CN117635565A - Semiconductor surface defect detection system based on image recognition - Google Patents
Semiconductor surface defect detection system based on image recognition Download PDFInfo
- Publication number
- CN117635565A CN117635565A CN202311620522.7A CN202311620522A CN117635565A CN 117635565 A CN117635565 A CN 117635565A CN 202311620522 A CN202311620522 A CN 202311620522A CN 117635565 A CN117635565 A CN 117635565A
- Authority
- CN
- China
- Prior art keywords
- image
- value
- flaw
- module
- semiconductor wafer
- 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
- 239000004065 semiconductor Substances 0.000 title claims abstract description 140
- 230000007547 defect Effects 0.000 title claims abstract description 83
- 238000001514 detection method Methods 0.000 title claims abstract description 71
- 238000011156 evaluation Methods 0.000 claims abstract description 46
- 238000007781 pre-processing Methods 0.000 claims abstract description 22
- 230000002950 deficient Effects 0.000 claims abstract description 10
- 235000012431 wafers Nutrition 0.000 claims description 156
- 238000004458 analytical method Methods 0.000 claims description 76
- 238000005286 illumination Methods 0.000 claims description 68
- 238000004519 manufacturing process Methods 0.000 claims description 49
- 238000004364 calculation method Methods 0.000 claims description 29
- 238000000034 method Methods 0.000 claims description 23
- 238000001228 spectrum Methods 0.000 claims description 22
- 230000008569 process Effects 0.000 claims description 17
- 238000012544 monitoring process Methods 0.000 claims description 12
- 230000002159 abnormal effect Effects 0.000 claims description 11
- 238000005065 mining Methods 0.000 claims description 11
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000005516 engineering process Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 238000005070 sampling Methods 0.000 claims description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 238000004891 communication Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 5
- 238000013136 deep learning model Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims 1
- XUIMIQQOPSSXEZ-UHFFFAOYSA-N Silicon Chemical group [Si] XUIMIQQOPSSXEZ-UHFFFAOYSA-N 0.000 description 7
- 229910052710 silicon Inorganic materials 0.000 description 7
- 239000010703 silicon Substances 0.000 description 7
- 230000002411 adverse Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 239000013078 crystal Substances 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000003749 cleanliness Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 229910021421 monocrystalline silicon Inorganic materials 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 229910021420 polycrystalline silicon Inorganic materials 0.000 description 1
- -1 scratches Substances 0.000 description 1
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
- G01N21/9501—Semiconductor wafers
-
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/09—Supervised learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- H01L21/67005—Apparatus not specifically provided for elsewhere
- H01L21/67242—Apparatus for monitoring, sorting or marking
- H01L21/67271—Sorting devices
-
- 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
- H01L21/67005—Apparatus not specifically provided for elsewhere
- H01L21/67242—Apparatus for monitoring, sorting or marking
- H01L21/67282—Marking devices
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L22/00—Testing or measuring during manufacture or treatment; Reliability measurements, i.e. testing of parts without further processing to modify the parts as such; Structural arrangements therefor
- H01L22/10—Measuring as part of the manufacturing process
- H01L22/12—Measuring as part of the manufacturing process for structural parameters, e.g. thickness, line width, refractive index, temperature, warp, bond strength, defects, optical inspection, electrical measurement of structural dimensions, metallurgic measurement of diffusions
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L22/00—Testing or measuring during manufacture or treatment; Reliability measurements, i.e. testing of parts without further processing to modify the parts as such; Structural arrangements therefor
- H01L22/20—Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30148—Semiconductor; IC; Wafer
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Manufacturing & Machinery (AREA)
- Software Systems (AREA)
- Power Engineering (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Pathology (AREA)
- Biochemistry (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Immunology (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Geometry (AREA)
- Signal Processing (AREA)
- Quality & Reliability (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
Abstract
The invention belongs to the technical field of semiconductor detection, in particular to a semiconductor surface defect detection system based on image recognition, which comprises a server, an image scanning acquisition module, an image preprocessing module, an image positioning registration module, a pixel defect detection recognition module and a surface defect comprehensive evaluation module; according to the invention, the collected target image is preprocessed through the image preprocessing module, the image positioning registration module carries out positioning registration on the target image and the standard template image, the pixel point detection and identification module accurately identifies defective pixel in the target image, the surface defect condition of the semiconductor wafer is accurately analyzed according to the defective pixel, the quality grade of the semiconductor wafer is accurately judged, the factors influencing the image quality are comprehensively detected through the image acquisition condition detection and early warning module before the image scanning acquisition, the influence degree of the current image acquisition condition on the image quality is judged, and the image acquisition quality is effectively ensured.
Description
Technical Field
The invention relates to the technical field of semiconductor detection, in particular to a semiconductor surface defect detection system based on image recognition.
Background
The semiconductor wafer is a silicon wafer used for manufacturing a silicon semiconductor circuit, the original material is silicon, high-purity polycrystalline silicon is dissolved and then is doped with silicon crystal seeds, then the silicon crystal seeds are slowly pulled out to form cylindrical monocrystalline silicon, a silicon crystal rod is ground, polished and sliced to form the silicon wafer, namely the wafer, and the main processing modes of the wafer are wafer processing and batch processing; in the production and processing process of a semiconductor wafer, defect detection is an essential link to the surface of the semiconductor wafer;
at present, when surface defect detection of a semiconductor wafer is carried out, defective point pixels are difficult to quickly and accurately identify through an image identification technology, grade accurate judgment of the semiconductor wafer is realized, the adverse influence degree of related factors on the surface image of the semiconductor wafer cannot be comprehensively and reasonably estimated before the surface image of the semiconductor wafer is acquired, the acquired image quality is difficult to effectively ensure, and the accuracy of grade judgment results is not facilitated;
in view of the above technical drawbacks, a solution is now proposed.
