CN107831173A - Photovoltaic component defect detection method and system - Google Patents
Photovoltaic component defect detection method and system Download PDFInfo
- Publication number
- CN107831173A CN107831173A CN201710964119.4A CN201710964119A CN107831173A CN 107831173 A CN107831173 A CN 107831173A CN 201710964119 A CN201710964119 A CN 201710964119A CN 107831173 A CN107831173 A CN 107831173A
- Authority
- CN
- China
- Prior art keywords
- images
- photovoltaic
- module
- obtains
- effective
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the present invention, which provides a kind of photovoltaic component defect detection method and system, methods described, to be included:Obtain the first EL images of the first photovoltaic cells to be detected;The first EL images are input in photovoltaic module defects detection model, obtain the first defect classification of first photovoltaic cells, the photovoltaic component defect detection method can realize automatic identification and the classification of photovoltaic cells defect, the workload of worker is reduced, the cost of detection is reduced, improves accuracy of detection and efficiency.
Description
Technical field
The present embodiments relate to solar photovoltaic assembly technical field, more particularly to a kind of photovoltaic module defects detection side
Method and system.
Background technology
Photovoltaic generation turns into the generation of electricity by new energy form being most widely used in various new energy at present.Whole photovoltaic electric
Most crucial part is exactly solar photovoltaic cell panel in standing, by shadow the defects of inevitable generation in its production, installation process
Its operating efficiency is rung, therefore is extremely necessary to carry out defects detection to solar photovoltaic cell panel.
Existing solar photovoltaic cell panel defects detection mode application more maturation mainly has infrared thermal imaging and electricity
Photoluminescence two ways, wherein infrared thermal imaging are applied to defects detection in a wide range of, and this mode can only typically detect hot spot
Defect, therefore accuracy of detection is relatively low.Electroluminescent (Electroluminescent, hereinafter referred to as:EL) mode is applied to monolithic
The defects of photovoltaic module, is detected, it can it is more clearer than infrared thermal imaging displaying defect details, be generally used for monolithic element
It is hidden to split defects detection.
However, in the prior art, when carrying out defects detection to solar photovoltaic assembly using electroluminescent mode, it can only adopt
With the mode of sampling observation, type the defects of artificial judgment photovoltaic module, for fairly large photovoltaic plant, need in this way
Expend substantial amounts of artificial and time cost.
The content of the invention
For problems of the prior art, the embodiment of the present invention provides a kind of photovoltaic component defect detection method and is
System.
In a first aspect, the embodiment of the present invention provides a kind of photovoltaic component defect detection method, methods described includes:
Obtain the first EL images of the first photovoltaic cells to be detected;
The first EL images are input in photovoltaic module defects detection model, obtain first photovoltaic cells
The first defect classification.
Second aspect, the embodiment of the present invention provide a kind of photovoltaic module defect detecting system, and the system includes:
First acquisition module, for obtaining the first EL images of the first photovoltaic cells to be detected;
Detection module, for the first EL images to be input into photovoltaic module defects detection model, obtain described first
First defect classification of photovoltaic cells.
The third aspect, the embodiment of the present invention provide a kind of photovoltaic module defect detection equipment, and the equipment includes memory
And processor, the processor and the memory complete mutual communication by bus;The memory storage has can quilt
The programmed instruction of the computing device, the processor call described program instruction to be able to carry out above-mentioned photovoltaic module defect inspection
Survey method.
Fourth aspect, the embodiment of the present invention provide a kind of computer-readable recording medium, are stored thereon with computer program,
The computer program realizes above-mentioned photovoltaic component defect detection method when being executed by processor.
Photovoltaic component defect detection method provided in an embodiment of the present invention and system, by obtaining the first photovoltaic to be detected
First EL images of battery unit, the first EL images are input in photovoltaic module defects detection model, obtain described
First defect classification of one photovoltaic cells, it is possible to achieve the automatic identification of photovoltaic cells defect and classification, reduce
Artificial workload, reduce the cost of detection, improve accuracy of detection and efficiency.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are this hairs
Some bright embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can be with root
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is photovoltaic component defect detection method flow chart provided in an embodiment of the present invention;
Fig. 2 is the structural representation of photovoltaic module defect detecting system provided in an embodiment of the present invention;
Fig. 3 is the structural representation of photovoltaic module defect detection equipment provided in an embodiment of the present invention.
