel in the effective area of the device to obtain a comparison value M' i; wherein i is a natural number greater than 1;
comparing the comparison value M 'i with a gray value Ni at a position corresponding to the preset reference image, and if the deviation value S= |M' i-Ni| is larger than a preset threshold value, marking pixels in an effective area of the device represented by Mi as dust foreign matter point pixels; and calculating the area of a communication area of each dust and foreign object according to the dust and foreign object point pixels to obtain a dust and foreign object image, wherein the central point of the dust and foreign object image is the position information of the dust and foreign object.
Further, the step of obtaining the device effective area in the real-time image from the black-and-white image by using the connected area extraction algorithm includes:
scanning a first row of a black-and-white image corresponding to the real-time image, defining a continuous white pixel sequence as a block, initializing an empty forest, allocating a label for each block as a tree of a single node, adding the single node into the forest, and simultaneously recording the labels, the line numbers, the starting points and the end points of the blocks;
sequentially scanning black-and-white images corresponding to the real-time images from a second row, distributing new labels for each current block, recording labels, row numbers, starting points and end points of the current blocks, adding the current blocks into a forest, judging whether the current blocks and the blocks of the previous row have intersections, and if so, merging the tree where the blocks of the previous row having the intersections and the tree where the current blocks are located;
traversing all trees in the forest, wherein the number of the trees is the number of the connected areas; traversing the nodes in each tree, and recording the minimum line number, the maximum line number, the minimum value of the starting point and the maximum value of the end point, wherein the four values are the upper, lower, left and right boundaries of each communication area; traversing the nodes of each tree, and adding the sizes of the blocks represented by each node to obtain the size of a connected area;
and finding out the size and the boundary of each connected region, and cutting the original image of the real-time image according to the boundary value of the maximum connected region to obtain the required device effective region.
Further, before the step of capturing a real-time image of the semiconductor device, the method includes:
setting an XYZ reference coordinate system, wherein a plurality of semiconductor devices are placed on a device placing table, and the position information of each semiconductor device on the device placing table is Xsi, YI, zsi; arranging a plurality of dust removing rods with different outer diameter specifications in groups, and then placing the dust removing rods on a dust removing rod placing table, wherein the position information of each dust removing rod on the dust removing rod placing table is Xcij, YIij and Zcij; where i, j are natural numbers greater than 1.
Further, the capturing the real-time image of the semiconductor device includes:
according to the position information Xsi, ysi and Zsi of each semiconductor device on the device placing table, a CCD visual detection module is arranged right above each semiconductor device, and the CCD visual detection module is adjusted to a preset height and then acquires a real-time image of each semiconductor device;
and according to the position information Xcij, YIij and Zcij of each dust removing rod on the dust removing rod placing table, picking up the dust removing rods with corresponding specifications from the dust removing rod placing table by utilizing a dust pick-up dust sticking module according to the area size of the dust and foreign matter images, and driving the dust removing rods to be contacted with and stuck to remove the dust and foreign matter.
The invention provides a semiconductor device surface dust removing system based on visual detection, which comprises a working platform, wherein a device placing table and a dust removing rod placing table are arranged on the working platform, and a CCD visual detection module and a dust pick-up and sticking module are respectively arranged corresponding to the device placing table and the dust removing rod placing table;
the CCD visual detection module is used for moving to the device placement table to acquire real-time images of each semiconductor device;
the dust collecting and adhering module is used for collecting the dust collecting rod from the dust collecting rod placing table, and driving the dust collecting rod to contact with and adhere to the dust and foreign matters according to the position information of the dust and foreign matters; when the dust and foreign matters are stuck and removed, driving the dust removing rod to reset and replacing the new dust removing rod to contact and stick and remove the other dust and foreign matters until all the dust and foreign matters are removed;
and the processing control terminal is used for receiving the real-time image of the semiconductor device, matching and comparing the real-time image with a preset reference image to obtain the position information of each dust and foreign matter on the semiconductor device and the dust and foreign matter image, matching the dust removing rods with corresponding specifications according to the area size of the dust and foreign matter image, and issuing an execution instruction to the CCD visual detection module and the pick-up dust sticking module.
