CN117132822A - Laminated cell of cadmium telluride perovskite and manufacturing method thereof - Google Patents
Laminated cell of cadmium telluride perovskite and manufacturing method thereof Download PDFInfo
- Publication number
- CN117132822A CN117132822A CN202311094836.8A CN202311094836A CN117132822A CN 117132822 A CN117132822 A CN 117132822A CN 202311094836 A CN202311094836 A CN 202311094836A CN 117132822 A CN117132822 A CN 117132822A
- Authority
- CN
- China
- Prior art keywords
- tangent plane
- classification
- local
- vector
- feature vectors
- 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.)
- Withdrawn
Links
- MARUHZGHZWCEQU-UHFFFAOYSA-N 5-phenyl-2h-tetrazole Chemical compound C1=CC=CC=C1C1=NNN=N1 MARUHZGHZWCEQU-UHFFFAOYSA-N 0.000 title claims abstract description 90
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 46
- 239000013598 vector Substances 0.000 claims description 279
- 238000000465 moulding Methods 0.000 claims description 53
- 238000000034 method Methods 0.000 claims description 42
- 238000012545 processing Methods 0.000 claims description 30
- 238000005457 optimization Methods 0.000 claims description 28
- 238000001514 detection method Methods 0.000 claims description 25
- 238000009826 distribution Methods 0.000 claims description 25
- 230000000903 blocking effect Effects 0.000 claims description 18
- 238000007781 pre-processing Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 13
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 abstract description 8
- 238000013473 artificial intelligence Methods 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 abstract description 4
- 239000010410 layer Substances 0.000 description 32
- 238000010586 diagram Methods 0.000 description 23
- 230000008569 process Effects 0.000 description 14
- 230000006870 function Effects 0.000 description 13
- 238000013461 design Methods 0.000 description 11
- 238000004590 computer program Methods 0.000 description 10
- 238000007747 plating Methods 0.000 description 8
- 239000000463 material Substances 0.000 description 7
- 230000008901 benefit Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000013527 convolutional neural network Methods 0.000 description 6
- 238000001755 magnetron sputter deposition Methods 0.000 description 6
- 239000002356 single layer Substances 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000010521 absorption reaction Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 239000010408 film Substances 0.000 description 4
- 238000005065 mining Methods 0.000 description 4
- 238000003058 natural language processing Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 239000010409 thin film Substances 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000003475 lamination Methods 0.000 description 3
- 230000031700 light absorption Effects 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 229910004613 CdTe Inorganic materials 0.000 description 2
- GWEVSGVZZGPLCZ-UHFFFAOYSA-N Titan oxide Chemical compound O=[Ti]=O GWEVSGVZZGPLCZ-UHFFFAOYSA-N 0.000 description 2
- 230000002411 adverse Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 239000011248 coating agent Substances 0.000 description 2
- 238000000576 coating method Methods 0.000 description 2
- 238000005094 computer simulation Methods 0.000 description 2
- 238000013434 data augmentation Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000005525 hole transport Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009022 nonlinear effect Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 239000006096 absorbing agent Substances 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000010549 co-Evaporation Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000000623 plasma-assisted chemical vapour deposition Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000004528 spin coating Methods 0.000 description 1
- 239000004408 titanium dioxide Substances 0.000 description 1
- 230000005641 tunneling Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
-
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/776—Validation; Performance evaluation
-
- H—ELECTRICITY
- H10—SEMICONDUCTOR DEVICES; ELECTRIC SOLID-STATE DEVICES NOT OTHERWISE PROVIDED FOR
- H10K—ORGANIC ELECTRIC SOLID-STATE DEVICES
- H10K39/00—Integrated devices, or assemblies of multiple devices, comprising at least one organic radiation-sensitive element covered by group H10K30/00
- H10K39/10—Organic photovoltaic [PV] modules; Arrays of single organic PV cells
- H10K39/15—Organic photovoltaic [PV] modules; Arrays of single organic PV cells comprising both organic PV cells and inorganic PV cells
-
- 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
- Y02P70/00—Climate change mitigation technologies in the production process for final industrial or consumer products
- Y02P70/50—Manufacturing or production processes characterised by the final manufactured product
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Chemical & Material Sciences (AREA)
- Inorganic Chemistry (AREA)
- Secondary Cells (AREA)
Abstract
A cadmium telluride perovskite laminated cell and a manufacturing method thereof are provided, wherein a section image of the sampled cadmium telluride perovskite laminated cell is obtained; and fully expressing the quality semantic characteristic information about the laminated battery in the section image by adopting an artificial intelligence technology based on deep learning, so as to accurately detect the forming quality of the laminated battery, thereby optimizing the manufacturing quality and efficiency of the laminated battery.
Description
Technical Field
The application relates to the technical field of intelligent manufacturing, in particular to a cadmium telluride perovskite laminated cell and a manufacturing method thereof.
Background
The cadmium telluride perovskite is a novel photovoltaic material and has the characteristics of high efficiency, low cost, environmental protection and the like, so that the cadmium telluride perovskite has a wide application prospect in the field of solar cells. In order to further improve the conversion efficiency of the solar cell, the prior art scheme carries out special structure lamination design on cadmium telluride and perovskite to prepare the laminated cell.
However, in the current manufacturing process of the laminated cell, due to the special properties of materials, there is a case that the molding quality does not meet the predetermined requirements, which affects the performance and lifetime of the solar cell. The traditional quality detection mode of the laminated battery by means of manpower not only consumes a great deal of manpower and energy, but also is often lower in detection precision and efficiency, and is difficult to meet due requirements.
Accordingly, an optimized cadmium telluride perovskite laminate cell and a method of making the same are desired.
Disclosure of Invention
The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides a laminated cell of cadmium telluride perovskite and a manufacturing method thereof, wherein a section image of the sampled laminated cell of cadmium telluride perovskite is acquired; and fully expressing the quality semantic characteristic information about the laminated battery in the section image by adopting an artificial intelligence technology based on deep learning, so as to accurately detect the forming quality of the laminated battery, thereby optimizing the manufacturing quality and efficiency of the laminated battery.
In a first aspect, there is provided a stacked cell of cadmium telluride perovskite comprising:
the image acquisition module is used for acquiring a section image of the sampled cadmium telluride perovskite laminated battery;
the image preprocessing module is used for carrying out image preprocessing on the section image to obtain a preprocessed section image;
the image blocking module is used for carrying out image blocking processing on the preprocessed section image to obtain a sequence of local section image blocks;
the section semantic feature extraction module is used for enabling the sequence of the local section image blocks to pass through a ViT model containing an embedded layer to obtain a plurality of context local section semantic feature vectors;
The similarity correlation module is used for calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors so as to obtain a classification feature vector consisting of the cosine similarity;
the feature optimization module is used for carrying out feature distribution optimization on the classified feature vectors to obtain optimized classified feature vectors; and
and the molding quality detection module is used for passing the optimized classification feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the molding quality of the laminated battery meets the preset requirement.