Embodiment
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is explicitly described, it is clear that described embodiment be the present invention
Part of the embodiment, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not having
The every other embodiment obtained under the premise of creative work is made, belongs to the scope of protection of the invention.
Fig. 1 is photovoltaic component defect detection method flow chart provided in an embodiment of the present invention, as shown in figure 1, methods described
Including:
Step 10, the first EL images for obtaining the first photovoltaic cells to be detected;
Step 11, the first EL images are input in photovoltaic module defects detection model, obtain first photovoltaic
First defect classification of battery unit.
Specifically, for the first photovoltaic cells of existing defects, want to detect its defect classification, firstly, it is necessary to
The first EL images of first photovoltaic cells are obtained, then, the first EL images are input to photovoltaic module defect
In detection model, according to the photovoltaic module defects detection model and the first EL images, described first can be calculated
The defects of photovoltaic cells classification.Using photovoltaic cells A as the first photovoltaic cells, the photovoltaic cells A
EL images as the first EL images exemplified by, the defects of wanting to detect the photovoltaic cells A classification, first, described in acquisition
Photovoltaic cells A EL images, then, the EL images of the photovoltaic cells A are input to photovoltaic module defects detection
In model, it is possible to the defects of obtaining photovoltaic cells A classification.
Photovoltaic component defect detection method provided in an embodiment of the present invention, by obtaining the first photovoltaic cell list to be detected
First EL images of member, the first EL images are input in photovoltaic module defects detection model, obtain first photovoltaic
First defect classification of battery unit, it is possible to achieve the automatic identification of photovoltaic cells defect and classification, reduce artificial
Workload, reduce the cost of detection, improve accuracy of detection and efficiency.
Optionally, on the basis of above-described embodiment, methods described also includes the photovoltaic module defects detection model
Process is established, the process includes:
Obtain the 2nd EL images of the second photovoltaic cells, to the 2nd EL images according to predetermined registration operation at
Reason, obtains second effective EL images;
Obtain the second defect classification of described second effective EL images;
Described second effective EL images and the second defect classification are input in the convolutional neural networks, carried out deep
Degree study, obtains the photovoltaic module defects detection model.
Specifically, it is as follows that the photovoltaic module defects detection model referred in above method embodiment establishes process:First, need
The 2nd EL images of multiple second photovoltaic cells are obtained, then, to the 2nd all EL images according to default operation
Handled, obtain multiple second effective EL images.Then, the defects of each second effective EL image corresponds to classification is obtained,
Such as can according to the feature of each second effective EL image, the defects of artificial judgment second effective EL images correspond to class
Not, so as to obtain multi-group data, every group of data include a second effective EL image and its it is corresponding the defects of classification.
Then, using every group of data as sample data, it is input in convolutional neural networks and carries out deep learning, have with second
Imitate input of the EL images as model, the defects of being corresponded to using described second effective EL images classification as model output, to light
Volt component defects detection model is trained.In order to improve the training effectiveness of the model, can rule of thumb data, to described
The key parameter of photovoltaic module defects detection model, such as weight and deviation, initial experience value is set respectively.The convolutional Neural
Structure in network is according to the convolutional neural networks model of classics, such as AlexNet and VGG etc., is designed.The present invention
Convolutional neural networks structure in embodiment can include two convolutional layers, two pond layers and three full articulamentums, network
Input can be 100 × 100 gray level image, wherein, the convolutional layer can use 5x5 and 3x3 convolution kernel, can pass through
Multiple small convolution kernels are superimposed, to improve the performance of network.
During being trained to photovoltaic module defects detection model, supervised learning algorithm can be used, such as instead
To propagation algorithm (back propagation), model training process is exercised supervision, specifically can use L2 regularizations and
Dropout algorithm limited model training process, prevents over-fitting.The over-fitting refers to, by the sample data
Second effective EL images are input to photovoltaic module defects detection model, and the prediction defect classification and actual defects classification of output are very
It is close, still, after second effective EL images outside sample data are input into the photovoltaic module defects detection model, output
Prediction defect classification and actual defects classification difference it is very big.If second effective EL images are input to the photovoltaic module after training
After defects detection model, the prediction defect classification of output and it is actual the defects of classification between loss function value be less than default threshold
During value, deconditioning process, final photovoltaic module defects detection model is obtained.