The technical scheme of the invention has the beneficial effects that:
according to the semiconductor device surface dust removing method and the system based on visual detection, the dust and foreign matters on the surface of the semiconductor device are rapidly identified by adopting an image identification mode, the position information and the dust and foreign matters image of each dust and foreign matters are calculated, then the area size of the dust and foreign matters image is continuously collected and grouped by adopting a machine learning and statistical analysis mode, dust removing rods with a plurality of outer diameter specifications are arranged according to grouping conditions and are preset to be prepared and then are placed on a working platform, and the best matched dust removing rods are selected to contact and adhere according to the dust and foreign matters image obtained in real time, so that the dust removing work can be rapidly completed, meanwhile, the consumption of the dust removing rods is minimum, and the contact times of the semiconductor device and the dust removing rods are minimized during dust removing, so that the foreign matters on the surface of the device are rapidly removed, and the damage or secondary pollution on the surface of the device is also maximally avoided.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
Fig. 1 is a flow chart of a method for removing dust on a surface of a semiconductor device based on visual inspection according to an embodiment of the present invention;
FIG. 2 is a flowchart of a method for obtaining position information of dust and foreign matter on a surface of a semiconductor device according to an embodiment of the present invention;
FIG. 3 is a flow chart of obtaining an effective area of a device in a real-time image based on a connected area in accordance with an embodiment of the present invention;
fig. 4 is a schematic diagram of a hardware structure of a surface dust removing system of a semiconductor device based on visual inspection according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a hardware structure of a surface dust removing system of a semiconductor device based on visual inspection according to an embodiment of the present invention;
fig. 6 is a schematic view of a dust removing rod with different outer diameter specifications according to an embodiment of the present invention.
Description of the embodiments
The following describes the embodiments of the present invention further with reference to the drawings. The description of these embodiments is provided to assist understanding of the present invention, but is not intended to limit the present invention. In addition, the technical features of the embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.
Example 1
As shown in fig. 1, an embodiment of the present invention provides a method for removing dust on a surface of a semiconductor device based on visual inspection, the method including:
s101, acquiring a real-time image of a semiconductor device, and matching and comparing the real-time image with a preset reference image to obtain position information of dust and foreign matters on the surface of the semiconductor device and a dust and foreign matter image;
s102, picking up a dust removing rod with corresponding specification according to the area of the dust and foreign matter image, and driving the dust removing rod to contact and adhere to the dust and foreign matter according to the position information of the dust and foreign matter;
specifically, before the step of picking up the dust removing rods with corresponding specifications according to the area size of the dust and foreign matter image, the method further comprises: calculating and counting the area sizes of all the dust and foreign matter images in the historical data, dividing the dust and foreign matter images into a plurality of numerical intervals according to the area sizes of the dust and foreign matter images, and setting the outer diameter value of the viscous semisolid part in the dust removing rod according to the threshold value of the numerical intervals;
and preparing the dust removing rod with corresponding specification in advance according to the external diameter value of the viscous semi-solid part.
In production practice, the external diameter value of the viscous semisolid part in the dust removing rod aims at different dust and foreign matters, and is not larger and better for the dust and foreign matters with different sizes, if the external diameter value is larger, the contact surface is larger, secondary pollution is easy to be caused, and in addition, the dust removing rod is used as a consumable, so that material waste is caused. If the outer diameter value is smaller, more contact times are needed to be completely stuck and removed, so that secondary pollution is easy to bring, in addition, the work efficiency is low due to the fact that the contact times are more, the dust removing rods are too much to use, and material waste is caused.
In the embodiment of the invention, historical data obtained by detecting dust and foreign matters of the same batch and type of semiconductor devices are statistically analyzed in the dust removing process of the semiconductor devices, the dust removing rods with a plurality of outer diameter values are grouped according to the area size distribution condition of dust and foreign matters images, the number of the dust removing rods with the outer diameter values is determined according to the grouping condition, and the dust removing rods with corresponding specifications are prepared in advance according to the area size distribution condition.