In the above-mentioned laminated cell of cadmium telluride perovskite, the section semantic feature extraction module includes: the embedding unit is used for carrying out vector embedding on each local section image block in the sequence of the local section image blocks by using an embedding layer of the ViT model so as to obtain a sequence of section image block embedded vectors; and a transform coding unit, configured to input the sequence of the slice image block embedding vectors into a transformer of the ViT model to obtain the plurality of context local slice semantic feature vectors.
In the above-described stacked cell of cadmium telluride perovskite, the conversion coding unit includes: the vector construction subunit is used for one-dimensionally arranging the sequence of the embedding vectors of the tangent plane image blocks to obtain a global tangent plane feature vector; a self-attention subunit, configured to calculate a product between the global tangent plane feature vector and a transpose vector of each tangent plane image block embedding vector in the sequence of tangent plane image block embedding vectors to obtain a plurality of self-attention correlation matrices; the normalization subunit is used for respectively performing normalization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of normalized self-attention correlation matrices; the attention calculating subunit is used for obtaining a plurality of probability values through a Softmax classification function by each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and the attention applying subunit is used for weighting each section image block embedded vector in the sequence of section image block embedded vectors by taking each probability value in the plurality of probability values as a weight so as to obtain the plurality of context local section semantic feature vectors.
In the above-mentioned stacked cell of cadmium telluride perovskite, the similarity-related module includes: the cosine calculating unit is used for calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors according to the following cosine formula to obtain a plurality of cosine similarity; wherein, the cosine formula is:
wherein V is i And V j Representing each two contextual local-section semantic feature vectors of the plurality of contextual local-section semantic feature vectors,and->Representing the context local tangent plane semantic feature vector V i And V j Characteristic value of each position, d (V i ,V j ) Representing cosine distances between any two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors; and an arrangement unit configured to arrange the plurality of similarities to obtain the classification feature vector.
In the above-described stacked cell of cadmium telluride perovskite, the feature optimization module is configured to: carrying out feature distribution optimization on the classification feature vector by using the following optimization formula to obtain the optimized classification feature vector; wherein, the optimization formula is:
Wherein v is i Is the characteristic value of the ith position of the classification characteristic vector, mu and sigma are the mean value and standard deviation of the characteristic value set of each position in the classification characteristic vector, and v i ' is the eigenvalue of the ith position of the optimized classification eigenvector.
In the above-mentioned laminated cell of cadmium telluride perovskite, the molding quality detection module includes: the full-connection coding unit is used for carrying out full-connection coding on the optimized classification feature vector by using a plurality of full-connection layers of the classifier so as to obtain a coding classification feature vector; and the classification unit is used for passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
In a second aspect, there is provided a method of manufacturing a stacked cell of cadmium telluride perovskite, comprising:
acquiring a section image of a sampled cadmium telluride perovskite laminated cell;
performing image preprocessing on the section image to obtain a preprocessed section image;
performing image blocking processing on the preprocessed section image to obtain a sequence of local section image blocks;
passing the sequence of local tangent plane image blocks through a ViT model comprising an embedded layer to obtain a plurality of context local tangent plane semantic feature vectors;
Calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors to obtain a classification feature vector consisting of the cosine similarity;
performing feature distribution optimization on the classification feature vectors to obtain optimized classification feature vectors; and
and the optimized classification feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the molding quality of the laminated battery meets the preset requirement.
In the above method for manufacturing a laminated cell of cadmium telluride perovskite, passing the sequence of partial cut image blocks through a ViT model containing an embedded layer to obtain a plurality of context partial cut semantic feature vectors, comprising: performing vector embedding on each local section image block in the sequence of local section image blocks by using an embedding layer of the ViT model to obtain a sequence of section image block embedded vectors; and inputting the sequence of the tangent plane image block embedded vectors into a converter of the ViT model to obtain the plurality of context local tangent plane semantic feature vectors.
In the above method for manufacturing a laminated cell of cadmium telluride perovskite, inputting the sequence of the cut image block embedding vectors into the converter of the ViT model to obtain the plurality of context local cut semantic feature vectors includes: one-dimensional arrangement is carried out on the sequence of the embedding vectors of the section image blocks so as to obtain a global section feature vector; calculating the product between the global tangent plane feature vector and the transpose vector of each tangent plane image block embedding vector in the sequence of the tangent plane image block embedding vectors to obtain a plurality of self-attention association matrixes; respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and weighting each tangent plane image block embedded vector in the sequence of the tangent plane image block embedded vectors by taking each probability value in the plurality of probability values as a weight to obtain the plurality of context local tangent plane semantic feature vectors.
In the above method for manufacturing a laminated cell of cadmium telluride perovskite, calculating cosine similarity between every two context local tangential semantic feature vectors in the plurality of context local tangential semantic feature vectors to obtain a classification feature vector composed of a plurality of cosine similarities, includes: calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors by using the following cosine formula to obtain a plurality of cosine similarity; wherein, the cosine formula is:
wherein V is i And V j Representing each two contextual local-section semantic feature vectors of the plurality of contextual local-section semantic feature vectors,and->Representing the context local tangent plane semantic feature vector V i And V j Characteristic value of each position, d (V i ,V j ) Representing cosine distances between any two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors; and arranging the plurality of similarities to obtain the classification feature vector.
Compared with the prior art, the cadmium telluride perovskite laminated cell and the manufacturing method thereof provided by the application acquire the section images of the sampled cadmium telluride perovskite laminated cell; and fully expressing the quality semantic characteristic information about the laminated battery in the section image by adopting an artificial intelligence technology based on deep learning, so as to accurately detect the forming quality of the laminated battery, thereby optimizing the manufacturing quality and efficiency of the laminated battery.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic structural design of a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
Fig. 2 is an application scenario diagram of a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
Fig. 3 is a block diagram of a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
FIG. 4 is a block diagram of the cut-plane semantic feature extraction module in a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
Fig. 5 is a block diagram of the transcoding unit in a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
Fig. 6 is a block diagram of the similarity correlation module in a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
Fig. 7 is a block diagram of the molding quality detection module in a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
Fig. 8 is a flow chart of a method of manufacturing a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
Fig. 9 is a schematic diagram of a system architecture of a method of manufacturing a stacked cell of cadmium telluride perovskite according to an embodiment of the application.
Detailed Description
The following description of the technical solutions according to the embodiments of the present application will be given with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
Unless defined otherwise, all technical and scientific terms used in the embodiments of the application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to limit the scope of the present application.
In describing embodiments of the present application, unless otherwise indicated and limited thereto, the term "connected" should be construed broadly, for example, it may be an electrical connection, or may be a communication between two elements, or may be a direct connection, or may be an indirect connection via an intermediate medium, and it will be understood by those skilled in the art that the specific meaning of the term may be interpreted according to circumstances.
It should be noted that, the term "first\second\third" related to the embodiment of the present application is merely to distinguish similar objects, and does not represent a specific order for the objects, it is to be understood that "first\second\third" may interchange a specific order or sequence where allowed. It is to be understood that the "first\second\third" distinguishing objects may be interchanged where appropriate such that embodiments of the application described herein may be practiced in sequences other than those illustrated or described herein.