Photovoltaic component defect detection method provided in an embodiment of the present invention, pass through the 2nd EL to the second photovoltaic cells
Image is handled according to predetermined registration operation, filters out second effective EL images, and by described second effective EL images and its correspondingly
It is actual the defects of classification as sample data, be input in convolutional neural networks, photovoltaic module defects detection model carried out
Training, the photovoltaic module defects detection model after being optimized, this mode cause the photovoltaic component defect detection method more
Add science, improve the precision of photovoltaic module defects detection.
It is optionally, described that the 2nd EL images are handled according to predetermined registration operation on the basis of above-described embodiment,
Second effective EL images are obtained, including:
Rotation process is carried out to the 2nd EL images, obtains the second rotation EL images;
Light and shade operation is carried out to the 2nd EL images, obtains the second light and shade EL images;
Horizontal transformation operation is carried out to the 2nd EL images, obtains the second horizontal transformation EL images;
Fuzzy operation is carried out to the 2nd EL images, obtains the second fuzzy EL images;
Random cropping operation is carried out to the 2nd EL images, obtains the second random cropping EL images;
To the 2nd EL images, the second rotation EL images, the second light and shade EL images, the second horizontal change
EL images, the second fuzzy EL images and the second random cropping EL images are changed, carries out normalizing according to preset algorithm respectively
Change is handled, and obtains described second effective EL images.
Specifically, the second effective EL images referred in above-described embodiment are obtained according to predetermined registration operation as described below
's:Rotation process is carried out to the 2nd EL images, corresponding second rotation EL images can be obtained;The 2nd EL images are carried out
Light and shade operates, and can obtain corresponding second light and shade EL images;Horizontal transformation operation is carried out to the 2nd EL images, can be obtained
To corresponding second horizontal transformation EL images;Fuzzy operation is carried out to the 2nd EL images, corresponding second mould can be obtained
Paste EL images;Random cropping operation is carried out to the 2nd EL images, corresponding second random cropping EL images can be obtained.Through
Cross after aforesaid operations, each the 2nd EL image can be expanded into multiple the 2nd EL images with identical defect classification,
So as to add sample data volume.
Then, by the 2nd EL images and its corresponding second rotation EL images, the second light and shade EL images, the second level
EL images, the second fuzzy EL images and the second random cropping EL images are converted, is normalized, obtains according to default algorithm
To corresponding second effective EL images.For example Regularization, the Regularization can be carried out to each above-mentioned image
The formula used is:
Wherein, wherein x is the image pixel of before processing, and mean is image pixel average, and stddev is image pixel standard
Difference, y are the image pixel after processing.
Photovoltaic component defect detection method provided in an embodiment of the present invention, by being rotated respectively to the 2nd EL images,
Light and shade, horizontal transformation, the operation of fuzzy and random cropping, and to the 2nd EL images and its corresponding second rotation EL figures
Picture, the second light and shade EL images, the second horizontal transformation EL images, the second fuzzy EL images and the second random cropping EL images are returned
One change is handled, and adds the quantity of sample, improves the precision and generalization of photovoltaic defects detection.It is provided in an embodiment of the present invention
Photovoltaic component defect detection method, it is possible to achieve between being -20 °~20 ° with any anglec of rotation and in angle of inclination
Photovoltaic module carry out defect classification detection.
Optionally, on the basis of above-described embodiment, the first EL images are being input to photovoltaic module defects detection
Before model, in addition to:
According to default Morphology Algorithm, processing is filled to the first EL images;
According to default filtering algorithm, processing is filtered to the first EL images by the filling processing;
According to default edge detection algorithm, edge extracting is carried out to the first EL images Jing Guo the filtering process
Processing.
Specifically, the first EL images of the first photovoltaic cells to be detected are typically from the EL of a photovoltaic module
Image, still, the EL images of multiple photovoltaic cells can be included in the EL images of each photovoltaic module.It is therefore desirable to first
From the EL images of the photovoltaic module, the first EL images of first photovoltaic cells are extracted.Detailed process is:It is first
First according to default Morphology Algorithm such as closing operation of mathematical morphology, processing is filled to the first EL images, described in filling
Small cavity in first EL display foreground colors, especially weaken the influence of two main gate lines on the first photovoltaic cells;So
Afterwards, according to default filtering algorithm such as bilateral filtering algorithm, the first EL images by the filling processing are carried out
Filtering process;Finally, according to default edge detection algorithm such as Canny edge detection algorithms, to passing through the filtering process
The first EL images carry out edge extracting processing, from the EL images of the photovoltaic module, extract first battery
First EL images of unit.