Specifically, the dust removing rod is driven to reach the upper part of the dust and foreign matters according to the position information of the dust and foreign matters, and the dust removing rod is moved downwards according to the initial height value of the dust removing rod and the surface of the semiconductor device and the maximum depth value of the viscous semisolid part in the dust removing rod so that the viscous semisolid part is covered and contacted with the dust and foreign matters and is adhered and removed. Because the outer diameter values of the viscous semisolid parts in the dust removing rod are different, the maximum depth value is the height required to move when the section where the outer diameter value is located contacts with dust and foreign matters when the dust removing rod moves downwards, so that the viscous semisolid parts can be fully utilized, and dust and foreign matters are removed by sticking.
S103, detecting the position information and the dust and foreign matter images of the dust and foreign matters on the surface of the semiconductor device again, and replacing the dust removing rod to contact and adhere the dust and foreign matters until all the dust and foreign matters are removed.
Optionally, the step of picking up the dust removing rod with the corresponding specification according to the area size of the dust and foreign matter image includes:
and picking up the dust removing rod according to the area of the dust and foreign matter image, so that the dust and foreign matter is covered on the section of the dust removing rod where the outer diameter value of the viscous semisolid part is located.
When the viscous semisolid part of the dust removing rod with the largest specification cannot cover the dust and foreign matters, the dust removing rod with the largest specification is picked up to contact and adhere the dust and foreign matters from the dust and foreign matters center position;
and detecting the position information and the dust and foreign matter images of the dust and foreign matters on the surface of the semiconductor device again, and replacing the dust removing rod to contact and adhere the dust and foreign matters until all the dust and foreign matters are removed.
As shown in fig. 2, the step of matching and comparing the real-time image with a preset reference image to obtain position information of dust and foreign matters on the surface of the semiconductor device and a dust and foreign matters image includes:
s201, carrying out gray level processing and binarization processing on the real-time image of the semiconductor device to obtain a corresponding black-and-white image, and obtaining a device effective area in the real-time image from the black-and-white image by using a connected area extraction algorithm;
the real-time image of the semiconductor device comprises a device effective area of the semiconductor device and a background area, and what is needed is the device effective area of the semiconductor device to perform gray-scale processing and binarization processing on the image so as to accelerate the image processing speed.
Initially, the real-time image of the semiconductor device is a color image, each pixel point in the image contains three color components, and the image size is large, so that the calculation amount of subsequent processing is large. For this purpose, three-channel color images are converted into single-channel gray images, each pixel in the gray image only needs to occupy one byte to store gray values, and the gray values range from 0 to 255.
The embodiment of the invention adopts a maximum value method to carry out gray processing on a real-time image of a semiconductor device, wherein the maximum value method takes the maximum value of three channel components on each pixel point as the gray value of the current pixel point, and the formula is as follows:
。
the image binarization is a process of resetting the pixel value of a gray image to 0 or 255 according to a certain rule so that only black and white colors exist in the image. The binarization greatly reduces the data volume of the image under the condition of keeping a certain information volume, and is beneficial to shortening the subsequent image processing process. According to the embodiment of the invention, the real-time image of the semiconductor device and the preset reference image are segmented through the fixed threshold G to obtain the corresponding black-and-white image.