The CdTe energy gap is about 1.45eV, which is an important film material, and the energy gap is very close to the ideal energy gap of the photovoltaic material, and has very high light absorption coefficient. It is found that a single layer CdTe film of several microns thickness can absorb most of sunlight, but the wavelength range of the absorption of the cadmium telluride film is 300 nm to 800 nm, while near infrared bands above 900 nm in the long band are hardly absorbed, which is certainly wasteful for sunlight utilization, and is single layer.
According to the application, the solar cell structure is mainly designed, so that full-band sunlight can be fully absorbed, and the conversion efficiency is improved.
Drawbacks of the prior art include: 1) The single-layer thin film battery cannot fully utilize sunlight; 2) The absorption wave band of the single-layer cadmium telluride is 300-900 nanometers, and almost no absorption is generated for the near infrared wave band; 3) The single-layer perovskite thin film battery has poor stability and rapid attenuation; 4) The efficiency of a single-layer thin film battery is low; 5) Solar cells employing a single absorber material have very limited potential in improving conversion efficiency.
In view of the drawbacks, the object of the present application is: 1) Through structural design, the cadmium telluride thin film battery and the perovskite battery are laminated, so that the two batteries are in a series structure; 2) Through the structural design, the utilization of full-wave-band sunlight can be increased, and the conversion efficiency is improved; 3) Through the structural design, the problem of battery attenuation can be effectively reduced, and the stability of the battery is improved; 4) The lower temperature coefficient can be obtained, and the generated energy is improved; 5) The cost can be reduced, and the application advantage can be improved.
In one embodiment of the present application, as shown in fig. 1, the technical design of the stacked cell of cadmium telluride perovskite is described as: 1. through structural design, the lamination of cadmium telluride and perovskite is realized; 2. plating a TCO film by adopting a PECVD method, wherein the thickness is 250-350 nanometers; 3. plating a hole transport layer in a slit coating mode, wherein the thickness of the hole transport layer is 5-10 nanometers; 4. coating a perovskite absorption layer in a slit spin coating mode, wherein the thickness is 100-200 nanometers, and the forbidden band width is 1.7-2.1 ev; 5. plating a porous bracket layer by a magnetron sputtering method, wherein the thickness is 50-100 nanometers; 6. plating a titanium dioxide compact layer by a magnetron sputtering method, wherein the thickness is 10-20 nanometers; 7. plating a tunneling junction by adopting magnetron sputtering, wherein the thickness is 2-10 nanometers; 8. plating an HRT layer by adopting magnetron sputtering, wherein the thickness is 10-20 nanometers; 9. plating a tellurium-selenium-cadmium-magnesium light absorption layer in a co-evaporation mode, wherein the thickness of the tellurium-selenium-cadmium-magnesium light absorption layer is 1-2 microns; 10. plating a back contact layer by a magnetron sputtering method, wherein the thickness is 5-15 nanometers; 11. the back electrode layer is formed by a magnetron sputtering method, and the thickness of the back electrode layer is 150-200 nanometers;
Further, the application has the advantages that: 1) Through structural design, the double-junction battery of the cadmium telluride perovskite lamination is realized; 2) Through the structural design, the laminated double-junction battery can effectively absorb sunlight in all wave bands, and the utilization rate of the sunlight is improved; 3) Through the structural design, the laminated battery is more stable and has lower attenuation; 4) Through simulation and actual measurement, the designed structure has better efficiency and improves the generated energy in the system application; 5) Through structural design, the laminated battery has a lower temperature coefficient, and still realizes a good power generation effect after the temperature rises.
As described above, in the current manufacturing process of the laminated cell, there is a case where the molding quality does not meet the predetermined requirement due to the special properties of the material, which affects the performance and lifetime of the solar cell. The traditional quality detection mode of the laminated battery by means of manpower not only consumes a great deal of manpower and energy, but also is often lower in detection precision and efficiency, and is difficult to meet due requirements. Thus, an optimized fabrication scheme for a stacked cell of cadmium telluride perovskite is desired.
Accordingly, in order to be able to detect the quality of the production of the laminated cell of cadmium telluride perovskite in the actual production process of the laminated cell of cadmium telluride perovskite, it is desirable to judge whether the quality of the laminated cell meets the predetermined requirement by analyzing the cut-surface image of the laminated cell of cadmium telluride perovskite. However, since a large amount of interference information may exist in the tangential plane image of the stacked cell of the cadmium telluride perovskite, the semantic feature information about the quality of the stacked cell in the tangential plane image is an implicit small-scale feature, and it is difficult to sufficiently capture and extract. Therefore, in this process, it is difficult to perform sufficient expression of the quality semantic feature information on the laminated battery in the sectional image, thereby accurately performing the molding quality detection of the laminated battery, and thereby optimizing the manufacturing quality and efficiency of the laminated battery.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. The development of deep learning and neural networks provides new solutions and schemes for mining quality semantic feature information about stacked cells in the slice images.
Specifically, in the technical scheme of the application, firstly, a section image of a sampled laminated cell of cadmium telluride perovskite is obtained. It should be appreciated that in the actual process of acquiring the slice images, the acquired original slice images may be affected by noise, deformation or edge distortion, which may adversely affect the accuracy and stability of the subsequent processing algorithm. Therefore, before the feature extraction of the cut surface image and the molding quality analysis of the laminated battery are performed, the image preprocessing is performed on the cut surface image to obtain a preprocessed cut surface image, so that image noise is removed, edges are smoothed, nonlinear effects are corrected, and the like. That is, through such preprocessing, a clearer, more accurate and stable tangent plane image can be obtained, which is helpful for improving the diagnosis and evaluation effects of the subsequent algorithm.
Further, feature mining of the pre-processed tangent plane images is performed using a convolutional neural network model with excellent performance in implicit feature extraction of the images, but the pure CNN method has difficulty in learning explicit global and remote semantic information interactions due to inherent limitations of convolution operations. Further, it is considered that capturing and extracting are difficult because the quality implicit features on the laminated battery in the pre-processed tangential plane images are fine features of a small scale. Therefore, in order to improve the expression capability of the micro-features of the pre-processed section image, which are related to the quality of the laminated battery, in a hidden manner, so as to improve the accuracy of the molding quality detection, in the technical scheme of the application, the pre-processed section image is subjected to image blocking processing and then is encoded in a ViT model containing an embedded layer, so as to extract hidden context semantic association feature distribution information related to the quality of the laminated battery in the pre-processed section image, thereby obtaining a plurality of context local section semantic feature vectors. It should be understood that the image blocking processing is performed on the preprocessed section image to obtain the characteristic information about the quality small-scale hidden characteristic of the laminated battery in each local section image block of the sequence of local section image blocks, which is not the small-scale characteristic information any more, so as to facilitate the subsequent forming quality detection of the laminated battery. In particular, here, the embedding layer linearly projects the individual local slice image blocks as one-dimensional embedding vectors via a learnable embedding matrix. The embedding process is realized by firstly arranging the pixel values of all pixel positions in each local section image block into one-dimensional vectors, and then carrying out full-connection coding on the one-dimensional vectors by using a full-connection layer so as to realize embedding. Here, the ViT model may directly process the respective local section image blocks through a self-attention mechanism like a transducer, so as to extract, from the respective local section image blocks, semantic association feature information about the molding quality of the laminated battery based on the whole of the preprocessed section image.