Photovoltaic component defect detection method provided in an embodiment of the present invention, pass through to the first battery unit to be detected
One EL images are filled processing, filtering process and edge extracting processing successively, are extracted from photovoltaic module image described
First EL images so that the photovoltaic component defect detection method more science.
Fig. 2 is the structural representation of photovoltaic module defect detecting system provided in an embodiment of the present invention, as shown in Fig. 2 institute
The system of stating includes:First acquisition module 20 and detection module 21, wherein:
First acquisition module 20 is used for the first EL images for obtaining the first photovoltaic cells to be detected;Detection module 21
For the first EL images to be input into photovoltaic module defects detection model, the first of first photovoltaic cells is obtained
Defect classification.
Specifically, for the first photovoltaic cells of existing defects, want to detect its defect classification, firstly, it is necessary to
First acquisition module 20 obtains the first EL images of first photovoltaic cells, it is then detected that module 21 is by described first
The first EL images that acquisition module 20 is got are input in photovoltaic module defects detection model, according to the photovoltaic module
Defects detection model and the first EL images, the defects of calculating first photovoltaic cells classification.With photovoltaic electric
Pool unit A exemplified by the EL images of the photovoltaic cells A are as the first EL images, is wanted as the first photovoltaic cells
The defects of detecting photovoltaic cells A classification, first, the first acquisition module 20 obtain the EL of the photovoltaic cells A
Image, it is then detected that the EL images input for the photovoltaic cells A that module 21 gets first acquisition module 20
Into photovoltaic module defects detection model, it is possible to the defects of obtaining photovoltaic cells A classification.
Photovoltaic module defect detecting system provided in an embodiment of the present invention, its function referring in particular to above method embodiment,
Here is omitted.
Photovoltaic module defect detecting system provided in an embodiment of the present invention, by obtaining the first photovoltaic cell list to be detected
First EL images of member, the first EL images are input in photovoltaic module defects detection model, obtain first photovoltaic
First defect classification of battery unit, it is possible to achieve the automatic identification of photovoltaic cells defect and classification, reduce artificial
Workload, reduce the cost of detection, improve accuracy of detection and efficiency.
Optionally, on the basis of above-described embodiment, the system includes:First acquisition module, detection module, second are obtained
Modulus block, the 3rd acquisition module and training module, wherein:
Second acquisition module be used for obtain the second photovoltaic cells the 2nd EL images, to the 2nd EL images according to
Predetermined registration operation is handled, and obtains second effective EL images;3rd acquisition module is used to obtain described second effective EL images
Second defect classification;Training module is used to described second effective EL images and the second defect classification being input to the convolution
In neutral net, deep learning is carried out, obtains the photovoltaic module defects detection model.
Specifically, first acquisition module and the detection module are described in detail in the above-described embodiments, herein not
Repeat again.Second acquisition module can obtain the 2nd EL images of the second photovoltaic cells, to the 2nd EL images
Handled according to default operation, obtain second effective EL images.3rd acquisition module can obtain described second effective EL
The defects of image corresponds to classification, such as, can be according to the feature of each second effective EL image, this is second effective for artificial judgment
The defects of EL images correspond to classification, obtains multi-group data, and every group of data include a second effective EL image and its corresponding lacked
Fall into classification.
Training module can be input in convolutional neural networks using every group of data as sample data and carries out deep learning,
Using second effective EL images as the input of model, the defects of being corresponded to using described second effective EL images classification be used as the defeated of model
Go out, photovoltaic module defects detection model is trained.In order to improve the training effectiveness of the model, can rule of thumb count
According to the key parameter of the photovoltaic module defects detection model, such as weight and deviation, setting initial experience value respectively.Institute
It is according to the convolutional neural networks model of classics, such as AlexNet and VGG etc. to state the structure in convolutional neural networks, is set
Meter.Convolutional neural networks structure in the embodiment of the present invention can connect entirely comprising two convolutional layers, two pond layers and three
Layer is connect, the input of network can be 100 × 100 gray level image, wherein, the convolutional layer can use 5x5 and 3x3 convolution
Core, can be by being superimposed multiple small convolution kernels, to improve the performance of network.