Practice shows that the effective area of the device in the real-time image is larger, the area of the residual noise area is smaller, the effective area of the device in the real-time image is obtained based on the connected area, and the connected area refers to a set formed by points which have the same gray value and are adjacent in position in the image. As shown in fig. 3, the specific steps are as follows:
s301, scanning a first row of a black-and-white image corresponding to the real-time image, defining a continuous white pixel sequence as a block (BDob), initializing an empty forest (a set of trees), allocating a label to each block as a tree with a single node, adding the label into the forest, and recording the label, the line number, the starting point and the end point of the block;
s302, sequentially scanning black-and-white images corresponding to the real-time images from a second row, distributing new labels for each current block, recording labels, row numbers, starting points and end points of the current block, adding the current block into a forest, judging whether the current block and the previous row of blocks have an intersection, and merging (Union) a tree where the block with the intersection in the previous row and a tree where the current block are located if the intersection exists;
s303, traversing all trees in the forest, wherein the number of the trees is the number of the connected areas; traversing the nodes in each tree, and recording the minimum line number, the maximum line number, the minimum value of the starting point and the maximum value of the end point, wherein the four values are the upper, lower, left and right boundaries of each communication area; traversing the nodes of each tree, and adding the sizes (the difference between the end point and the starting point) of the blocks represented by each node to obtain the size of a connected area;
s304, finding out the size and the boundary of each connected region, and cutting the original image of the real-time image according to the boundary value of the maximum connected region to obtain the required device effective region.
S202, calculating and obtaining the maximum value Mmax in a set M formed by gray values of all pixels in the real-time image and the maximum value Nmax in a set N formed by gray values of all pixels in a preset reference image, wherein the difference D=Mmax-Nmax; adding D to the gray value Mi of each pixel in the effective area of the device to obtain a comparison value M' i; wherein i is a natural number greater than 1;
s203, comparing the comparison value M 'i with a gray value Ni at a position corresponding to the preset reference image, and if the deviation value S= I M' i-Ni is greater than a preset threshold value, marking pixels in an effective area of the device represented by Mi as dust foreign matter point pixels; and calculating the area of a communication area of each dust and foreign object according to the dust and foreign object point pixels to obtain a dust and foreign object image, wherein the central point of the dust and foreign object image is the position information of the dust and foreign object.
Specifically, before the step of capturing a real-time image of the semiconductor device, the method includes:
setting an XYZ reference coordinate system, wherein a plurality of semiconductor devices are placed on a device placing table, and the position information of each semiconductor device on the device placing table is Xsi, YI, zsi; arranging a plurality of dust removing rods with different outer diameter specifications in groups, and then placing the dust removing rods on a dust removing rod placing table, wherein the position information of each dust removing rod on the dust removing rod placing table is Xcij, YIij and Zcij; where i, j are natural numbers greater than 1.
Specifically, the capturing the real-time image of the semiconductor device includes:
according to the position information Xsi, ysi and Zsi of each semiconductor device on the device placing table, a CCD visual detection module is arranged right above each semiconductor device, and the CCD visual detection module is adjusted to a preset height and then acquires a real-time image of each semiconductor device;
and according to the position information Xcij, YIij and Zcij of each dust removing rod on the dust removing rod placing table, picking up the dust removing rods with corresponding specifications from the dust removing rod placing table by utilizing a dust pick-up dust sticking module according to the area size of the dust and foreign matter images, and driving the dust removing rods to be contacted with and stuck to remove the dust and foreign matter.
Example 2
As shown in fig. 4 and 5, the embodiment of the invention further provides a semiconductor device surface dust removing system based on visual detection, the system comprises a working platform 10, a device placing table 20 and a dust removing rod placing table 30 are arranged on the working platform 10, and a CCD visual detection module 40 and a dust picking and sticking module 50 are respectively arranged corresponding to the device placing table 20 and the dust removing rod placing table 30;
the CCD visual detection module 40 is used for moving to the device placement table 20 to acquire real-time images of each semiconductor device 90;
the dust pick-up and sticking module 50 is configured to pick up the dust removing rod 100 from the dust removing rod placement table 30, and drive the dust removing rod to contact with and stick out the dust and foreign matters according to the position information of the dust and foreign matters; when the dust and foreign matters are stuck and removed, driving the dust removing rod 100 to reset and replacing the dust removing rod 100 to contact and stick and remove the other dust and foreign matters until all the dust and foreign matters are removed;
the processing control terminal 80 is configured to receive the real-time image of the semiconductor device, match and compare the real-time image with a preset reference image, obtain location information of each dust and foreign matter on the semiconductor device 90 and a dust and foreign matter image, match a dust removing bar 100 with a corresponding specification according to the area size of the dust and foreign matter image, and send an execution instruction to the CCD vision detection module 40 and the pick-up dust sticking module 50.