Then, considering that in the section image of the laminated cell of the cadmium telluride perovskite, if the molding quality of the laminated cell meets the preset requirement, the local section image blocks have similarity or consistency between the semantic association characteristic information of the quality implication context of the laminated cell. Therefore, in order to accurately detect the molding quality of the laminated battery, the similarity association feature information between the quality hidden features of the laminated battery in each two local tangent plane image blocks can be described by calculating the cosine similarity between every two local tangent plane semantic feature vectors in the multiple local tangent plane semantic feature vectors, so that a classification feature vector consisting of multiple cosine similarities is obtained.
And then, the classification feature vector is further passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the molding quality of the laminated battery meets the preset requirement. That is, in the technical solution of the present application, the label of the classifier includes that the molding quality of the laminated battery meets a predetermined requirement (first label) and that the molding quality of the laminated battery does not meet a predetermined requirement (second label), wherein the classifier determines to which classification label the classification feature vector belongs through a soft maximum function. It should be noted that the first label p1 and the second label p2 do not include a manually set concept, and in fact, during the training process, the computer model does not have a concept of "whether the molding quality of the laminated battery meets the predetermined requirement", which is simply that there are two kinds of classification labels and the probability that the output feature is under the two kinds of classification labels, that is, the sum of p1 and p2 is one. Therefore, the classification result of whether the molding quality of the laminated battery meets the preset requirement is actually that the classification label is converted into the classification probability distribution of the two classifications meeting the natural rule, and the physical meaning of the natural probability distribution of the label is essentially used instead of the language text meaning of whether the molding quality of the laminated battery meets the preset requirement. It should be understood that, in the technical solution of the present application, the classification label of the classifier is a detection and evaluation label for whether the molding quality of the laminated battery meets the predetermined requirement, so after the classification result is obtained, the molding quality detection of the laminated battery can be performed based on the classification result, thereby optimizing the manufacturing quality and efficiency of the laminated battery.
Particularly, in the technical scheme of the application, because the cosine similarity between context local tangent plane semantic feature vectors expresses the similarity of the overall feature distribution of the context local tangent plane semantic feature vectors in a high-dimensional feature space, when the tangent plane image is subjected to image blocking processing, the semantic segmentation of the image blocks is uneven, so that a relatively obvious similarity difference exists between the overall distribution of the image semantic features expressed by the context local tangent plane semantic feature vectors (for example, the background similarity is high and the background similarity is low, the object similarity is low), the problem of higher correlation exists in some local distributions of the classification feature vectors formed by the obtained cosine similarity, the problem of the classification feature vectors in the probability density dimension exists, the class probability expression of the classification feature vectors under the classification task is influenced, and the accuracy of classification results obtained by the classification feature vectors through the classifier is reduced.
Therefore, the applicant of the present application orthogonalizes the manifold surface dimensions of the gaussian probability density for the classification feature vector V, specifically expressed as:
Wherein μ and σ are the feature value set v i E means and standard deviation of V, and V i ' is the feature value of the ith position of the classification feature vector after optimization.
Here, by characterizing the unit tangent vector modulo length and the unit normal vector modulo length of the curved surface with the square root of the mean value and standard deviation of the high-dimensional feature set expressing the manifold curved surface, the manifold curved surface of the high-dimensional feature manifold of the classification feature vector V can be subjected to orthogonal projection based on the unit modulo length on the tangent plane and the normal plane, so that the dimensional reconstruction of the probability density of the high-dimensional feature is performed based on the basic structure of the gaussian feature manifold geometry, and the accuracy of class probability expression of the optimized classification feature vector under the classification task is improved by improving the dimensional orthogonalization of the probability density, thereby improving the accuracy of the classification result obtained by the classifier of the optimized classification feature vector. In this way, the molding quality of the laminated battery can be accurately detected, thereby optimizing the manufacturing quality and efficiency of the laminated battery.
Fig. 2 is an application scenario diagram of a stacked cell of cadmium telluride perovskite according to an embodiment of the application. As shown in fig. 2, in this application scenario, first, a cut-plane image (e.g., C as illustrated in fig. 2) of a stacked cell (e.g., M as illustrated in fig. 2) of sampled cadmium telluride perovskite is acquired; the obtained cut surface image is then input into a server (e.g., S as illustrated in fig. 2) that deploys a laminated cell algorithm of cadmium telluride perovskite, wherein the server is capable of processing the cut surface image based on the laminated cell algorithm of cadmium telluride perovskite to generate a classification result for indicating whether the molding quality of the laminated cell meets a predetermined requirement.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
In one embodiment of the application, fig. 3 is a block diagram of a stacked cell of cadmium telluride perovskite according to an embodiment of the application. As shown in fig. 3, a stacked cell 100 of cadmium telluride perovskite according to an embodiment of the present application includes: an image acquisition module 110 for acquiring a tangential image of the sampled cadmium telluride perovskite laminate cell; an image preprocessing module 120, configured to perform image preprocessing on the slice image to obtain a preprocessed slice image; the image blocking module 130 is configured to perform image blocking processing on the preprocessed section image to obtain a sequence of local section image blocks; the section semantic feature extraction module 140 is configured to pass the sequence of the local section image block through a ViT model including an embedding layer to obtain a plurality of context local section semantic feature vector similarity association modules 150, and calculate cosine similarities between every two context local section semantic feature vectors in the plurality of context local section semantic feature vectors to obtain a classification feature vector composed of a plurality of cosine similarities; the feature optimization module 160 is configured to perform feature distribution optimization on the classification feature vector to obtain an optimized classification feature vector; and a molding quality detection module 170, configured to pass the optimized classification feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether the molding quality of the laminated battery meets a predetermined requirement.
Specifically, in the embodiment of the present application, the image acquisition module 110 is configured to acquire a tangential image of the sampled stacked cell of cadmium telluride perovskite. As described above, in the current manufacturing process of the laminated cell, there is a case where the molding quality does not meet the predetermined requirement due to the special properties of the material, which affects the performance and lifetime of the solar cell. The traditional quality detection mode of the laminated battery by means of manpower not only consumes a great deal of manpower and energy, but also is often lower in detection precision and efficiency, and is difficult to meet due requirements. Thus, an optimized fabrication scheme for a stacked cell of cadmium telluride perovskite is desired.
Accordingly, in order to be able to detect the quality of the production of the laminated cell of cadmium telluride perovskite in the actual production process of the laminated cell of cadmium telluride perovskite, it is desirable to judge whether the quality of the laminated cell meets the predetermined requirement by analyzing the cut-surface image of the laminated cell of cadmium telluride perovskite. However, since a large amount of interference information may exist in the tangential plane image of the stacked cell of the cadmium telluride perovskite, the semantic feature information about the quality of the stacked cell in the tangential plane image is an implicit small-scale feature, and it is difficult to sufficiently capture and extract. Therefore, in this process, it is difficult to perform sufficient expression of the quality semantic feature information on the laminated battery in the sectional image, thereby accurately performing the molding quality detection of the laminated battery, and thereby optimizing the manufacturing quality and efficiency of the laminated battery.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. The development of deep learning and neural networks provides new solutions and schemes for mining quality semantic feature information about stacked cells in the slice images.