During training module is trained to photovoltaic module defects detection model, supervised learning algorithm can be used,
Such as back-propagation algorithm (back propagation), model training process is exercised supervision, can specifically use L2 canonicals
Change and dropout algorithm limited model training process, prevent over-fitting.The over-fitting refers to, by the sample data
In second effective EL images be input to photovoltaic module defects detection model, prediction defect classification and the actual defects classification of output
Closely, still, after second effective EL images outside sample data being input into the photovoltaic module defects detection model,
The prediction defect classification of output and actual defects classification difference are very big.If second effective EL images are input to training by training module
After photovoltaic module defects detection model afterwards, the prediction defect classification of output and it is actual the defects of classification between loss function value
During less than default threshold value, deconditioning process, so as to obtain final photovoltaic module defects detection model.
Photovoltaic module defective system provided in an embodiment of the present invention, pass through the 2nd EL images to the second photovoltaic cells
Handled according to predetermined registration operation, filter out second effective EL images, and by described second effective EL images and its corresponding reality
Classification was input in convolutional neural networks, photovoltaic module defects detection model was trained as sample data the defects of border,
Photovoltaic module defects detection model after being optimized, this mode cause the photovoltaic module defect detecting system more section
Learn, improve the precision of photovoltaic module defects detection.
Optionally, on the basis of above-described embodiment, second acquisition module includes:Rotate submodule, light and shade submodule
Block, horizontal transformation submodule, fuzzy submodule, cutting submodule and normalization submodule, wherein:
Rotate submodule to be used to carry out rotation process to the 2nd EL images, obtain the second rotation EL images;Light and shade
Module is used to carry out light and shade operation to the 2nd EL images, obtains the second light and shade EL images;Horizontal transformation submodule be used for pair
The 2nd EL images carry out horizontal transformation operation, obtain the second horizontal transformation EL images;Fuzzy submodule is used for described the
Two EL images carry out fuzzy operation, obtain the second fuzzy EL images;Cut submodule be used for the 2nd EL images carry out with
Machine trimming operation, obtain the second random cropping EL images;Submodule is normalized to be used for the described second rotation EL images, described the
Two light and shade EL images, the second horizontal transformation EL images, the second fuzzy EL images and the second random cropping EL figures
Picture, it is normalized respectively according to preset algorithm, obtains described second effective EL images.
Specifically, the second acquisition module referred in above-described embodiment includes rotation submodule, light and shade submodule, horizontal change
Change submodule, fuzzy submodule, cut submodule and normalization submodule, wherein, rotation submodule can scheme to the 2nd EL
As carrying out rotation process, corresponding second rotation EL images are obtained;Light and shade submodule can carry out bright to the 2nd EL images
Dark operation, obtain corresponding second light and shade EL images;Horizontal transformation submodule can carry out horizontal change to the 2nd EL images
Operation is changed, obtains corresponding second horizontal transformation EL images;Fuzzy submodule can carry out fuzzy behaviour to the 2nd EL images
Make, obtain corresponding second fuzzy EL images;Random cropping operation can be carried out to the 2nd EL images by cutting submodule, be obtained
To corresponding second random cropping EL images.
Normalizing submodule can scheme to the 2nd EL images and its corresponding second rotation EL images, the second light and shade EL
Picture, the second horizontal transformation EL images, the second fuzzy EL images and the second random cropping EL images, are returned according to default algorithm
One change is handled, and obtains corresponding second effective EL images.For example Regularization can be carried out to each above-mentioned image, it is described
The formula that Regularization is used is:
Wherein, wherein x is the image pixel of before processing, and mean is image pixel average, and stddev is image pixel standard
Difference, y are the image pixel after processing.
Photovoltaic module defect detecting system provided in an embodiment of the present invention, by being rotated respectively to the 2nd EL images,
Light and shade, horizontal transformation, the operation of fuzzy and random cropping, and to the 2nd EL images and its corresponding second rotation EL figures
Picture, the second light and shade EL images, the second horizontal transformation EL images, the second fuzzy EL images and the second random cropping EL images are returned
One change is handled, and adds the quantity of sample, improves the precision and universality of photovoltaic defects detection.
Optionally, on the basis of above-described embodiment, the system includes:First acquisition module, detection module, fill mould
Block, filtration module and edge extracting module, wherein:
Fill module to be used for according to default Morphology Algorithm, processing is filled to the first EL images;Filter mould
Block is used for according to default filtering algorithm, and processing is filtered to the first EL images by the filling processing;Edge
Extraction module is used for according to default edge detection algorithm, and edge is carried out to the first EL images Jing Guo the filtering process
Extraction process.
