[go: up one dir, main page]

WO2024190191A1 - Procédé de traitement d'image - Google Patents

Procédé de traitement d'image Download PDF

Info

Publication number
WO2024190191A1
WO2024190191A1 PCT/JP2024/004036 JP2024004036W WO2024190191A1 WO 2024190191 A1 WO2024190191 A1 WO 2024190191A1 JP 2024004036 W JP2024004036 W JP 2024004036W WO 2024190191 A1 WO2024190191 A1 WO 2024190191A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
processed image
image processing
processing method
defects
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/JP2024/004036
Other languages
English (en)
Japanese (ja)
Inventor
直輝 卜部
篤志 杉谷
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Toray Engineering Co Ltd
Tasmit Inc
Original Assignee
Toray Engineering Co Ltd
Tasmit Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Toray Engineering Co Ltd, Tasmit Inc filed Critical Toray Engineering Co Ltd
Priority to CN202480014661.1A priority Critical patent/CN120752518A/zh
Publication of WO2024190191A1 publication Critical patent/WO2024190191A1/fr
Anticipated expiration legal-status Critical
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
    • G01N21/956—Inspecting patterns on the surface of objects
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration using local operators
    • G06T5/30—Erosion or dilatation, e.g. thinning

Definitions

  • the present invention relates to an image processing method.
  • Images are used to detect defects in the manufacturing process of various products.
  • Patent Document 1 discloses a highly sensitive defect inspection method for inspecting patterns to detect defects in memory mats and peripheral circuits of semiconductor wafers.
  • Various processes are involved in the manufacture of semiconductors, and various defects occur in each process.
  • cleaning processes are performed frequently, and defects caused by residual cleaning fluid occur in many cleaning processes, so it is necessary to detect them to eliminate defective products and provide feedback to the manufacturing conditions of the process.
  • Semiconductor chips have wiring patterns, so in order to detect defects, it is necessary to distinguish between the wiring pattern and the defect.
  • an image of only the wiring pattern is prepared, and this is compared with an image used for defect detection, and image processing is performed to detect the defects.
  • the brightness value of the wiring will decrease due to the widening of the wiring caused by manufacturing variations, and if the wiring being inspected is narrow (brightness is high), the surrounding area will be judged to be defective, resulting in the problem of false defects. Also, a defect that exists across a wiring will be detected as separate defects on both sides of the wiring, making it impossible to grasp the actual number and size of the defects.
  • the present invention has been made in consideration of these points, and its purpose is to provide an image processing method that detects defect images present in an image while eliminating the influence of the wiring pattern when performing image-based inspection of work-in-progress or finished products such as semiconductor chips that have wiring patterns.
  • the image processing method of the present invention is an image processing method for detecting an irregular image displayed on a photographic screen on which at least one straight line extending in one direction is displayed, and includes the steps of: obtaining a first processed image by filtering the photographic screen using a smoothing filter with a segment defined by a first predetermined number a of pixels aligned along the one direction and a second predetermined number b of pixels aligned along another direction different from the one direction; obtaining a second processed image by filtering the photographic screen using a smoothing filter with a segment defined by a third predetermined number c of pixels aligned along the one direction and a fourth predetermined number d of pixels aligned along the other direction; obtaining a third processed image, which is a difference image, by subtracting the first processed image from the second processed image; and detecting the irregular image using the third processed image, wherein the number a is greater than any of the numbers b, c, and d.
  • the first predetermined number a may be any number less than the total number of pixels aligned along the one direction and greater than b, c, and d.
  • the second predetermined number b, the third predetermined number c, and the fourth predetermined number d are preferably 1/5 or less of the total number of pixels aligned in one direction, and more preferably 1/10 or less.
  • the detection step it is preferable to perform binarization processing.
  • the detection step it is preferable to perform a closing process after the binarization process to combine the multiple irregular images divided by the straight lines.
  • the detection step it is preferable to perform a labeling process after the closing process.
  • FIG. 1 is a diagram showing a photographed image for inspection according to an embodiment
  • FIG. 2 is a diagram showing a first processed image that is a result of performing averaging filter processing 1 on a captured image for inspection.
  • 13 is a diagram showing a second processed image that is a result of performing averaging filter processing 2 on the captured image for inspection.
  • FIG. 13 is a diagram showing an image obtained by performing binarization and closing processing on a third processed image, which is a difference image obtained by subtracting the first processed image from the second processed image.
  • FIG. FIG. 2 is an example of a flow diagram according to an embodiment.
  • the first embodiment relates to image processing performed on captured images for inspecting an individual chip or a wafer in the intermediate or completed stage of manufacturing a semiconductor device (semiconductor chip).
  • Figure 1 shows a photograph of a wafer after completion of a certain process in the manufacture of a semiconductor device.
  • the wafer On the wafer, there are five wiring patterns 10, 12, 14, 16, and 18, as well as defects 20, 22, 24, 26, and 28 that result from some of the cleaning solution not being completely removed, and in addition, the photographed image contains tiny dots 30 and 31 caused by noise at the time of photography.
  • Each of the wiring patterns 10, 12, 14, 16, and 18 extends horizontally in the photographed image.
  • the defects 20, 22, 24, 26, and 28 are caused by residual cleaning fluid and are therefore irregular in shape, with the edge portions 20a, 22a, 24a, 26a, and 28a rising higher than the central portions 20b, 22b, 24b, 2b6, and 28b.
  • a lookup table conversion is performed on the captured image data.
  • an output value is assigned to the input data (brightness value in this case) using a table (assignment table).
  • a table assignment table.
  • Smoothing filter processing 1 and smoothing filter processing 2 are performed separately on the data that has undergone lookup table conversion.
  • Smoothing filter processing is a low-pass filter process that reduces and smooths the changes in brightness values between nearby pixels. Specifically, image processing is performed on a group of adjacent pixels called a segment (for example, a square section of 4 pixels vertical by 4 pixels horizontal). Smoothing filter processing is performed for the purposes of reducing noise and emphasizing features such as edges. There are various methods that use filters for averaging filter processing, so the appropriate filter processing can be selected according to the purpose of the judgment.
  • the number of pixels aligned along the direction in which the wiring pattern extends is defined as a first predetermined number a
  • the number of pixels aligned along the direction perpendicular to the direction in which the wiring pattern extends is defined as a second predetermined number b