Specifically, in the technical scheme of the application, firstly, a section image of a sampled laminated cell of cadmium telluride perovskite is obtained.
Specifically, in the embodiment of the present application, the image preprocessing module 120 is configured to perform image preprocessing on the slice image to obtain a preprocessed slice image. It should be appreciated that in the actual process of acquiring the slice images, the acquired original slice images may be affected by noise, deformation or edge distortion, which may adversely affect the accuracy and stability of the subsequent processing algorithm. Therefore, before the feature extraction of the cut surface image and the molding quality analysis of the laminated battery are performed, the image preprocessing is performed on the cut surface image to obtain a preprocessed cut surface image, so that image noise is removed, edges are smoothed, nonlinear effects are corrected, and the like. That is, through such preprocessing, a clearer, more accurate and stable tangent plane image can be obtained, which is helpful for improving the diagnosis and evaluation effects of the subsequent algorithm.
Specifically, in the embodiment of the present application, the image blocking module 130 is configured to perform image blocking processing on the preprocessed slice image to obtain a sequence of local slice image blocks. Further, feature mining of the pre-processed tangent plane images is performed using a convolutional neural network model with excellent performance in implicit feature extraction of the images, but the pure CNN method has difficulty in learning explicit global and remote semantic information interactions due to inherent limitations of convolution operations. Further, it is considered that capturing and extracting are difficult because the quality implicit features on the laminated battery in the pre-processed tangential plane images are fine features of a small scale. Therefore, in order to improve the expression capability of the microscopic features of the hidden small scale of the molding quality of the laminated battery in the pre-processed section image, so as to improve the accuracy of molding quality detection, in the technical scheme of the application, the pre-processed section image is subjected to image blocking processing
Specifically, in the embodiment of the present application, the section semantic feature extraction module 140 is configured to pass the sequence of the local section image blocks through a ViT model including an embedded layer to obtain a plurality of context local section semantic feature vectors. And then, coding in a ViT model containing an embedded layer to extract implicit context semantic association feature distribution information about the quality of the laminated battery in the preprocessed tangent plane image, thereby obtaining a plurality of context local tangent plane semantic feature vectors. It should be understood that the image blocking processing is performed on the preprocessed section image to obtain the characteristic information about the quality small-scale hidden characteristic of the laminated battery in each local section image block of the sequence of local section image blocks, which is not the small-scale characteristic information any more, so as to facilitate the subsequent forming quality detection of the laminated battery.
In particular, here, the embedding layer linearly projects the individual local slice image blocks as one-dimensional embedding vectors via a learnable embedding matrix. The embedding process is realized by firstly arranging the pixel values of all pixel positions in each local section image block into one-dimensional vectors, and then carrying out full-connection coding on the one-dimensional vectors by using a full-connection layer so as to realize embedding. Here, the ViT model may directly process the respective local section image blocks through a self-attention mechanism like a transducer, so as to extract, from the respective local section image blocks, semantic association feature information about the molding quality of the laminated battery based on the whole of the preprocessed section image.
Fig. 4 is a block diagram of the tangential plane semantic feature extraction module in the cadmium telluride perovskite stacked cell according to the embodiment of the present application, as shown in fig. 4, the tangential plane semantic feature extraction module 140 includes: an embedding unit 141, configured to use an embedding layer of the ViT model to perform vector embedding on each local tangential image block in the sequence of local tangential image blocks to obtain a sequence of tangential image block embedded vectors; and a transform coding unit 142, configured to input the sequence of embedding vectors of the slice image blocks into the converter of the ViT model to obtain the plurality of context local slice semantic feature vectors.
Fig. 5 is a block diagram of the transcoding unit in the stacked cell of cadmium telluride perovskite according to an embodiment of the present application, as shown in fig. 5, the transcoding unit 142 includes: a vector construction subunit 1421, configured to perform one-dimensional arrangement on the sequence of the embedding vectors of the tangent plane image blocks to obtain a global tangent plane feature vector; a self-attention subunit 1422, configured to calculate a product between the global tangent plane feature vector and a transpose vector of each tangent plane image block embedding vector in the sequence of tangent plane image block embedding vectors to obtain a plurality of self-attention correlation matrices; a normalization subunit 1423, configured to perform normalization processing on each of the plurality of self-attention correlation matrices to obtain a plurality of normalized self-attention correlation matrices; a attention calculating subunit 1424, configured to obtain a plurality of probability values by using a Softmax classification function from each normalized self-attention correlation matrix in the plurality of normalized self-attention correlation matrices; and an attention applying subunit 1425, configured to weight each slice image block embedding vector in the sequence of slice image block embedding vectors with each probability value in the plurality of probability values as a weight to obtain the plurality of context local slice semantic feature vectors.
It should be understood that since the transducer structure proposed by Google in 2017, a wave of hot surge is rapidly initiated, and for the NLP field, the self-attention mechanism replaces the conventional cyclic neural network structure adopted when processing sequence data, so that not only is parallel training realized, but also the training efficiency is improved, and meanwhile, good results are obtained in application. In NLP, a sequence is input into a transducer, but in the field of vision, how to convert a 2d picture into a 1d sequence needs to be considered, and the most intuitive idea is to input pixels in the picture into the transducer, but the complexity is too high.
While the ViT model can reduce the complexity of input, the picture is cut into image blocks, each image block is projected as a fixed length vector into the transducer, and the operation of the subsequent encoder is identical to that of the original transducer. However, because the pictures are classified, a special mark is added into the input sequence, and the output corresponding to the mark is the final class prediction. ViT exhibits quite excellent performance over many visual tasks, but the lack of inductive biasing allows ViT to be applied to small data sets with very much dependence on model regularization (model regularization) and data augmentation (data augmentation) compared to CNN (Convolutional Neural Network ).
Specifically, in the embodiment of the present application, the similarity association module 150 is configured to calculate cosine similarity between each two context local tangential semantic feature vectors in the context local tangential semantic feature vectors to obtain a classification feature vector composed of the cosine similarities. Then, considering that in the section image of the laminated cell of the cadmium telluride perovskite, if the molding quality of the laminated cell meets the preset requirement, the local section image blocks have similarity or consistency between the semantic association characteristic information of the quality implication context of the laminated cell.
Therefore, in order to accurately detect the molding quality of the laminated battery, the similarity association feature information between the quality hidden features of the laminated battery in each two local tangent plane image blocks can be described by calculating the cosine similarity between every two local tangent plane semantic feature vectors in the multiple local tangent plane semantic feature vectors, so that a classification feature vector consisting of multiple cosine similarities is obtained.
Fig. 6 is a block diagram of the similarity correlation module in the stacked cell of cadmium telluride perovskite according to the embodiment of the present application, as shown in fig. 6, the similarity correlation module 150 includes: a cosine calculating unit 151, configured to calculate cosine similarities between every two context local tangential plane semantic feature vectors in the context local tangential plane semantic feature vectors according to the following cosine formula to obtain a plurality of cosine similarities; wherein, the cosine formula is:
Wherein V is i And V j Representing each two contextual local-section semantic feature vectors of the plurality of contextual local-section semantic feature vectors,and->Representing the context local tangent plane semantic feature vector V i And V j Characteristic value of each position, d (V i ,V j ) Representing cosine distances between any two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors; and an arrangement unit 152 configured to arrange the plurality of similarities to obtain the classification feature vector.