Disclosure of Invention
The invention aims to provide a semiconductor surface defect detection system based on image recognition, which solves the problems that in the prior art, defective point pixels are difficult to rapidly and accurately recognize through an image recognition technology, the grade of a semiconductor wafer is accurately judged, the adverse influence degree of relevant factors on the semiconductor wafer cannot be comprehensively and reasonably estimated before the surface image of the semiconductor wafer is acquired, and the grade judgment result accuracy is not facilitated to be improved.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a semiconductor surface defect detection system based on image recognition comprises a server, an image scanning acquisition module, an image preprocessing module, an image positioning registration module, a pixel defect detection recognition module and a surface defect comprehensive evaluation module; the image scanning acquisition module acquires a surface image of the semiconductor wafer through a scanning electron microscope or an atomic force microscope, marks the acquired surface image as a target image, and sends the target image to the image preprocessing module through the server; the image preprocessing module is used for preprocessing the received target image by utilizing an image preprocessing technology, and the preprocessed target image is sent to the image positioning registration module through the server;
the image positioning registration module is used for calling a standard template image of the semiconductor wafer from the server, performing feature point matching on the target image and the standard template image through a feature matching technology, and performing geometric transformation on the images based on corresponding feature points to align the target image and the standard template image of the semiconductor wafer, so that positioning registration of the target image and the standard template image of the semiconductor wafer is realized; the pixel point detection and identification module performs pixel flaw identification analysis on the surface of the semiconductor wafer based on the target image and the standard template image after positioning registration so as to capture flaw point pixels in the target image, and sends the captured flaw point pixels to the surface flaw comprehensive evaluation module through the server; the surface defect comprehensive evaluation module is used for performing grade judgment on the semiconductor wafer through analysis, marking the semiconductor wafer as a top grade wafer, a good grade wafer or a obsolete wafer, and sending grade judgment information of the corresponding semiconductor wafer to the server.
Further, the specific analysis process of the pixel defect identification analysis is as follows:
respectively calculating gray-scale histograms of the registered target image and the standard template image, finding out the maximum value in each histogram, marking the two maximum values as a target maximum value and a standard maximum value, calculating the difference value between the target maximum value and the standard maximum value and marking the difference value as a direct difference value to obtain the difference of the target image and the standard template image in brightness;
adding the above-mentioned straight variance difference value to each pixel in the effective area of the target image, comparing the obtained result with the gray scale value of the corresponding position of the standard template image, if the difference between the two is greater than the set corresponding threshold value, the corresponding pixel is marked as a defective pixel; and the defective point pixels in the target image are sent to a surface defect comprehensive evaluation module through a server.
Further, the specific operation process of the surface defect comprehensive evaluation module comprises the following steps:
based on all the flaw point pixels, a connected region marking algorithm in image processing is used for finding out the connected region of each flaw, determining the size and shape of each flaw, and calculating the area size of the connected region of each flaw; marking and classifying the connected areas of each flaw to obtain a picture of each flaw, and taking the picture of each flaw as input data of a deep learning model;
the flaw pictures are sequentially input into a convolutional neural network classifier trained in advance, the category of each flaw is identified through the convolutional neural network classifier, and grade judgment is carried out according to the identified flaw category and the number, so that the corresponding semiconductor wafer is marked as a top grade wafer, a good grade wafer or a obsolete wafer.
Further, the specific analysis and judgment process of the grade judgment is as follows:
retrieving all defect categories of the semiconductor wafer from the server, marking the corresponding defect categories as i, i= {1,2, …, n }, wherein n represents the number of the defect categories and n is a natural number greater than 1; the number of the corresponding flaw class i in the target image is collected and marked as a flaw distribution value, the flaw distribution value of the flaw class i is compared with a corresponding preset flaw distribution threshold value in a numerical mode, and if the flaw distribution value exceeds the preset flaw distribution value, the corresponding flaw class i in the target image is judged to be abnormal;
if the defect type with abnormal performance exists in the semiconductor wafer, marking the corresponding semiconductor wafer as a obsolete wafer; if no defect type with abnormal appearance exists in the semiconductor wafer, the defect influence value of each defect type is called from the server, the numerical value of the defect influence value is larger than zero, and the defect influence value of each defect type is recorded in advance by a manager and stored in the server;
multiplying the flaw distribution value of the corresponding flaw class i in the target image with the corresponding flaw influence value, marking the product of the flaw distribution value and the flaw influence value as flaw detection values, and carrying out summation calculation on flaw detection values of all flaw classes in the target image to obtain flaw evaluation values; comparing the flaw evaluation value with a preset flaw evaluation value range in a numerical value mode, and marking the corresponding semiconductor wafer as a obsolete wafer if the flaw evaluation value exceeds the maximum value of the preset flaw evaluation value range; if the flaw evaluation value is within the preset flaw evaluation value range, marking the corresponding semiconductor wafer as a good-grade wafer; and if the flaw evaluation value does not exceed the minimum value of the preset flaw evaluation value range, marking the corresponding semiconductor wafer as a top-grade wafer.
Further, the server is in communication connection with the image condition detection early warning module, before the image scanning acquisition module performs image scanning acquisition of the semiconductor wafer, the image condition detection early warning module detects and analyzes factors influencing image quality, judges the influence degree of the current image acquisition condition on the image quality, generates a condition detection unqualified signal or a condition detection qualified signal, sends the condition detection unqualified signal or the condition detection qualified signal to the server, and sends corresponding early warning when the condition detection unqualified signal is generated.
Further, the specific operation process of the image sampling condition detection and early warning module comprises the following steps:
comparing the position deviation value and the angle deviation value of the acquired semiconductor wafer with those of the image scanning acquisition module before image acquisition, carrying out numerical calculation on the position deviation value and the angle deviation value to obtain a potential angle analysis value, carrying out numerical comparison on the potential angle analysis value and a preset potential angle analysis threshold value, and generating a sampling condition detection failure signal if the potential angle analysis value exceeds the preset potential angle analysis threshold value;
if the potential angle analysis value does not exceed the preset potential angle analysis threshold value, acquiring an illumination analysis value through light source performance analysis, comparing the illumination analysis value with the preset illumination analysis threshold value in a numerical mode, and if the illumination analysis value exceeds the preset illumination analysis threshold value, generating a sampling condition detection failure signal; and if the illumination analysis value does not exceed the preset illumination analysis threshold value, carrying out loop acquisition auxiliary decision analysis.