  • smoothing filter process is performed on a x b segment.
  • This segment is a horizontally long rectangular segment, with a>b.
  • b is not particularly limited as long as it is smaller than a, but is preferably 1/4 or less of a, and more preferably 1/10 or less. b may be 1 or more and 30 or less. Alternatively, b is preferably the number of pixels that is 1/2 or less of the width of the wiring pattern image.
  • the more pixels in a segment the more time it will take to perform averaging filter processing 1, so the number of pixels in a segment can be selected according to the purpose. Also, when performing averaging filter processing on one segment and then moving to the next segment and performing averaging filter processing on that, for example, if the movement to the next segment is a/m pixels in the horizontal direction (m is any integer), the time it takes to process an entire image will increase as m becomes larger, so in this case too, the amount of segment movement can be selected according to the purpose.
  • the first processed image obtained by performing averaging filter processing 1 is shown in Figure 2.
  • averaging filter processing 1 minute points 30, 31 and defects 20, 22, 24, 26, 28 caused by noise disappear, leaving only wiring patterns 10, 12, 14, 16, 18.
  • the segment is a rectangle that is long in the horizontal direction, wiring patterns 10, 12, 14, 16, 18 that extend from one end to the other in the horizontal direction remain, but the other images disappear.
  • b is the number of pixels that is half or less of the width of the wiring pattern image
  • the images of wiring patterns 10, 12, 14, 16, 18 are output as images with approximately the same width as the original.
  • averaging filter process 2 the number of pixels lined up along the extension direction of the wiring pattern (horizontal direction in the figure) is set as a third predetermined number c, and the number of pixels lined up along the direction perpendicular to the extension direction of the wiring pattern (vertical direction in the figure) is set as a fourth predetermined number d, and smoothing filter process is performed on a segment of c x d.
  • a>c, a>d There are no particular restrictions on c and d as long as they are smaller than a, but c and d are preferably 1/4 or less of a, and more preferably 1/10 or less. Also, c and d may be 1 or more and 30 or less. It is preferable that d is the same as b.
  • the second processed image obtained by performing averaging filter processing 2 is shown in Figure 3.
  • averaging filter processing 2 minute points 30 and 31 caused by noise are eliminated, and wiring patterns 10, 12, 14, 16, and 18 and defects 20, 22, 24, 26, and 28 remain.
  • a difference process (inter-image subtraction process) is performed to subtract the first processed image from the second processed image, to obtain a third processed image, which is a difference image.
  • the wiring patterns 10, 12, 14, 16, and 18 disappear and only the defects 20, 22, 24, 26, and 28 remain, eliminating the problem of false defects.
  • This third processed image may be used as is to perform a detection step to detect irregular defects.
  • the third processed image becomes an image of the defects 20, 22, 24, 26, 28 separated by the wiring pattern portions.
  • the edge portions 20a, 22a, 24a, 26a, 28a of the defects 20, 22, 24, 26, 28 remain displayed with a brightness different from that of the central portions 20b, 22b, 24b, 2b6, 28b.
  • the third processed image in such a state may be usable without any problems, but if it is desired to know the exact number and size (area) of the defects, further image processing may be performed in a detection step using the third processed image, and then defect inspection may be performed.
  • Binarization is a process in which the display of each pixel is converted to white or black using a predetermined threshold value for the brightness value of each pixel. For example, it is a process in which defective parts are made white and other parts are made black.
  • the brightness difference between the edge parts 20a, 22a, 24a, 26a, 28a of the defects 20, 22, 24, 26, 28 and the central parts 20b, 22b, 24b, 2b6, 28b disappears, and only the differences between the defects 20, 22, 24, 26, 28 and the other parts are displayed.
  • Closing is a process performed on an image that has been binarized, and is a process that fills in gaps between adjacent defects (white areas) that are separated by the wiring pattern, for example, and joins the separated parts. If a single defect in the third processed image is separated into multiple parts by the wiring pattern, this process will return it to the original single defect. Note that other image processing that achieves the same results may also be used.
  • Figure 4 shows the image obtained by performing the closing process after binarization. It is believed that this image almost accurately represents only the detected defect image.
  • image processing may be applied depending on the purpose of defect detection and processing.
  • the labeling process shown in Figure 5 may be performed. Labeling is a process in which white pixels are searched for in the binarized image, and if a white pixel found is surrounded by other white pixels, both white pixels are considered to be the same object and assigned the same label number. Labeling makes it possible to recognize that defects that are far apart from each other are separate defects by their label numbers, and the number and positions of the defects can be determined.
  • Size filtering is a process that picks up only images of a specified size (predetermined length range). Three size filtering processes are shown in Figure 5, including a process that picks up pixel clusters (defects) whose horizontal length is within a specified range, a process that picks up pixel clusters (defects) whose vertical length is within a specified range, and a process that picks up pixel clusters (defects) whose area is within a specified range. This makes it possible to determine, for example, that defects larger than a specified size must be treated as defective products, but that defects smaller than that size will disappear in a later process and there is no problem.
  • the image processing of this embodiment makes it possible to easily avoid and recover from false defects that may be caused by the presence of a wiring pattern, and from the division or deformation of an existing defect image caused by the wiring pattern, when performing defect detection and judgment on a captured image that displays a wiring pattern that is at least one straight line extending in one direction.
  • image processing is performed by performing two different averaging filter processing on a single captured image, image processing can be completed in a short time. After performing the two different averaging filter processing and obtaining an image of the difference between them, conventional image processing methods can be applied to accurately detect and judge defects.
  • image processing is performed on a captured image in which wires extending horizontally are displayed, but a similar method may be used to perform image processing on a captured image in which wires extend vertically.
  • the segments used in averaging filter processing 1 should be vertically long rectangles.
  • the image that is the subject of the present invention is not limited to a captured image of a semiconductor device as long as it is a captured image showing at least one straight line extending in one direction.
  • captured images of wiring boards are also subject to the present invention.
  • the subject of image processing is not limited to a captured image showing only a straight line extending in one direction, but may show multiple straight lines extending in two or more directions.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Image Processing (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)