Specifically, in the embodiment of the present application, the feature optimization module 160 is configured to perform feature distribution optimization on the classification feature vector to obtain an optimized classification feature vector. Particularly, in the technical scheme of the application, because the cosine similarity between context local tangent plane semantic feature vectors expresses the similarity of the overall feature distribution of the context local tangent plane semantic feature vectors in a high-dimensional feature space, when the tangent plane image is subjected to image blocking processing, the semantic segmentation of the image blocks is uneven, so that a relatively obvious similarity difference exists between the overall distribution of the image semantic features expressed by the context local tangent plane semantic feature vectors (for example, the background similarity is high and the background similarity is low, the object similarity is low), the problem of higher correlation exists in some local distributions of the classification feature vectors formed by the obtained cosine similarity, the problem of the classification feature vectors in the probability density dimension exists, the class probability expression of the classification feature vectors under the classification task is influenced, and the accuracy of classification results obtained by the classification feature vectors through the classifier is reduced.
Therefore, the applicant of the present application orthogonalizes the manifold surface dimensions of the gaussian probability density for the classification feature vector V, specifically expressed as: carrying out feature distribution optimization on the classification feature vector by using the following optimization formula to obtain the optimized classification feature vector; wherein, the optimization formula is:
wherein v is i Is the characteristic value of the ith position of the classification characteristic vector, mu and sigma are the mean value and standard deviation of the characteristic value set of each position in the classification characteristic vector, and v i ' is the eigenvalue of the ith position of the optimized classification eigenvector.
Here, by characterizing the unit tangent vector modulo length and the unit normal vector modulo length of the curved surface with the square root of the mean value and standard deviation of the high-dimensional feature set expressing the manifold curved surface, the manifold curved surface of the high-dimensional feature manifold of the classification feature vector V can be subjected to orthogonal projection based on the unit modulo length on the tangent plane and the normal plane, so that the dimensional reconstruction of the probability density of the high-dimensional feature is performed based on the basic structure of the gaussian feature manifold geometry, and the accuracy of class probability expression of the optimized classification feature vector under the classification task is improved by improving the dimensional orthogonalization of the probability density, thereby improving the accuracy of the classification result obtained by the classifier of the optimized classification feature vector. In this way, the molding quality of the laminated battery can be accurately detected, thereby optimizing the manufacturing quality and efficiency of the laminated battery.
Specifically, in the embodiment of the present application, the molding quality detection module 170 is configured to pass the optimized classification feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether the molding quality of the laminated battery meets a predetermined requirement. And then, the classification feature vector is further passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the molding quality of the laminated battery meets the preset requirement. That is, in the technical solution of the present application, the label of the classifier includes that the molding quality of the laminated battery meets a predetermined requirement (first label) and that the molding quality of the laminated battery does not meet a predetermined requirement (second label), wherein the classifier determines to which classification label the classification feature vector belongs through a soft maximum function.
It should be noted that the first label p1 and the second label p2 do not include a manually set concept, and in fact, during the training process, the computer model does not have a concept of "whether the molding quality of the laminated battery meets the predetermined requirement", which is simply that there are two kinds of classification labels and the probability that the output feature is under the two kinds of classification labels, that is, the sum of p1 and p2 is one. Therefore, the classification result of whether the molding quality of the laminated battery meets the preset requirement is actually that the classification label is converted into the classification probability distribution of the two classifications meeting the natural rule, and the physical meaning of the natural probability distribution of the label is essentially used instead of the language text meaning of whether the molding quality of the laminated battery meets the preset requirement.
It should be understood that, in the technical solution of the present application, the classification label of the classifier is a detection and evaluation label for whether the molding quality of the laminated battery meets the predetermined requirement, so after the classification result is obtained, the molding quality detection of the laminated battery can be performed based on the classification result, thereby optimizing the manufacturing quality and efficiency of the laminated battery.
Fig. 7 is a block diagram of the molding quality detection module in the stack cell of cadmium telluride perovskite according to the embodiment of the present application, as shown in fig. 7, the molding quality detection module 170 includes: a full-connection encoding unit 171, configured to perform full-connection encoding on the optimized classification feature vector by using a plurality of full-connection layers of the classifier to obtain an encoded classification feature vector; and a classification unit 172, configured to pass the encoded classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
In summary, a stacked cell 100 of cadmium telluride perovskite according to an embodiment of the present application is illustrated that acquires a cut-out image of a sampled stacked cell of cadmium telluride perovskite; and fully expressing the quality semantic characteristic information about the laminated battery in the section image by adopting an artificial intelligence technology based on deep learning, so as to accurately detect the forming quality of the laminated battery, thereby optimizing the manufacturing quality and efficiency of the laminated battery.
In one embodiment of the application, fig. 8 is a flow chart of a method of manufacturing a stacked cell of cadmium telluride perovskite according to an embodiment of the application. As shown in fig. 8, a method for manufacturing a stacked cell of cadmium telluride perovskite according to an embodiment of the present application includes: 210, obtaining a section image of a sampled cadmium telluride perovskite laminated cell; 220, performing image preprocessing on the section image to obtain a preprocessed section image; 230, performing image blocking processing on the preprocessed section image to obtain a sequence of local section image blocks; 240, passing the sequence of local tangent plane image blocks through a ViT model comprising an embedded layer to obtain a plurality of context local tangent plane semantic feature vectors; 250, calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors to obtain a classification feature vector composed of the cosine similarity; 260, performing feature distribution optimization on the classification feature vector to obtain an optimized classification feature vector; and 270, passing the optimized classification feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the molding quality of the laminated battery meets the preset requirement.
Fig. 9 is a schematic diagram of a system architecture of a method of manufacturing a stacked cell of cadmium telluride perovskite according to an embodiment of the application. As shown in fig. 9, in the system architecture of the method for manufacturing a laminated cell of cadmium telluride perovskite, first, a cut-plane image of a sampled laminated cell of cadmium telluride perovskite is acquired; then, carrying out image preprocessing on the section image to obtain a preprocessed section image; then, carrying out image blocking processing on the preprocessed section image to obtain a sequence of local section image blocks; then, the sequence of the local section image blocks passes through a ViT model containing an embedded layer to obtain a plurality of context local section semantic feature vectors; then, calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors to obtain a classification feature vector composed of the cosine similarity; then, carrying out feature distribution optimization on the classification feature vectors to obtain optimized classification feature vectors; and finally, the optimized classification feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the molding quality of the laminated battery meets the preset requirement.
In a specific example, in the above method for manufacturing a stacked cell of cadmium telluride perovskite, passing the sequence of local tangential image blocks through a ViT model containing an embedded layer to obtain a plurality of contextual local tangential semantic feature vectors comprises: performing vector embedding on each local section image block in the sequence of local section image blocks by using an embedding layer of the ViT model to obtain a sequence of section image block embedded vectors; and inputting the sequence of the tangent plane image block embedded vectors into a converter of the ViT model to obtain the plurality of context local tangent plane semantic feature vectors.