Further, the specific analysis process of the loop-picking auxiliary detection analysis is as follows:
acquiring the region temperature, the region humidity and the region ambiguity of an image acquisition region, calling a preset suitable image acquisition temperature range from a server, carrying out average value calculation on the maximum value and the minimum value of the preset suitable image acquisition temperature range to obtain an image acquisition temperature standard value, carrying out difference value calculation on the region temperature and the image acquisition temperature standard value, and taking an absolute value to obtain an image acquisition temperature difference value, and acquiring the image acquisition humidity difference value in the same way; carrying out numerical calculation on a picked-up image temperature difference value, a picked-up image humidity difference value and an area ambiguity of an image acquisition area to obtain a picked-up loop auxiliary value, carrying out numerical comparison on the picked-up loop auxiliary value and a preset picked-up loop auxiliary threshold value, and generating a picked-up condition detection unqualified signal if the picked-up loop auxiliary value exceeds the preset picked-up loop auxiliary threshold value; and if the mining ring auxiliary value does not exceed the preset mining ring auxiliary threshold value, generating a mining condition detection qualified signal.
Further, the specific analysis process of the light source performance analysis is as follows:
acquiring a light source spectrum actual range of an image acquisition area, calling a preset light source spectrum proper range from a server, and performing superposition comparison on the preset light source spectrum proper range and the light source spectrum actual range to obtain a spectrum superposition coefficient; a plurality of illumination monitoring points are distributed in an image acquisition area, illumination intensity values corresponding to the illumination monitoring points are acquired, illumination intensity values of all the illumination monitoring points are summed up and calculated, an average value is obtained, and variance calculation is carried out on the illumination intensity values of all the illumination monitoring points to obtain an illumination non-uniformity coefficient;
the method comprises the steps of calling a preset proper image capturing illumination intensity range from a server, carrying out numerical calculation on the maximum value and the minimum value of the preset proper image capturing illumination intensity range to obtain an image capturing cursor value, carrying out difference calculation on an illumination average value and the image capturing cursor value, and taking an absolute value to obtain an illumination deviation coefficient; and carrying out numerical calculation on the spectrum coincidence degree coefficient, the illumination deviation coefficient and the illumination non-uniformity coefficient to obtain an illumination analysis value.
Further, the server is in communication connection with the semiconductor production line production early warning module, the server sends grade judging information of all semiconductor wafers produced in unit time of the corresponding semiconductor production line to the semiconductor production line production early warning module, the semiconductor production line gathers the grade judging information of all semiconductor wafers, the number of the superior wafers, the number of the good wafers and the number of the obsolete wafers produced in the unit time of the corresponding semiconductor production line are obtained, and numerical calculation is carried out on the number of the superior wafers, the number of the good wafers and the number of the obsolete wafers to obtain a production line early warning coefficient; and comparing the production line early warning coefficient with a preset production line early warning coefficient threshold value in a numerical mode, if the production line early warning coefficient exceeds the preset production line early warning coefficient threshold value, generating a production line early warning signal, sending the signal to a server, and sending a corresponding early warning when the production line early warning signal is generated.
Compared with the prior art, the invention has the beneficial effects that:
1. according to the invention, the image scanning acquisition module acquires the surface image of the semiconductor wafer, the image preprocessing module carries out preprocessing on the received target image to ensure the quality of the image, the image positioning registration module carries out positioning registration on the target image and the standard template image, the pixel point detection identification module carries out pixel defect identification analysis on the surface of the semiconductor wafer based on the positioned and registered target image and the standard template image, the defect point pixel in the target image can be accurately identified, the surface defect comprehensive evaluation module carries out grade judgment on the semiconductor wafer through analysis, the accurate analysis on the surface defect condition of the semiconductor wafer is realized, the quality grade of the semiconductor wafer can be accurately judged, and corresponding treatment measures can be carried out on the semiconductor wafers with different grades;
2. according to the invention, before the image scanning and acquisition module performs image scanning and acquisition of the semiconductor wafer, the image acquisition condition detection and early warning module detects and analyzes factors influencing the image quality, and judges the influence degree of the current image acquisition condition on the image quality, so that an operator can timely make targeted improvement measures, the image acquisition quality is guaranteed, and the accuracy of the subsequent semiconductor wafer grade judgment result is improved; and summarizing grade judging information of the semiconductor wafers produced in unit time of the semiconductor production line through the semiconductor production line production early warning module, and judging abnormal production conditions of the corresponding production line, so that equipment supervision and operator supervision of the corresponding production line are timely enhanced by a production line manager, and the product quality of the semiconductor wafers produced subsequently is ensured.
Drawings
For the convenience of those skilled in the art, the present invention will be further described with reference to the accompanying drawings;
FIG. 1 is a first system block diagram of a first embodiment of the present invention;
FIG. 2 is a second system block diagram according to a first embodiment of the invention;
fig. 3 is a system block diagram of a second embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Embodiment one: as shown in fig. 1-2, the semiconductor surface defect detection system based on image recognition provided by the invention comprises a server, an image scanning acquisition module, an image preprocessing module, an image positioning registration module, a pixel defect detection recognition module and a surface defect comprehensive evaluation module, wherein the server is in communication connection with the image scanning acquisition module, the image preprocessing module, the image positioning registration module, the pixel defect detection recognition module and the surface defect comprehensive evaluation module; the image scanning acquisition module acquires a surface image of the semiconductor wafer by using high-precision image acquisition equipment (such as a scanning electron microscope or an atomic force microscope), marks the acquired surface image as a target image, and sends the target image to the image preprocessing module through the server;
the image preprocessing module preprocesses the received target image by utilizing an image preprocessing technology to ensure the quality of the image, and sends the preprocessed target image to the image positioning registration module through the server; it should be noted that, the preprocessing operation mainly includes gray conversion, noise removal, contrast enhancement, and normalization; wherein, gray scale conversion: converting the original color image into a gray scale image, which can reduce the data throughput; noise removal: filters (e.g., gaussian filters) are used to reduce noise in the image that may be generated due to environmental or device factors during image acquisition; contrast enhancement: increasing the contrast of the image by stretching the pixel value range may make the details of the image more apparent; normalization: the pixel value range of the image is scaled to a uniform range.