First acquisition module and the detection module are described in detail in the above-described embodiments, and here is omitted.
Due to the first EL images of the first photovoltaic cells to be detected, it is generally from the EL images of a photovoltaic module, often
The EL images of multiple photovoltaic cells can be included in the EL images of individual photovoltaic module.So, it is necessary first to first from the light
In the EL images for lying prostrate component, the first EL images of first photovoltaic cells are extracted.Specifically, filling module can be by
According to default Morphology Algorithm such as closing operation of mathematical morphology, processing, filling described first are filled to the first EL images
Small cavity in EL display foreground colors, especially weaken the influence of two main gate lines on the first photovoltaic cells;Filter mould
Block can enter according to default filtering algorithm such as bilateral filtering algorithm to the first EL images by the filling processing
Row filtering process;Edge extracting module can be according to default edge detection algorithm such as Canny edge detection algorithms, to passing through
The first EL images of the filtering process carry out edge extracting processing, from the EL images of the photovoltaic module, extract institute
State the first EL images of the first battery unit.
Photovoltaic module defect detecting system provided in an embodiment of the present invention, pass through to the first battery unit to be detected
One EL images are filled processing, filtering process and edge extracting processing successively, are extracted from photovoltaic module image described
First EL images so that the photovoltaic module detecting system more science.
Fig. 3 is the structural representation of photovoltaic module defect detection equipment provided in an embodiment of the present invention, as shown in figure 3, institute
Stating equipment includes:Processor (processor) 31, memory (memory) 32 and bus 33, wherein:
The processor 31 and the memory 32 complete mutual communication by the bus 33;The processor 31
For calling the programmed instruction in the memory 32, to perform the method that above-mentioned each method embodiment is provided, such as including:
Obtain the first EL images of the first photovoltaic cells to be detected;The first EL images are input to the inspection of photovoltaic module defect
Survey in model, obtain the first defect classification of first photovoltaic cells.
The embodiment of the present invention discloses a kind of computer program product, and the computer program product is non-transient including being stored in
Computer program on computer-readable recording medium, the computer program include programmed instruction, when described program instructs quilt
When computer performs, computer is able to carry out the method that above-mentioned each method embodiment is provided, such as including:Obtain to be detected
First EL images of the first photovoltaic cells;The first EL images are input in photovoltaic module defects detection model, obtained
To the first defect classification of first photovoltaic cells.
The embodiment of the present invention provides a kind of non-transient computer readable storage medium storing program for executing, the non-transient computer readable storage
Medium storing computer instructs, and the computer instruction makes the computer perform the side that above-mentioned each method embodiment is provided
Method, such as including:Obtain the first EL images of the first photovoltaic cells to be detected;The first EL images are input to light
Lie prostrate in component defects detection model, obtain the first defect classification of first photovoltaic cells.
One of ordinary skill in the art will appreciate that:Realizing all or part of step of above method embodiment can pass through
Programmed instruction related hardware is completed, and foregoing program can be stored in a computer read/write memory medium, the program
Upon execution, the step of execution includes above method embodiment;And foregoing storage medium includes:ROM, RAM, magnetic disc or light
Disk etc. is various can be with the medium of store program codes.
The embodiments such as photovoltaic module defect detection equipment described above are only schematical, wherein described be used as is divided
Unit from part description can be or may not be it is physically separate, can be as the part that unit is shown or
It may not be physical location, you can with positioned at a place, or can also be distributed on multiple NEs.Can basis
It is actual to need to select some or all of module therein to realize the purpose of this embodiment scheme.Ordinary skill people
Member is not in the case where paying performing creative labour, you can to understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
Realized by the mode of software plus required general hardware platform, naturally it is also possible to pass through hardware.Based on such understanding, on
The part that technical scheme substantially in other words contributes to prior art is stated to embody in the form of software product, should
Computer software product can store in a computer-readable storage medium, such as ROM/RAM, magnetic disc, CD, including some fingers
Make to cause a computer equipment (can be personal computer, server, or network equipment etc.) to perform each implementation
Method described in some parts of example or embodiment.