In a specific example, in the above method for manufacturing a stacked cell of cadmium telluride perovskite, inputting the sequence of the slice image block embedding vectors into the converter of the ViT model to obtain the plurality of context local slice semantic feature vectors includes: one-dimensional arrangement is carried out on the sequence of the embedding vectors of the section image blocks so as to obtain a global section feature vector; calculating the product between the global tangent plane feature vector and the transpose vector of each tangent plane image block embedding vector in the sequence of the tangent plane image block embedding vectors to obtain a plurality of self-attention association matrixes; respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and weighting each tangent plane image block embedded vector in the sequence of the tangent plane image block embedded vectors by taking each probability value in the plurality of probability values as a weight to obtain the plurality of context local tangent plane semantic feature vectors.
In a specific example, in the above method for manufacturing a stacked cell of cadmium telluride perovskite, calculating cosine similarities between every two contextual local tangential semantic feature vectors in the plurality of contextual local tangential semantic feature vectors to obtain a classification feature vector composed of the plurality of cosine similarities includes: calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors by using the following cosine formula to obtain a plurality of cosine similarity; wherein, the cosine formula is:
wherein V is i And V j Representing each two contextual local-section semantic feature vectors of the plurality of contextual local-section semantic feature vectors,and->Representing the context local tangent plane semantic feature vector V i And V j Characteristic value of each position, d (V i ,V j ) Representing cosine distances between any two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors; and arranging the plurality of similarities to obtain the classification feature vector.
In a specific example, in the above method for manufacturing a stacked cell of cadmium telluride perovskite, the feature distribution optimization of the classification feature vector to obtain an optimized classification feature vector includes: carrying out feature distribution optimization on the classification feature vector by using the following optimization formula to obtain the optimized classification feature vector; wherein, the optimization formula is:
Wherein v is i Is the characteristic value of the ith position of the classification characteristic vector, mu and sigma are the mean value and standard deviation of the characteristic value set of each position in the classification characteristic vector, and v i ' is the eigenvalue of the ith position of the optimized classification eigenvector.
In a specific example, in the above method for manufacturing a laminated cell of cadmium telluride perovskite, the optimizing classification feature vector is passed through a classifier to obtain a classification result, the classification result being used to indicate whether the molding quality of the laminated cell meets a predetermined requirement, including: performing full-connection coding on the optimized classification feature vector by using a plurality of full-connection layers of the classifier to obtain a coding classification feature vector; and passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
It will be appreciated by those skilled in the art that the specific operations of the respective steps in the above-described method of manufacturing a laminated cell of cadmium telluride perovskite have been described in detail in the above description of the laminated cell of cadmium telluride perovskite with reference to fig. 1 to 7, and thus, repetitive descriptions thereof will be omitted.
The present application also provides a computer program product comprising instructions which, when executed, cause an apparatus to perform operations corresponding to the above-described method.
In one embodiment of the present application, there is also provided a computer-readable storage medium storing a computer program for executing the above-described method.
It should be appreciated that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the forms of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects may be utilized. Furthermore, the computer program product may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
Methods, systems, and computer program products of embodiments of the present application are described in the flow diagrams and/or block diagrams. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The basic principles of the present application have been described above in connection with specific embodiments, however, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not intended to be limiting, and these advantages, benefits, effects, etc. are not to be considered as essential to the various embodiments of the present application. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, as the application is not necessarily limited to practice with the above described specific details.
The block diagrams of the devices, apparatuses, devices, systems referred to in the present application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.
It is also noted that in the apparatus, devices and methods of the present application, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent aspects of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Finally, it is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or terminal device comprising the element.
The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit embodiments of the application to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.
Claims (10)
1. A cadmium telluride perovskite stacked cell comprising:
the image acquisition module is used for acquiring a section image of the sampled cadmium telluride perovskite laminated battery;
the image preprocessing module is used for carrying out image preprocessing on the section image to obtain a preprocessed section image;
the image blocking module is used for carrying out image blocking processing on the preprocessed section image to obtain a sequence of local section image blocks;
the section semantic feature extraction module is used for enabling the sequence of the local section image blocks to pass through a ViT model containing an embedded layer to obtain a plurality of context local section semantic feature vectors;
the similarity correlation module is used for calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors so as to obtain a classification feature vector consisting of the cosine similarity;
the feature optimization module is used for carrying out feature distribution optimization on the classified feature vectors to obtain optimized classified feature vectors; and
and the molding quality detection module is used for passing the optimized classification feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the molding quality of the laminated battery meets the preset requirement.
2. The cadmium telluride perovskite stack cell of claim 1, wherein the cut surface semantic feature extraction module comprises:
the embedding unit is used for carrying out vector embedding on each local section image block in the sequence of the local section image blocks by using an embedding layer of the ViT model so as to obtain a sequence of section image block embedded vectors; and
and the conversion coding unit is used for inputting the sequence of the embedding vectors of the section image blocks into the converter of the ViT model to obtain the plurality of context local section semantic feature vectors.
3. The cadmium telluride perovskite stack cell of claim 2, wherein the conversion coding unit comprises:
the vector construction subunit is used for one-dimensionally arranging the sequence of the embedding vectors of the tangent plane image blocks to obtain a global tangent plane feature vector;
a self-attention subunit, configured to calculate a product between the global tangent plane feature vector and a transpose vector of each tangent plane image block embedding vector in the sequence of tangent plane image block embedding vectors to obtain a plurality of self-attention correlation matrices;
the normalization subunit is used for respectively performing normalization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of normalized self-attention correlation matrices;
The attention calculating subunit is used for obtaining a plurality of probability values through a Softmax classification function by each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and
and the attention applying subunit is used for weighting each slice image block embedded vector in the sequence of slice image block embedded vectors by taking each probability value in the plurality of probability values as a weight so as to obtain the plurality of context local slice semantic feature vectors.
4. A stacked cell as claimed in claim 3, wherein the similarity correlation module comprises:
the cosine calculating unit is used for calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors according to the following cosine formula to obtain a plurality of cosine similarity;
wherein, the cosine formula is:
wherein V is i And V j Representing each two contextual local-section semantic feature vectors of the plurality of contextual local-section semantic feature vectors,and->Representing the context local tangent plane semantic feature vector V i And V j Characteristic value of each position, d (V i ,V j ) Representing cosine distances between any two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors; and
and the arrangement unit is used for arranging the plurality of similarity degrees to obtain the classification characteristic vector.
5. The cadmium telluride perovskite stacked cell as defined in claim 4, wherein the feature optimization module is configured to: carrying out feature distribution optimization on the classification feature vector by using the following optimization formula to obtain the optimized classification feature vector;
wherein, the optimization formula is:
wherein v is i Is the characteristic value of the ith position of the classification characteristic vector, mu and sigma are the mean value and standard deviation of the characteristic value set of each position in the classification characteristic vector, and v i ' is the eigenvalue of the ith position of the optimized classification eigenvector.