Because the manufacturing process of the semiconductor wafer is complex, there may be some geometrical distortion and deviation caused by other factors, so that the collected target image and the standard template image need to be aligned; the image positioning registration module is used for calling a standard template image of the semiconductor wafer from the server, performing feature point matching on the target image and the standard template image through a feature matching technology in image processing, for example, performing feature point matching by using SIFT, SURF and other algorithms, and performing geometric transformation on the image based on the corresponding feature points to align the target image and the standard template image of the semiconductor wafer, so that positioning registration of the target image and the standard template image of the semiconductor wafer is realized;
it should be noted that SIFT (Scale-Invariant Feature Transform) refers to Scale-invariant feature transform, which is a method for detecting and describing local features in an image; the SIFT algorithm has the main advantages that the SIFT algorithm has invariance to scale and rotation, has good uniqueness, can be used for accurate matching, can be used for finding similar characteristic points between a wafer image and a standard template image in wafer image registration, and realizes geometric transformation alignment of two images by matching the characteristic points;
SURF (Speeded Up Robust Features) is an acceleration robust feature, which is a method of detecting and describing local features in an image, with high uniqueness and stability; compared with the SIFT algorithm, the SURF algorithm has faster calculation speed, so that the SURF algorithm has more advantages in some application scenes with higher real-time requirements; in the wafer image registration, the SURF algorithm may also be used to find similar feature points between the wafer image and the standard template image, and by matching these feature points, the geometric transformation alignment of the two images is achieved.
The pixel point detection and identification module performs pixel flaw identification analysis on the surface of the semiconductor wafer based on the target image and the standard template image after positioning registration to capture flaw point pixels in the target image, and sends the captured flaw point pixels to the surface flaw comprehensive evaluation module through the server, so that flaw point pixels in the target image can be accurately identified, and data support is provided for the follow-up accurate evaluation of the surface flaw grade condition of the semiconductor wafer; the specific analysis process of the pixel defect identification analysis is as follows:
respectively calculating gray-scale histograms of the registered target image and the standard template image, finding out the maximum value in each histogram, wherein the maximum value can be regarded as the brightness level of the image, marking the two maximum values as the target maximum value and the standard maximum value, calculating the difference value between the target maximum value and the standard maximum value and marking the difference value as a direct variance difference value, and obtaining the difference of the brightness of the target image and the standard template image;
adding the above-mentioned straight variance difference value to each pixel in the effective area of the target image, comparing the obtained result with the gray scale value of the corresponding position of the standard template image, if the difference between the two is greater than the set corresponding threshold value, the corresponding pixel is marked as a defective pixel; and the defective point pixels in the target image are sent to a surface defect comprehensive evaluation module through a server.
The surface defect comprehensive evaluation module is used for judging the grade of the semiconductor wafer through analysis, marking the semiconductor wafer as a superior grade wafer, a good grade wafer or an obsolete wafer, and sending grade judging information corresponding to the semiconductor wafer to the server, so that the accurate analysis of the surface defect condition of the semiconductor wafer is realized, the quality grade of the semiconductor wafer can be accurately judged, and corresponding treatment measures can be made for the semiconductor wafers with different grades; the specific operation process of the surface defect comprehensive evaluation module is as follows:
based on all the flaw point pixels, a connected region marking algorithm (such as connected region analysis of binary images) in image processing is used for finding out the connected region of each flaw, and the size and the shape of each flaw can be determined and the area size of the connected region of each flaw can be calculated; marking and classifying the connected areas of each flaw to obtain a picture of each flaw, and taking the picture of each flaw as input data of a deep learning model; the flaw pictures are sequentially input into a convolutional neural network Classifier (CNN) trained in advance, the category (such as particles, scratches, holes and the like) of each flaw is identified through the convolutional neural network classifier, grade judgment is carried out according to the identified flaw category and the number, and the specific analysis and judgment process of the grade judgment is as follows:
retrieving all defect categories of the semiconductor wafer from the server, marking the corresponding defect categories as i, i= {1,2, …, n }, wherein n represents the number of the defect categories and n is a natural number greater than 1; the number of the corresponding flaw class i in the target image is collected and marked as a flaw distribution value, the flaw distribution value of the flaw class i is compared with a corresponding preset flaw distribution threshold value in a numerical mode, if the flaw distribution value exceeds the preset flaw distribution value of the corresponding flaw class i, the fact that the corresponding flaw class i is abnormal is indicated, and the fact that the corresponding flaw class i in the target image is abnormal is judged;
if the defect types with abnormal performance exist in the semiconductor wafer, which indicates that the surface defect condition of the corresponding semiconductor wafer is serious, marking the corresponding semiconductor wafer as a obsolete wafer; if no defect type with abnormal appearance exists in the semiconductor wafer, the defect influence value of each defect type is called from the server, the numerical value of the defect influence value is larger than zero, and the defect influence value of each defect type is recorded in advance by a manager and stored in the server; it should be noted that, the larger the value of the flaw influence value is, the larger the adverse influence of the corresponding flaw class i on the quality of the semiconductor wafer is; multiplying the flaw distribution value of the corresponding flaw class i in the target image with the corresponding flaw influence value, marking the product of the flaw distribution value and the flaw influence value as flaw detection values, and carrying out summation calculation on flaw detection values of all flaw classes in the target image to obtain flaw evaluation values;
it should be noted that, the larger the value of the flaw evaluation value is, the more serious the surface defect condition of the corresponding semiconductor wafer is; comparing the flaw evaluation value with a preset flaw evaluation value range in a numerical value mode, and marking the corresponding semiconductor wafer as a obsolete wafer if the flaw evaluation value exceeds the maximum value of the preset flaw evaluation value range and indicates that the surface defect condition of the corresponding semiconductor wafer is serious; if the flaw evaluation value is within the preset flaw evaluation value range, indicating that the surface defect condition of the corresponding semiconductor wafer is good, marking the corresponding semiconductor wafer as a good-grade wafer; if the flaw evaluation value does not exceed the minimum value of the preset flaw evaluation value range, indicating that the surface defect condition of the corresponding semiconductor wafer is good, marking the corresponding semiconductor wafer as a top-grade wafer.