Finally it should be noted that:Various embodiments above is rather than right only illustrating the technical scheme of embodiments of the invention
It is limited;Although embodiments of the invention are described in detail with reference to foregoing embodiments, the ordinary skill of this area
Personnel should be understood:It can still modify to the technical scheme described in foregoing embodiments, or to which part
Or all technical characteristic carries out equivalent substitution;And these modifications or replacement, do not make the essence disengaging of appropriate technical solution
The scope of each embodiment technical scheme of embodiments of the invention.
Claims (10)
- A kind of 1. photovoltaic component defect detection method, it is characterised in that including:Obtain the first EL images of the first photovoltaic cells to be detected;The first EL images are input in photovoltaic module defects detection model, obtain the of first photovoltaic cells One defect classification.
- 2. according to the method for claim 1, it is characterised in that methods described also includes the photovoltaic module defects detection mould Type establishes process, and the process includes:The 2nd EL images of the second photovoltaic cells are obtained, the 2nd EL images are handled according to predetermined registration operation, obtained To second effective EL images;Obtain the second defect classification of described second effective EL images;Described second effective EL images and the second defect classification are input in the convolutional neural networks, carry out depth Practise, obtain the photovoltaic module defects detection model.
- 3. according to the method for claim 2, it is characterised in that described that the 2nd EL images are carried out according to predetermined registration operation Processing, obtains second effective EL images, including:Rotation process is carried out to the 2nd EL images, obtains the second rotation EL images;Light and shade operation is carried out to the 2nd EL images, obtains the second light and shade EL images;Horizontal transformation operation is carried out to the 2nd EL images, obtains the second horizontal transformation EL images;Fuzzy operation is carried out to the 2nd EL images, obtains the second fuzzy EL images;Random cropping operation is carried out to the 2nd EL images, obtains the second random cropping EL images;To the 2nd EL images, the second rotation EL images, the second light and shade EL images, the second horizontal transformation EL Image, the second fuzzy EL images and the second random cropping EL images, are normalized place according to preset algorithm respectively Reason, obtains described second effective EL images.
- 4. according to the method for claim 1, it is characterised in that the first EL images are being input to photovoltaic module defect Before detection model, in addition to:According to default Morphology Algorithm, processing is filled to the first EL images;According to default filtering algorithm, processing is filtered to the first EL images by the filling processing;According to default edge detection algorithm, the first EL images Jing Guo the filtering process are carried out at edge extracting Reason.
- A kind of 5. photovoltaic module defect detecting system, it is characterised in that including:First acquisition module, for obtaining the first EL images of the first photovoltaic cells to be detected;Detection module, for the first EL images to be input into photovoltaic module defects detection model, obtain first photovoltaic First defect classification of battery unit.
- 6. system according to claim 5, it is characterised in that also include:Second acquisition module, for obtaining the 2nd EL images of the second photovoltaic cells, to the 2nd EL images according to pre- If operation is handled, second effective EL images are obtained;3rd acquisition module, for obtaining the second defect classification of described second effective EL images;Training module, for described second effective EL images and the second defect classification to be input into the convolutional neural networks In, deep learning is carried out, obtains the photovoltaic module defects detection model.
- 7. system according to claim 6, it is characterised in that second acquisition module includes:Submodule is rotated, for carrying out rotation process to the 2nd EL images, obtains the second rotation EL images;Light and shade submodule, for carrying out light and shade operation to the 2nd EL images, obtain the second light and shade EL images;Horizontal transformation submodule, for carrying out horizontal transformation operation to the 2nd EL images, obtain the second horizontal transformation EL figures Picture;Fuzzy submodule, for carrying out fuzzy operation to the 2nd EL images, obtain the second fuzzy EL images;Submodule is cut, for carrying out random cropping operation to the 2nd EL images, obtains the second random cropping EL images;Submodule is normalized, for the described second rotation EL images, the second light and shade EL images, second horizontal transformation EL images, the second fuzzy EL images and the second random cropping EL images, are normalized according to preset algorithm respectively Processing, obtains described second effective EL images.
- 8. system according to claim 5, it is characterised in that also include:Module is filled, for according to default Morphology Algorithm, processing to be filled to the first EL images;Filtration module, for according to default filtering algorithm, being filtered to the first EL images by the filling processing Ripple processing;Edge extracting module, for according to default edge detection algorithm, scheming to the first EL Jing Guo the filtering process As carrying out edge extracting processing.