Specifically, the system further includes a first triaxial moving module 60 and a second triaxial moving module 70, where the first triaxial moving module 60 and the second triaxial moving module 70 respectively drive the CCD vision detecting module 40 and the dust-collecting and dust-sticking picking module 50 to spatially move on the working platform 10.
Optionally, the CCD vision detecting module 40 is further provided with a ranging sensor, which is used for acquiring a height value between the CCD camera and the surface of the semiconductor device on the device placing table.
As shown in fig. 5, the dust removing bars 100 having different outer diameter value specifications are arranged in groups on the dust removing bar placing table 30, and the number relationship of the dust removing bars 100 of different outer diameter value specifications is determined by the dust foreign matter area size statistical grouping result of the semiconductor device.
As shown in fig. 6, the dust removing rod 100 includes a dust removing body 110 and a viscous semisolid portion 120 formed by drying a dust-binding liquid attached to the dust removing body 110; the dust removing rods 100 with different outer diameter values and specifications have the same dust removing body 110, and the size of the area of the communication area of each dust and foreign matter in the dust and foreign matter image is obtained, so that the maximum depth value Hj of the viscous semisolid portion 120 and the cross-sectional area of the outer diameter Dj cover the area of the communication area of each dust and foreign matter, and the outer diameter Dj of the viscous semisolid portion 120 can be determined according to the size of the area of the communication area of each dust and foreign matter, and further, the natural number with the maximum depth value Hj being greater than 1, that is, the specification number of the dust removing rod 100 is determined.
According to the semiconductor device surface dust removing method and the system based on visual detection, the dust and foreign matters on the surface of the semiconductor device are rapidly identified by adopting an image identification mode, the position information and the dust and foreign matters image of each dust and foreign matters are calculated, then the area size of the dust and foreign matters image is continuously collected and grouped by adopting a machine learning and statistical analysis mode, dust removing rods with a plurality of outer diameter specifications are arranged according to grouping conditions and are preset to be prepared and then are placed on a working platform, and the best matched dust removing rods are selected to contact and adhere according to the dust and foreign matters image obtained in real time, so that the dust removing work can be rapidly completed, meanwhile, the consumption of the dust removing rods is minimum, and the contact times of the semiconductor device and the dust removing rods are minimized during dust removing, so that the foreign matters on the surface of the device are rapidly removed, and the damage or secondary pollution on the surface of the device is also maximally avoided.
The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. It will be apparent to those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, and yet fall within the scope of the invention.
Claims (10)
1. A method for removing dust from a surface of a semiconductor device based on visual inspection, the method comprising:
acquiring a real-time image of a semiconductor device, and matching and comparing the real-time image with a preset reference image to obtain position information of dust and foreign matters on the surface of the semiconductor device and a dust and foreign matter image;
and picking up a dust removing rod with corresponding specification according to the area of the dust and foreign matter image, and driving the dust removing rod to be contacted with the dust and foreign matter according to the position information of the dust and foreign matter, and sticking and removing the dust and the foreign matter.
2. The visual inspection-based semiconductor device surface dust removal method according to claim 1, further comprising: and detecting the position information and the dust and foreign matter images of the dust and foreign matters on the surface of the semiconductor device again, and replacing the dust removing rod to contact and adhere the dust and foreign matters until all the dust and foreign matters are removed.
3. The method for removing dust from a surface of a semiconductor device based on visual inspection according to claim 1, wherein before the step of picking up dust removing bars of corresponding specifications according to the area size of the dust and foreign matter image, the method further comprises:
calculating and counting the area sizes of all the dust and foreign matter images in the historical data, dividing the dust and foreign matter images into a plurality of numerical intervals according to the area sizes of the dust and foreign matter images, and setting the outer diameter value of the viscous semisolid part in the dust removing rod according to the threshold value of the numerical intervals;
and preparing the dust removing rod with corresponding specification in advance according to the external diameter value of the viscous semi-solid part.