Further, the server is in communication connection with the image condition detection and early warning module, before the image scanning and acquisition module performs image scanning and acquisition of the semiconductor wafer, the image condition detection and early warning module detects and analyzes factors influencing image quality, judges the influence degree of the current image acquisition condition on the image quality, generates a condition detection unqualified signal or a condition detection qualified signal, sends the condition detection unqualified signal or the condition detection qualified signal to the server, and sends corresponding early warning when the condition detection unqualified signal is generated, so that operators can timely make targeted improvement measures, and the image acquisition quality is guaranteed; the specific operation process of the image sampling condition detection and early warning module is as follows:
comparing the position deviation value and the angle deviation value of the collected semiconductor wafer with those of an image scanning and collecting module before image collection, and carrying out numerical calculation on the position deviation value WP and the angle deviation value JP through a formula WJ=ep1×WP+ep2×JP to obtain a potential angle analysis value WJ, wherein ep1 and ep2 are preset weight coefficients, and the values of ep1 and ep2 are both larger than zero; and, the larger the value of the potential angle analysis value WJ is, the more unfavorable the current is for guaranteeing the image acquisition quality; comparing the azimuth analysis value WJ with a preset azimuth analysis threshold value, and generating a sampling condition detection disqualification signal if the azimuth analysis value WJ exceeds the preset azimuth analysis threshold value;
if the potential angle analysis value WJ does not exceed the preset potential angle analysis threshold value, acquiring a light source spectrum actual range of the image acquisition area, retrieving a preset light source spectrum proper range from a server, and performing superposition comparison on the preset light source spectrum proper range and the light source spectrum actual range to obtain a spectrum superposition coefficient (for example, if the preset light source spectrum proper range is completely in the light source spectrum actual range, the spectrum superposition coefficient is 1), wherein the larger the value of the spectrum superposition coefficient is, the more favorable the quality of the acquired image is improved; a plurality of illumination monitoring points are distributed in an image acquisition area, illumination intensity values corresponding to the illumination monitoring points are acquired, illumination intensity values of all the illumination monitoring points are summed up and calculated, an average value is obtained, and variance calculation is carried out on the illumination intensity values of all the illumination monitoring points to obtain an illumination non-uniformity coefficient;
the method comprises the steps of calling a preset proper image capturing illumination intensity range from a server, carrying out numerical calculation on the maximum value and the minimum value of the preset proper image capturing illumination intensity range to obtain an image capturing cursor value, carrying out difference calculation on an illumination average value and the image capturing cursor value, and taking an absolute value to obtain an illumination deviation coefficient; it should be noted that, the smaller the value of the illumination non-uniformity coefficient is, the more uniform the illumination of the image acquisition area is, which is more beneficial to improving the quality of the acquired image; the smaller the value of the illumination deviation coefficient is, the more the illumination intensity of the image acquisition area tends to be at a proper level, and the quality of the acquired image is improved;
calculating the spectrum overlap ratio coefficient GY, the illumination deviation coefficient GR and the illumination non-uniformity coefficient GF by the formula GT=t1/(GY+0.126) +t2+t3, wherein tf1, tf2 and tf3 are preset proportional coefficients, and tf1 > tf2 > tf3 > 1; in addition, the smaller the value of the illumination analysis value GT is, the better the illumination performance of the image acquisition area is, and the quality of the acquired image is improved; comparing the illumination analysis value GT with a preset illumination analysis threshold value, and generating a sampling condition detection failure signal if the illumination analysis value GT exceeds the preset illumination analysis threshold value;
if the illumination analysis value GT does not exceed a preset illumination analysis threshold value, acquiring the region temperature, the region humidity and the region ambiguity of the image acquisition region, wherein the region ambiguity is a data value representing the region atmosphere cleanliness degree, and the greater the dust concentration of the image acquisition region is, the worse the region atmosphere cleanliness is and the greater the value of the region ambiguity is; the method comprises the steps of retrieving a preset suitable image-picking temperature range from a server, carrying out average value calculation on the maximum value and the minimum value of the preset suitable image-picking temperature range to obtain an image-picking temperature standard value, carrying out difference value calculation on the regional temperature and the image-picking temperature standard value, and taking an absolute value to obtain an image-picking temperature difference value, and obtaining an image-picking humidity difference value in the same way;
calculating the values of a picked-up temperature difference value QT, a picked-up humidity difference value QY and a region ambiguity QK of an image acquisition region through a formula CT=ed 1 qt+ed2 QY+ed3 QK to obtain a picked-up loop auxiliary value CT, wherein ed1, ed2 and ed3 are preset weight coefficients, and the values of ed1, ed2 and ed3 are all larger than zero; moreover, the larger the numerical value of the acquisition loop auxiliary value CT is, the more unfavorable the quality of the acquired image is ensured; comparing the mining ring auxiliary value CT with a preset mining ring auxiliary threshold value, and generating a mining condition detection failure signal if the mining ring auxiliary value CT exceeds the preset mining ring auxiliary threshold value; if the acquisition loop auxiliary value CT does not exceed the preset acquisition loop auxiliary threshold value, generating an acquisition condition detection qualified signal.
Embodiment two: as shown in fig. 3, the difference between the present embodiment and embodiment 1 is that the server is communicatively connected to the semiconductor production line production pre-warning module, and the server sends the grade determination information of all the semiconductor wafers produced in the unit time of the corresponding semiconductor production line to the semiconductor production line production pre-warning module, and the semiconductor production line gathers the grade determination information of all the semiconductor wafers to obtain the number of the top-grade wafers, the number of the good-grade wafers and the number of the eliminated wafers produced in the unit time of the corresponding semiconductor production line;
performing numerical calculation on the number YK of the superior wafers, the number YT of the good wafers and the number YG of the obsolete wafers by a formula CY= (a1+a2+YT+a3)/(YK+YT+YG) to obtain a line early warning coefficient CY; wherein a1, a2 and a3 are preset weight coefficients, a3 > a2 > a1 > 0; and the larger the value of the product line early warning coefficient CY is, the worse the product quality of the semiconductor wafer produced by the corresponding product line in unit time is, and the higher the possibility of abnormality of the corresponding product line is; and comparing the production line early-warning coefficient CY with a preset production line early-warning coefficient threshold value, if the production line early-warning coefficient CY exceeds the preset production line early-warning coefficient threshold value, generating a production line early-warning signal and sending the signal to a server, and sending corresponding early warning when generating the production line early-warning signal, so that a production line manager can master the quality of semiconductor wafers produced by the corresponding production line in detail, thereby enhancing the equipment supervision and the operator supervision of the corresponding production line in time, eliminating the existing production line abnormality, and ensuring the product quality of the semiconductor wafers produced subsequently.