Abstract

L'invention concerne un procédé de traitement d'image permettant de détecter, lors d'une inspection par imagerie d'une puce semi-conductrice ou analogue munie d'un câblage, un défaut présent dans une image en éliminant l'influence du câblage.　Plus précisément, l'invention concerne un procédé de traitement d'image permettant de détecter une image irrégulière en effectuant un traitement d'image sur un écran d'imagerie sur lequel est affichée une ligne droite s'étendant dans une direction, le procédé de traitement d'image comprenant : une étape d'acquisition d'une première image traitée obtenue en effectuant un traitement par filtre de lissage sur l'écran d'imagerie à l'aide d'un segment défini par un premier nombre a de pixels agencés selon une direction et un deuxième nombre b de pixels agencés selon une direction différente de la direction ; une étape d'acquisition d'une deuxième image traitée, obtenue en effectuant un traitement par filtre de lissage sur l'écran d'imagerie à l'aide d'un segment défini par un troisième nombre c de pixels agencés dans une direction et par un quatrième nombre d de pixels agencés dans une autre direction ; une étape d'acquisition d'une troisième image traitée, qui consiste en une image obtenue en soustrayant la première image traitée de la deuxième image traitée ; et une étape de détection pour détecter une image irrégulière à l'aide de la troisième image traitée, a étant plus grand que b, c et d.
PCT/JP2024/004036 2023-03-13 2024-02-07 Procédé de traitement d'image Pending WO2024190191A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202480014661.1A CN120752518A (zh) 2023-03-13 2024-02-07 图像处理方法