4. The method for removing dust from a surface of a semiconductor device based on visual inspection according to claim 1, wherein the step of driving the dust removing bar to contact and adhere to the dust and foreign substances according to the positional information of the dust and foreign substances comprises:
and driving the dust removing rod to reach the position above the dust and foreign matters according to the position information of the dust and foreign matters, and downwards moving the dust removing rod according to the initial height value of the dust removing rod and the surface of the semiconductor device and the maximum depth value of the viscous semisolid part in the dust removing rod so that the viscous semisolid part is covered and contacted with the dust and foreign matters and is adhered and removed.
5. The method for removing dust from a surface of a semiconductor device based on visual inspection according to claim 3, wherein the step of picking up the dust removing bars of the corresponding specification according to the area size of the dust foreign matter image comprises:
and picking up the dust removing rod according to the area of the dust and foreign matter image, so that the dust and foreign matter is covered on the section of the dust removing rod where the outer diameter value of the viscous semisolid part is located.
6. The visual inspection-based surface dust removing method of a semiconductor device according to claim 3, wherein picking up a dust removing bar of a corresponding specification according to an area size of the dust foreign matter image further comprises:
when the viscous semisolid part of the dust removing rod with the largest specification cannot cover the dust and foreign matters, the dust removing rod with the largest specification is picked up to contact and adhere the dust and foreign matters from the dust and foreign matters center position;
and detecting the position information and the dust and foreign matter images of the dust and foreign matters on the surface of the semiconductor device again, and replacing the dust removing rod to contact and adhere the dust and foreign matters until all the dust and foreign matters are removed.
7. The method for removing dust from a surface of a semiconductor device based on visual inspection according to claim 1, wherein the step of matching and comparing the real-time image with a preset reference image to obtain position information of dust and foreign matter on the surface of the semiconductor device and the dust and foreign matter image comprises:
carrying out gray level processing and binarization processing on the real-time image of the semiconductor device to obtain a corresponding black-and-white image, and obtaining a device effective area in the real-time image from the black-and-white image by using a connected area extraction algorithm;
calculating and obtaining the maximum value Mmax in a set M formed by the gray values of all pixels in the real-time image and the maximum value Nmax in a set N formed by the gray values of all pixels in a preset reference image, wherein the difference D=Mmax-Nmax between the maximum value Mmax and the maximum value Nmax; adding D to the gray value Mi of each pixel in the effective area of the device to obtain a comparison value M' i; wherein i is a natural number greater than 1;
comparing the comparison value M 'i with a gray value Ni at a position corresponding to the preset reference image, and if the deviation value S= |M' i-Ni| is larger than a preset threshold value, marking pixels in an effective area of the device represented by Mi as dust foreign matter point pixels; and calculating the area of a communication area of each dust and foreign object according to the dust and foreign object point pixels to obtain a dust and foreign object image, wherein the central point of the dust and foreign object image is the position information of the dust and foreign object.
8. The method for removing dust from a surface of a semiconductor device based on visual inspection according to claim 7, wherein the step of obtaining a device effective area in a real-time image from the black-and-white image using a connected area extraction algorithm comprises:
scanning a first row of a black-and-white image corresponding to the real-time image, defining a continuous white pixel sequence as a block, initializing an empty forest, allocating a label for each block as a tree of a single node, adding the single node into the forest, and simultaneously recording the labels, the line numbers, the starting points and the end points of the blocks;
sequentially scanning black-and-white images corresponding to the real-time images from a second row, distributing new labels for each current block, recording labels, row numbers, starting points and end points of the current blocks, adding the current blocks into a forest, judging whether the current blocks and the blocks of the previous row have intersections, and if so, merging the tree where the blocks of the previous row having the intersections and the tree where the current blocks are located;
traversing all trees in the forest, wherein the number of the trees is the number of the connected areas; traversing the nodes in each tree, and recording the minimum line number, the maximum line number, the minimum value of the starting point and the maximum value of the end point, wherein the four values are the upper, lower, left and right boundaries of each communication area; traversing the nodes of each tree, and adding the sizes of the blocks represented by each node to obtain the size of a connected area;
and finding out the size and the boundary of each connected region, and cutting the original image of the real-time image according to the boundary value of the maximum connected region to obtain the required device effective region.