The working principle of the invention is as follows: when the image processing device is used, the image scanning acquisition module acquires the surface image of the semiconductor wafer by using high-precision image acquisition equipment, the image preprocessing module preprocesses the received target image by using an image preprocessing technology so as to ensure the quality of the image, and the image positioning registration module carries out positioning registration on the target image and a standard template image; the pixel point detection and identification module performs pixel flaw identification analysis on the surface of the semiconductor wafer based on the target image and the standard template image after positioning registration so as to capture flaw point pixels in the target image, accurately identify flaw point pixels in the target image and provide data support for the follow-up accurate assessment of the surface flaw grade condition of the semiconductor wafer; the surface defect comprehensive evaluation module is used for carrying out grade judgment on the semiconductor wafer through analysis, so that accurate analysis on the surface defect condition of the semiconductor wafer is realized, and the quality grade of the semiconductor wafer can be accurately judged, thereby being beneficial to carrying out corresponding treatment measures on the semiconductor wafers with different grades; and before the image scanning and acquisition module performs image scanning and acquisition of the semiconductor wafer, the image acquisition condition detection and early warning module detects and analyzes factors influencing the image quality, and judges the influence degree of the current image acquisition condition on the image quality, so that an operator can timely make targeted improvement measures, the image acquisition quality is guaranteed, and the accuracy of the subsequent semiconductor wafer grade judgment result is improved.
The above formulas are all formulas with dimensions removed and numerical values calculated, the formulas are formulas with a large amount of data collected for software simulation to obtain the latest real situation, and preset parameters in the formulas are set by those skilled in the art according to the actual situation. The preferred embodiments of the invention disclosed above are intended only to assist in the explanation of the invention. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention. The invention is limited only by the claims and the full scope and equivalents thereof.
Claims (9)
1. The semiconductor surface defect detection system based on image recognition is characterized by comprising a server, an image scanning acquisition module, an image preprocessing module, an image positioning registration module, a pixel defect detection recognition module and a surface defect comprehensive evaluation module; the image scanning acquisition module acquires a surface image of the semiconductor wafer through a scanning electron microscope or an atomic force microscope, marks the acquired surface image as a target image, and sends the target image to the image preprocessing module through the server; the image preprocessing module is used for preprocessing the received target image by utilizing an image preprocessing technology, and the preprocessed target image is sent to the image positioning registration module through the server;
the image positioning registration module is used for calling a standard template image of the semiconductor wafer from the server, performing feature point matching on the target image and the standard template image through a feature matching technology, and performing geometric transformation on the images based on corresponding feature points to align the target image and the standard template image of the semiconductor wafer, so that positioning registration of the target image and the standard template image of the semiconductor wafer is realized; the pixel point detection and identification module performs pixel flaw identification analysis on the surface of the semiconductor wafer based on the target image and the standard template image after positioning registration so as to capture flaw point pixels in the target image, and sends the captured flaw point pixels to the surface flaw comprehensive evaluation module through the server; the surface defect comprehensive evaluation module is used for performing grade judgment on the semiconductor wafer through analysis, marking the semiconductor wafer as a top grade wafer, a good grade wafer or a obsolete wafer, and sending grade judgment information of the corresponding semiconductor wafer to the server.
2. The system for detecting surface defects of a semiconductor based on image recognition according to claim 1, wherein the specific analysis process of the pixel defect recognition analysis is as follows:
respectively calculating gray-scale histograms of the registered target image and the standard template image, finding out the maximum value in each histogram, marking the two maximum values as a target maximum value and a standard maximum value, calculating the difference value between the target maximum value and the standard maximum value and marking the difference value as a direct difference value to obtain the difference of the target image and the standard template image in brightness;
adding the above-mentioned straight variance difference value to each pixel in the effective area of the target image, comparing the obtained result with the gray scale value of the corresponding position of the standard template image, if the difference between the two is greater than the set corresponding threshold value, the corresponding pixel is marked as a defective pixel; and the defective point pixels in the target image are sent to a surface defect comprehensive evaluation module through a server.
3. The system for detecting surface defects of a semiconductor based on image recognition according to claim 2, wherein the specific operation process of the integrated surface defect evaluation module comprises:
based on all the flaw point pixels, a connected region marking algorithm in image processing is used for finding out the connected region of each flaw, determining the size and shape of each flaw, and calculating the area size of the connected region of each flaw; marking and classifying the connected areas of each flaw to obtain a picture of each flaw, and taking the picture of each flaw as input data of a deep learning model;
the flaw pictures are sequentially input into a convolutional neural network classifier trained in advance, the category of each flaw is identified through the convolutional neural network classifier, and grade judgment is carried out according to the identified flaw category and the number, so that the corresponding semiconductor wafer is marked as a top grade wafer, a good grade wafer or a obsolete wafer.
4. A semiconductor surface defect inspection system based on image recognition according to claim 3, wherein the specific analysis determination process of the grade determination is as follows:
retrieving all defect categories of the semiconductor wafer from the server, marking the corresponding defect categories as i, i= {1,2, …, n }, wherein n represents the number of the defect categories and n is a natural number greater than 1; the number of the corresponding flaw class i in the target image is collected and marked as a flaw distribution value, the flaw distribution value of the flaw class i is compared with a corresponding preset flaw distribution threshold value in a numerical mode, and if the flaw distribution value exceeds the preset flaw distribution threshold value, the corresponding flaw class i in the target image is judged to be abnormal;
if the defect type with abnormal performance exists in the semiconductor wafer, marking the corresponding semiconductor wafer as a obsolete wafer; if no defect type with abnormal appearance exists in the semiconductor wafer, the defect influence value of each defect type is called from the server, the numerical value of the defect influence value is larger than zero, and the defect influence value of each defect type is recorded in advance by a manager and stored in the server;
multiplying the flaw distribution value of the corresponding flaw class i in the target image with the corresponding flaw influence value, marking the product of the flaw distribution value and the flaw influence value as flaw detection values, and carrying out summation calculation on flaw detection values of all flaw classes in the target image to obtain flaw evaluation values; if the flaw evaluation value exceeds the maximum value of the preset flaw evaluation value range, marking the corresponding semiconductor wafer as a obsolete wafer; if the flaw evaluation value is within the preset flaw evaluation value range, marking the corresponding semiconductor wafer as a good-grade wafer; and if the flaw evaluation value does not exceed the minimum value of the preset flaw evaluation value range, marking the corresponding semiconductor wafer as a top-grade wafer.
5. The semiconductor surface defect detection system based on image recognition according to claim 1, wherein the server is in communication connection with the image condition detection and early warning module, the image condition detection and early warning module detects and analyzes factors affecting image quality before the image scanning and acquisition module performs image scanning and acquisition of the semiconductor wafer, judges the influence degree of the current image acquisition condition on the image quality, generates a condition detection unqualified signal or a condition detection qualified signal, sends the condition detection unqualified signal or the condition detection qualified signal to the server, and sends corresponding early warning when the condition detection unqualified signal is generated.