6. The cadmium telluride perovskite stack cell of claim 5, wherein the molding quality detection module comprises:
the full-connection coding unit is used for carrying out full-connection coding on the optimized classification feature vector by using a plurality of full-connection layers of the classifier so as to obtain a coding classification feature vector; and
And the classification unit is used for passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
7. A method for manufacturing a cadmium telluride perovskite laminated cell, comprising:
acquiring a section image of a sampled cadmium telluride perovskite laminated cell;
performing image preprocessing on the section image to obtain a preprocessed section image;
performing image blocking processing on the preprocessed section image to obtain a sequence of local section image blocks;
passing the sequence of local tangent plane image blocks through a ViT model comprising an embedded layer to obtain a plurality of context local tangent plane semantic feature vectors;
calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors to obtain a classification feature vector consisting of the cosine similarity;
performing feature distribution optimization on the classification feature vectors to obtain optimized classification feature vectors; and
and the optimized classification feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the molding quality of the laminated battery meets the preset requirement.
8. The method of claim 7, wherein passing the sequence of partial cut image blocks through a ViT model comprising an embedded layer to obtain a plurality of contextual partial cut semantic feature vectors comprises:
performing vector embedding on each local section image block in the sequence of local section image blocks by using an embedding layer of the ViT model to obtain a sequence of section image block embedded vectors; and
the sequence of the tangent plane image block embedding vectors is input to a converter of the ViT model to obtain the plurality of context local tangent plane semantic feature vectors.
9. The method of claim 8, wherein inputting the sequence of slice image block embedding vectors into the ViT model converter to obtain the plurality of contextual local slice semantic feature vectors comprises:
one-dimensional arrangement is carried out on the sequence of the embedding vectors of the section image blocks so as to obtain a global section feature vector;
calculating the product between the global tangent plane feature vector and the transpose vector of each tangent plane image block embedding vector in the sequence of the tangent plane image block embedding vectors to obtain a plurality of self-attention association matrixes;
Respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices;
obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and
and weighting each tangent plane image block embedded vector in the sequence of the tangent plane image block embedded vector by taking each probability value in the plurality of probability values as a weight so as to obtain the plurality of context local tangent plane semantic feature vectors.
10. The method of manufacturing a stacked cell of cadmium telluride perovskite according to claim 9, wherein calculating cosine similarities between every two of the plurality of context local tangential semantic feature vectors to obtain a classification feature vector consisting of the plurality of cosine similarities comprises:
calculating cosine similarity between every two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors by using the following cosine formula to obtain a plurality of cosine similarity;
wherein, the cosine formula is:
Wherein V is i And V j Representing each two contextual local-section semantic feature vectors of the plurality of contextual local-section semantic feature vectors,and->Representing the context local tangent plane semantic feature vector V i And V j Characteristic value of each position, d (V i ,V j ) Representing cosine distances between any two context local tangent plane semantic feature vectors in the context local tangent plane semantic feature vectors; and
and arranging the plurality of similarities to obtain the classification feature vector.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311094836.8A CN117132822A (en) | 2023-08-29 | 2023-08-29 | Laminated cell of cadmium telluride perovskite and manufacturing method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311094836.8A CN117132822A (en) | 2023-08-29 | 2023-08-29 | Laminated cell of cadmium telluride perovskite and manufacturing method thereof |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117132822A true CN117132822A (en) | 2023-11-28 |
Family
ID=88862420
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311094836.8A Withdrawn CN117132822A (en) | 2023-08-29 | 2023-08-29 | Laminated cell of cadmium telluride perovskite and manufacturing method thereof |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117132822A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117350088A (en) * | 2023-12-06 | 2024-01-05 | 苏州易来科得科技有限公司 | Method, device, storage medium and equipment for generating simulation grid of battery pole piece |
CN117876366A (en) * | 2024-03-11 | 2024-04-12 | 宝鸡子扬双金属材料有限公司 | Titanium tube quality detection method and system based on image processing |
-
2023
- 2023-08-29 CN CN202311094836.8A patent/CN117132822A/en not_active Withdrawn
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117350088A (en) * | 2023-12-06 | 2024-01-05 | 苏州易来科得科技有限公司 | Method, device, storage medium and equipment for generating simulation grid of battery pole piece |
CN117350088B (en) * | 2023-12-06 | 2024-02-23 | 苏州易来科得科技有限公司 | Method, device, storage medium and equipment for generating simulation grid of battery pole piece |
CN117876366A (en) * | 2024-03-11 | 2024-04-12 | 宝鸡子扬双金属材料有限公司 | Titanium tube quality detection method and system based on image processing |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN117132822A (en) | Laminated cell of cadmium telluride perovskite and manufacturing method thereof | |
US12068717B2 (en) | Photovoltaic array fault diagnosis method based on composite information | |
US20230114877A1 (en) | Unsupervised Latent Low-Rank Projection Learning Method for Feature Extraction of Hyperspectral Images | |
Chen et al. | Accurate modeling of photovoltaic modules using a 1-D deep residual network based on IV characteristics | |
CN110321777B (en) | Face recognition method based on stacked convolution sparse denoising autoencoder | |
CN103177265A (en) | High-definition image classification method based on kernel function and sparse coding | |
CN115759435A (en) | Photovoltaic power generation power prediction method based on improved CNN-LSTM | |
Zheng et al. | Recognition method of voltage sag causes based on two‐dimensional transform and deep learning hybrid model | |
CN115564092A (en) | Short-time wind power prediction system and method for wind power plant | |
CN115660182A (en) | Photovoltaic output prediction method based on maximum expected sample weighted neural network model | |
CN111539482A (en) | RBF kernel function-based space multidimensional wind power data dimension reduction and reconstruction method | |
Ji et al. | Few-shot scene classification using auxiliary objectives and transductive inference | |
Yu et al. | Short-term solar irradiance prediction based on spatiotemporal graph convolutional recurrent neural network | |
Liu et al. | Scene classification for remote sensing images with self‐attention augmented CNN | |
Wu et al. | Ghost-RetinaNet: Fast Shadow DetectionMethod for Photovoltaic Panels Based on Improved RetinaNet. | |
An et al. | Ultra-short-term prediction method of PV power output based on the CNN–LSTM hybrid learning model driven by EWT | |
Eastwood et al. | Self-supervised disentanglement by leveraging structure in data augmentations | |
Li et al. | LSA based multi-instance learning algorithm for image retrieval | |
CN116186350B (en) | Power transmission line engineering searching method and device based on knowledge graph and topic text | |
CN117113243A (en) | Photovoltaic equipment abnormality detection method | |
CN116994040A (en) | Image recognition-based deep sea wind power generation PQDs (pulse-height distribution system) classification method and system | |
Kong et al. | Feature extraction of load curve based on autoencoder network | |
Wang et al. | Multi-view deep metric learning for volumetric image recognition | |
CN114387466A (en) | Solar cell color difference detection method and system | |
CN113469269A (en) | Residual convolution self-coding wind-solar-charged scene generation method based on multi-channel fusion |
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 | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20231128 |