9. The method for removing dust from a surface of a semiconductor device based on visual inspection according to claim 1, wherein before the step of capturing a real-time image of the semiconductor device, the method comprises:
setting an XYZ reference coordinate system, wherein a plurality of semiconductor devices are placed on a device placing table, and the position information of each semiconductor device on the device placing table is Xsi, YI, zsi; arranging a plurality of dust removing rods with different outer diameter specifications in groups, and then placing the dust removing rods on a dust removing rod placing table, wherein the position information of each dust removing rod on the dust removing rod placing table is Xcij, YIij and Zcij; wherein i, j are natural numbers greater than 1;
according to the position information Xsi, ysi and Zsi of each semiconductor device on the device placing table, a CCD visual detection module is arranged right above each semiconductor device, and the CCD visual detection module is adjusted to a preset height and then acquires a real-time image of each semiconductor device;
and according to the position information Xcij, YIij and Zcij of each dust removing rod on the dust removing rod placing table, picking up the dust removing rods with corresponding specifications from the dust removing rod placing table by utilizing a dust pick-up dust sticking module according to the area size of the dust and foreign matter images, and driving the dust removing rods to be contacted with and stuck to remove the dust and foreign matter.
10. The surface dust removing system for the semiconductor device based on visual detection is characterized by comprising a working platform, wherein a device placing table and a dust removing rod placing table are arranged on the working platform, and a CCD visual detection module and a dust pick-up dust sticking module are respectively arranged corresponding to the device placing table and the dust removing rod placing table;
the CCD visual detection module is used for moving to the device placement table to acquire real-time images of each semiconductor device;
the dust collecting and adhering module is used for collecting the dust collecting rod from the dust collecting rod placing table, and driving the dust collecting rod to contact with and adhere to the dust and foreign matters according to the position information of the dust and foreign matters; when the dust and foreign matters are stuck and removed, driving the dust removing rod to reset and replacing the new dust removing rod to contact and stick and remove the other dust and foreign matters until all the dust and foreign matters are removed;
and the processing control terminal is used for receiving the real-time image of the semiconductor device, matching and comparing the real-time image with a preset reference image to obtain the position information of each dust and foreign matter on the semiconductor device and the dust and foreign matter image, matching the dust removing rods with corresponding specifications according to the area size of the dust and foreign matter image, and issuing an execution instruction to the CCD visual detection module and the pick-up dust sticking module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311781372.8A CN117470104B (en) | 2023-12-22 | 2023-12-22 | Semiconductor device surface dust removing method and system based on visual detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311781372.8A CN117470104B (en) | 2023-12-22 | 2023-12-22 | Semiconductor device surface dust removing method and system based on visual detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117470104A true CN117470104A (en) | 2024-01-30 |
| CN117470104B CN117470104B (en) | 2024-04-05 |
Family
ID=89638194
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311781372.8A Active CN117470104B (en) | 2023-12-22 | 2023-12-22 | Semiconductor device surface dust removing method and system based on visual detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117470104B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106391624A (en) * | 2016-07-01 | 2017-02-15 | 国网河南省电力公司漯河供电公司 | Self-induction solar photovoltaic dust removing system |
| CN111781215A (en) * | 2020-06-23 | 2020-10-16 | 深圳赋元软件有限公司 | Viscidity dust collector |
| CN213701071U (en) * | 2020-11-26 | 2021-07-16 | 吴珍 | Multipurpose dust-binding roller |