6. The system for detecting surface defects of a semiconductor based on image recognition according to claim 5, wherein the specific operation process of the image condition detection and early warning module comprises:
comparing the position deviation value and the angle deviation value of the acquired semiconductor wafer with those of the image scanning acquisition module before image acquisition, carrying out numerical calculation on the position deviation value and the angle deviation value to obtain a potential angle analysis value, and generating a sampling condition detection failure signal if the potential angle analysis value exceeds a preset potential angle analysis threshold; if the potential angle analysis value does not exceed the preset potential angle analysis threshold, acquiring an illumination analysis value through light source performance analysis, and if the illumination analysis value exceeds the preset illumination analysis threshold, generating a sampling condition detection failure signal; and if the illumination analysis value does not exceed the preset illumination analysis threshold value, carrying out loop acquisition auxiliary decision analysis.
7. The system for detecting surface defects of a semiconductor based on image recognition according to claim 6, wherein the specific analysis process of the loop-picking auxiliary detection analysis is as follows:
acquiring the region temperature, the region humidity and the region ambiguity of an image acquisition region, calling a preset suitable image acquisition temperature range from a server, carrying out average value calculation on the maximum value and the minimum value of the preset suitable image acquisition temperature range to obtain an image acquisition temperature standard value, carrying out difference value calculation on the region temperature and the image acquisition temperature standard value, and taking an absolute value to obtain an image acquisition temperature difference value, and acquiring the image acquisition humidity difference value in the same way; carrying out numerical calculation on a picked-up image temperature difference value, a picked-up image humidity difference value and an area ambiguity of an image acquisition area to obtain a picked-up loop auxiliary value, and generating a picked-up condition detection unqualified signal if the picked-up loop auxiliary value exceeds a preset picked-up loop auxiliary threshold; and if the mining ring auxiliary value does not exceed the preset mining ring auxiliary threshold value, generating a mining condition detection qualified signal.
8. The system for detecting surface defects of a semiconductor based on image recognition as recited in claim 6, wherein the specific analysis process of the light source performance analysis is as follows:
acquiring a light source spectrum actual range of an image acquisition area, calling a preset light source spectrum proper range from a server, and performing superposition comparison on the preset light source spectrum proper range and the light source spectrum actual range to obtain a spectrum superposition coefficient; a plurality of illumination monitoring points are distributed in an image acquisition area, illumination intensity values corresponding to the illumination monitoring points are acquired, illumination intensity values of all the illumination monitoring points are summed up and calculated, an average value is obtained, and variance calculation is carried out on the illumination intensity values of all the illumination monitoring points to obtain an illumination non-uniformity coefficient;
the method comprises the steps of calling a preset proper image capturing illumination intensity range from a server, carrying out numerical calculation on the maximum value and the minimum value of the preset proper image capturing illumination intensity range to obtain an image capturing cursor value, carrying out difference calculation on an illumination average value and the image capturing cursor value, and taking an absolute value to obtain an illumination deviation coefficient; and carrying out numerical calculation on the spectrum coincidence degree coefficient, the illumination deviation coefficient and the illumination non-uniformity coefficient to obtain an illumination analysis value.
9. The system for detecting the surface defects of the semiconductor based on the image recognition according to claim 5, wherein the server is in communication connection with the semiconductor production line production early warning module, the server sends grade judgment information of all the semiconductor wafers produced in unit time of the corresponding semiconductor production line to the semiconductor production line production early warning module, the semiconductor production line gathers the grade judgment information of all the semiconductor wafers to obtain the number of the superior wafers, the number of the good wafers and the number of the obsolete wafers produced in the corresponding semiconductor production line in unit time, and numerical calculation is carried out on the number of the superior wafers, the number of the good wafers and the number of the obsolete wafers to obtain a production line early warning coefficient; if the line early warning coefficient exceeds a preset line early warning coefficient threshold, generating a line early warning signal and sending the signal to a server, and sending a corresponding early warning when the line early warning signal is generated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311620522.7A CN117635565B (en) | 2023-11-29 | 2023-11-29 | Semiconductor surface defect detection system based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311620522.7A CN117635565B (en) | 2023-11-29 | 2023-11-29 | Semiconductor surface defect detection system based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117635565A true CN117635565A (en) | 2024-03-01 |
| CN117635565B CN117635565B (en) | 2024-05-24 |
Family
ID=90021035
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311620522.7A Active CN117635565B (en) | 2023-11-29 | 2023-11-29 | Semiconductor surface defect detection system based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117635565B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117890389A (en) * | 2024-03-14 | 2024-04-16 | 上海强华实业股份有限公司 | Online detection method and equipment for surface quality of refined products and storage medium thereof |
| CN117990163A (en) * | 2024-04-03 | 2024-05-07 | 赣州华京稀土新材料有限公司 | Automatic performance detection system and method for NdFeB (neodymium iron boron) thin sheet |
| CN118067740A (en) * | 2024-04-18 | 2024-05-24 | 华南理工大学 | Optical measurement and electrical measurement detection equipment for resonator and filter |
| CN118150584A (en) * | 2024-05-10 | 2024-06-07 | 太仓点石航空动力有限公司 | Turbine casing surface quality detection method based on image recognition |
| CN118501159A (en) * | 2024-07-16 | 2024-08-16 | 大连山丘科技有限公司 | Automobile part defect detection method and system based on machine vision |