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2023038536A JP2024129373A (ja) 2023-03-13 2023-03-13 画像処理方法
JP2023-038536 2023-03-13

Publications (1)

Publication Number Publication Date
WO2024190191A1 true WO2024190191A1 (fr) 2024-09-19

Family

ID=92755152

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2024/004036 Pending WO2024190191A1 (fr) 2023-03-13 2024-02-07 Procédé de traitement d'image

Country Status (4)

Country Link
JP (1) JP2024129373A (fr)
CN (1) CN120752518A (fr)
TW (1) TW202438876A (fr)
WO (1) WO2024190191A1 (fr)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007255959A (ja) * 2006-03-22 2007-10-04 Lasertec Corp 検査装置及び検査方法とその検査装置及び検査方法を用いたパターン基板の製造方法
JP2010121958A (ja) * 2008-11-17 2010-06-03 Seiko Epson Corp 欠陥検出方法および欠陥検出装置
WO2013038833A1 (fr) * 2011-09-16 2013-03-21 コニカミノルタホールディングス株式会社 Système de traitement d'image, procédé de traitement d'image et programme de traitement d'image

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007255959A (ja) * 2006-03-22 2007-10-04 Lasertec Corp 検査装置及び検査方法とその検査装置及び検査方法を用いたパターン基板の製造方法
JP2010121958A (ja) * 2008-11-17 2010-06-03 Seiko Epson Corp 欠陥検出方法および欠陥検出装置
WO2013038833A1 (fr) * 2011-09-16 2013-03-21 コニカミノルタホールディングス株式会社 Système de traitement d'image, procédé de traitement d'image et programme de traitement d'image

Also Published As

Publication number Publication date
CN120752518A (zh) 2025-10-03
TW202438876A (zh) 2024-10-01
JP2024129373A (ja) 2024-09-27

Similar Documents

Publication Publication Date Title
KR100744212B1 (ko) 인쇄회로기판상의 결함을 검출하기 위한 시스템 및 방법
JP4229767B2 (ja) 画像欠陥検査方法、画像欠陥検査装置及び外観検査装置
CN112669272B (zh) 一种aoi快速检测方法及快速检测系统
CN1331206C (zh) 缺陷检查方法
JP5088165B2 (ja) 欠陥検出方法および欠陥検出装置
JP5660861B2 (ja) 基板上の異物検査方法および異物検査装置
CN1898555B (zh) 基板检查装置
WO2024190191A1 (fr) Procédé de traitement d'image
JP5257063B2 (ja) 欠陥検出方法および欠陥検出装置
JP4629086B2 (ja) 画像欠陥検査方法および画像欠陥検査装置
JP5765713B2 (ja) 欠陥検査装置、欠陥検査方法、及び欠陥検査プログラム
CN107533994A (zh) 自动化的基于图像的过程监测及控制
JP4244046B2 (ja) 画像処理方法および画像処理装置
JP5239275B2 (ja) 欠陥検出方法および欠陥検出装置
JP2008292342A (ja) エッジ欠陥検出方法、そのプログラム、および検出装置
JP5096940B2 (ja) プリント配線板の検査方法及びその装置
JP2010243214A (ja) 欠陥検出方法および欠陥検出装置
JP2006329679A (ja) 製品パターン検査方法および装置
JP4889018B2 (ja) 外観検査方法
JPH11339040A (ja) マクロ検査方法
JP2009186330A (ja) 振動検出方法、外観検査方法、およびこれらの装置
JP2005258724A (ja) 製品検査装置
JP2007147320A (ja) 検査装置
WO2024166608A1 (fr) Procédé et dispositif d'inspection visuelle
TW202410230A (zh) 晶圓光學檢測方法及晶圓光學檢測裝置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 24770310

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: CN2024800146611

Country of ref document: CN

Ref document number: 202480014661.1

Country of ref document: CN

WWP Wipo information: published in national office

Ref document number: 202480014661.1

Country of ref document: CN

NENP Non-entry into the national phase

Ref country code: DE