| CN115690670A (en) * | 2022-09-23 | 2023-02-03 | 盛吉盛(宁波)半导体科技有限公司 | Intelligent identification method and system for wafer defects |
| CN115937065A (en) * | 2022-04-20 | 2023-04-07 | 深圳大学 | Foreign matter detection method, device and equipment of display module and storage medium |
| CN219740762U (en) * | 2023-05-22 | 2023-09-22 | 杭州荣宏电子科技有限公司 | PCB circuit board dust collector for electronic technology |
| CN117399366A (en) * | 2023-09-07 | 2024-01-16 | 中科见微智能装备(苏州)有限公司 | Method and system for detecting and processing surface flaws and foreign matters of semiconductor device |
-
2023
- 2023-12-22 CN CN202311781372.8A patent/CN117470104B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106391624A (en) * | 2016-07-01 | 2017-02-15 | 国网河南省电力公司漯河供电公司 | Self-induction solar photovoltaic dust removing system |
| CN111781215A (en) * | 2020-06-23 | 2020-10-16 | 深圳赋元软件有限公司 | Viscidity dust collector |
| CN213701071U (en) * | 2020-11-26 | 2021-07-16 | 吴珍 | Multipurpose dust-binding roller |
| CN115937065A (en) * | 2022-04-20 | 2023-04-07 | 深圳大学 | Foreign matter detection method, device and equipment of display module and storage medium |
| CN115690670A (en) * | 2022-09-23 | 2023-02-03 | 盛吉盛(宁波)半导体科技有限公司 | Intelligent identification method and system for wafer defects |
| CN219740762U (en) * | 2023-05-22 | 2023-09-22 | 杭州荣宏电子科技有限公司 | PCB circuit board dust collector for electronic technology |
| CN117399366A (en) * | 2023-09-07 | 2024-01-16 | 中科见微智能装备(苏州)有限公司 | Method and system for detecting and processing surface flaws and foreign matters of semiconductor device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117470104B (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107804514B (en) | Toothbrush sorting method based on image recognition | |
| CN110084241B (en) | Automatic ammeter reading method based on image recognition | |
| CN108960011B (en) | Partially-shielded citrus fruit image identification method | |
| CN108126914B (en) | Deep learning-based robot sorting method for scattered multiple objects in material frame | |
| CN112837290B (en) | Crack image automatic identification method based on seed filling algorithm | |
| CN107101598B (en) | Automatic detection method and device for concentricity quality of piezoelectric ceramic silver sheet | |
| CN109522767B (en) | One-dimensional code breaking visual detection method | |
| CN111570315A (en) | Tableware processing device, control method of processing device and control device | |
| WO2021135675A1 (en) | Automatic bar counting method and apparatus | |
| CN112884718A (en) | Method, device and system for detecting code spraying characters of package and storage medium | |
| CN111627004A (en) | Detection method and detection device for cleanliness of tableware, storage medium and processor | |
| CN114235837A (en) | LED packaging surface defect detection method, device, medium and equipment based on machine vision | |
| CN102660815A (en) | Cotton foreign fiber automatic removing method and system based on image acquisition and DSP (digital signal processing) algorithm | |
| CN117399366A (en) | Method and system for detecting and processing surface flaws and foreign matters of semiconductor device | |
| CN117470104B (en) | Semiconductor device surface dust removing method and system based on visual detection | |
| CN109509173A (en) | A kind of method of counting and counting device of bacterial clump | |
| CN110726720A (en) | Method for detecting suspended substances in drinking mineral water | |
| CN114332041A (en) | U-Net-based polished tile surface flaw detection method and device | |
| CN108734054B (en) | Non-shielding citrus fruit image identification method | |
| CN109596620A (en) | Product surface shape defect detection method and system based on machine vision | |
| CN112338898A (en) | Image processing method and device of object sorting system and object sorting system | |
| CN104574398B (en) | A kind of vanning cigarette filter stick number system based on line scan camera | |
| CN110909751B (en) | Visual identification method, system and medium for transformer substation insulator cleaning robot | |
| CN117132655A (en) | Filling barrel opening position measuring method based on machine vision | |
| CN112541367A (en) | Multiple two-dimensional code identification method based on deep learning and image processing |
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 |