| CN118608316A (en) * | 2024-08-06 | 2024-09-06 | 广东祺力电子有限公司 | Semiconductor device image data detection system and method based on artificial intelligence |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102305793A (en) * | 2011-05-11 | 2012-01-04 | 苏州天准精密技术有限公司 | Method and equipment for detecting appearance quality of product |
| CN110954553A (en) * | 2019-10-10 | 2020-04-03 | 段美华 | Lens optical detection system |
| CN111260647A (en) * | 2020-03-12 | 2020-06-09 | 南京安科医疗科技有限公司 | CT scanning auxiliary method based on image detection, computer readable storage medium and CT scanning device |
| CN112932522A (en) * | 2021-02-03 | 2021-06-11 | 上海联影医疗科技股份有限公司 | Medical image acquisition method and device and computer equipment |
| US20220223481A1 (en) * | 2019-05-07 | 2022-07-14 | Xuzhou Xinjing Semiconductor Technology Co., Ltd. | Method and system for automatically detecting and controlling defects on wafer |
| CN115295458A (en) * | 2022-08-25 | 2022-11-04 | 魅杰光电科技(上海)有限公司 | Wafer detection system and method |
| CN115345848A (en) * | 2022-08-12 | 2022-11-15 | 童雯 | Quality inspection method of display screen based on big data |
| 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 |
| CN117116791A (en) * | 2023-08-18 | 2023-11-24 | 江苏永鼎股份有限公司 | Flaw testing method and device for wafer processing |
-
2023
- 2023-11-29 CN CN202311620522.7A patent/CN117635565B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102305793A (en) * | 2011-05-11 | 2012-01-04 | 苏州天准精密技术有限公司 | Method and equipment for detecting appearance quality of product |
| US20220223481A1 (en) * | 2019-05-07 | 2022-07-14 | Xuzhou Xinjing Semiconductor Technology Co., Ltd. | Method and system for automatically detecting and controlling defects on wafer |
| CN110954553A (en) * | 2019-10-10 | 2020-04-03 | 段美华 | Lens optical detection system |
| CN111260647A (en) * | 2020-03-12 | 2020-06-09 | 南京安科医疗科技有限公司 | CT scanning auxiliary method based on image detection, computer readable storage medium and CT scanning device |
| CN112932522A (en) * | 2021-02-03 | 2021-06-11 | 上海联影医疗科技股份有限公司 | Medical image acquisition method and device and computer equipment |
| CN115345848A (en) * | 2022-08-12 | 2022-11-15 | 童雯 | Quality inspection method of display screen based on big data |
| CN115295458A (en) * | 2022-08-25 | 2022-11-04 | 魅杰光电科技(上海)有限公司 | Wafer detection system and method |
| 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 |
| CN117116791A (en) * | 2023-08-18 | 2023-11-24 | 江苏永鼎股份有限公司 | Flaw testing method and device for wafer processing |
Non-Patent Citations (2)
| Title |
|---|
| 倪天宇: "晶圆表面缺陷自动检测技术的研究", 中国优秀硕士学位论文全文数据库 (信息科技辑), no. 2023, 31 January 2023 (2023-01-31), pages 135 - 460 * |
| 刘西峰: "基于图像配准的晶圆表面缺陷检测", 仪器仪表与分析监测, no. 3, 27 August 2020 (2020-08-27), pages 1 - 4 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117890389A (en) * | 2024-03-14 | 2024-04-16 | 上海强华实业股份有限公司 | Online detection method and equipment for surface quality of refined products and storage medium thereof |
| CN117890389B (en) * | 2024-03-14 | 2024-05-24 | 上海强华实业股份有限公司 | Online detection method and equipment for surface quality of refined products and storage medium thereof |
| CN117990163A (en) * | 2024-04-03 | 2024-05-07 | 赣州华京稀土新材料有限公司 | Automatic performance detection system and method for NdFeB (neodymium iron boron) thin sheet |
| CN117990163B (en) * | 2024-04-03 | 2024-06-04 | 赣州华京稀土新材料有限公司 | Automatic performance detection system and method for NdFeB (neodymium iron boron) thin sheet |
| CN118067740A (en) * | 2024-04-18 | 2024-05-24 | 华南理工大学 | Optical measurement and electrical measurement detection equipment for resonator and filter |
| CN118150584A (en) * | 2024-05-10 | 2024-06-07 | 太仓点石航空动力有限公司 | Turbine casing surface quality detection method based on image recognition |
| CN118150584B (en) * | 2024-05-10 | 2024-08-02 | 太仓点石航空动力有限公司 | Turbine casing surface quality detection method based on image recognition |
| CN118501159A (en) * | 2024-07-16 | 2024-08-16 | 大连山丘科技有限公司 | Automobile part defect detection method and system based on machine vision |
| CN118608316A (en) * | 2024-08-06 | 2024-09-06 | 广东祺力电子有限公司 | Semiconductor device image data detection system and method based on artificial intelligence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117635565B (en) | 2024-05-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117635565B (en) | Semiconductor surface defect detection system based on image recognition | |
| CN109550712B (en) | Chemical fiber filament tail fiber appearance defect detection system and method | |
| CN109454006B (en) | Detection and classification method based on device for online detection and classification of chemical fiber spindle tripping defects | |
| CN117250208B (en) | Machine vision-based nano-imprint wafer defect accurate detection system and method | |
| CN109255787B (en) | System and method for detecting scratch of silk ingot based on deep learning and image processing technology | |
| CN108445011B (en) | Defect detection system and method based on deep learning | |
| CN112149543B (en) | Building dust recognition system and method based on computer vision | |
| WO2016070462A1 (en) | Histogram of oriented gradient-based display panel defect detection method | |
| CN111652098B (en) | Product surface defect detection method and device | |
| CN113222938A (en) | Chip defect detection method and system and computer readable storage medium | |
| CN113177924A (en) | Industrial production line product flaw detection method | |
| CN111275700A (en) | Terminal defect detection method and system based on deep learning | |
| CN117350967B (en) | Optical detection-based wafer appearance defect detection system and method | |
| CN117974601B (en) | Method and system for detecting surface defects of silicon wafer based on template matching | |
| CN112419261B (en) | Visual acquisition method and device with abnormal point removing function | |
| CN112014407A (en) | Method for detecting surface defects of integrated circuit wafer | |
| CN116309407A (en) | Method for detecting abnormal state of railway contact net bolt | |
| CN112288682A (en) | Electric power equipment defect positioning method based on image registration | |
| CN117455917B (en) | Establishment of false alarm library of etched lead frame and false alarm on-line judging and screening method | |
| CN107545565B (en) | Solar screen plate detection method | |
| CN117911353A (en) | Method for detecting surface defects of circular crystal oscillator wafer | |
| CN112102280B (en) | Method for detecting loosening and loss faults of small part bearing key nut of railway wagon | |
| CN112184619A (en) | Metal part surface defect detection method based on deep learning | |
| CN116071306A (en) | Machine vision detection method for riveting defect of rolling bearing retainer | |
| CN114120318B (en) | Dark field image target point accurate extraction method based on integrated decision tree |
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 |