- A kind of 9. photovoltaic module defect detection equipment, it is characterised in that including memory and processor, the processor and described Memory completes mutual communication by bus;The memory storage have can by the programmed instruction of the computing device, The processor calls described program instruction to be able to carry out the method as described in Claims 1-4 is any.
- 10. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that the computer program quilt The method as described in Claims 1-4 is any is realized during computing device.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710964119.4A CN107831173A (en) | 2017-10-17 | 2017-10-17 | Photovoltaic component defect detection method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710964119.4A CN107831173A (en) | 2017-10-17 | 2017-10-17 | Photovoltaic component defect detection method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107831173A true CN107831173A (en) | 2018-03-23 |
Family
ID=61648144
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710964119.4A Pending CN107831173A (en) | 2017-10-17 | 2017-10-17 | Photovoltaic component defect detection method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107831173A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108564577A (en) * | 2018-04-12 | 2018-09-21 | 重庆邮电大学 | Solar cell segment grid defect inspection method based on convolutional neural networks |
CN108562589A (en) * | 2018-03-30 | 2018-09-21 | 慧泉智能科技(苏州)有限公司 | A method of magnetic circuit material surface defect is detected |
CN108956653A (en) * | 2018-05-31 | 2018-12-07 | 广东正业科技股份有限公司 | A kind of quality of welding spot detection method, system, device and readable storage medium storing program for executing |
CN109239075A (en) * | 2018-08-27 | 2019-01-18 | 北京百度网讯科技有限公司 | Battery detection method and device |
CN110426395A (en) * | 2019-07-02 | 2019-11-08 | 广州大学 | A kind of solar energy EL cell silicon chip surface inspecting method and device |
CN110473806A (en) * | 2019-07-13 | 2019-11-19 | 河北工业大学 | The deep learning identification of photovoltaic cell sorting and control method and device |
CN110619343A (en) * | 2018-06-20 | 2019-12-27 | 亚摩信息技术股份有限公司 | Automatic defect classification method based on machine learning |
CN110736547A (en) * | 2019-10-17 | 2020-01-31 | 华能海南发电股份有限公司 | Photovoltaic panel fault intelligent diagnosis system based on infrared imaging technology |
CN113965163A (en) * | 2021-02-03 | 2022-01-21 | 苏州威华智能装备有限公司 | Battery piece defect detection method |
CN114581362A (en) * | 2021-07-22 | 2022-06-03 | 正泰集团研发中心(上海)有限公司 | Photovoltaic module defect detection method and device, electronic equipment and readable storage medium |
CN112396083B (en) * | 2019-08-19 | 2024-02-20 | 阿里巴巴集团控股有限公司 | Image recognition, model training and construction and detection methods, systems and equipment |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102590222A (en) * | 2012-03-06 | 2012-07-18 | 英利能源(中国)有限公司 | Photovoltaic component defect detection method and system |
CN107192759A (en) * | 2017-06-09 | 2017-09-22 | 湖南大学 | A kind of photovoltaic cell lossless detection method and system based on sensing optical heat radiation |
-
2017
- 2017-10-17 CN CN201710964119.4A patent/CN107831173A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102590222A (en) * | 2012-03-06 | 2012-07-18 | 英利能源(中国)有限公司 | Photovoltaic component defect detection method and system |
CN107192759A (en) * | 2017-06-09 | 2017-09-22 | 湖南大学 | A kind of photovoltaic cell lossless detection method and system based on sensing optical heat radiation |
Non-Patent Citations (2)
Title |
---|
李俊山 等: "《数字图像处理》", 30 April 2007, 清华大学出版社 * |
赵瑞: "基于Matlab的太阳电池片缺陷EL图像识别", 《万方数据库》 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108562589A (en) * | 2018-03-30 | 2018-09-21 | 慧泉智能科技(苏州)有限公司 | A method of magnetic circuit material surface defect is detected |
CN108562589B (en) * | 2018-03-30 | 2020-12-01 | 慧泉智能科技(苏州)有限公司 | Method for detecting surface defects of magnetic circuit material |
CN108564577A (en) * | 2018-04-12 | 2018-09-21 | 重庆邮电大学 | Solar cell segment grid defect inspection method based on convolutional neural networks |
CN108956653A (en) * | 2018-05-31 | 2018-12-07 | 广东正业科技股份有限公司 | A kind of quality of welding spot detection method, system, device and readable storage medium storing program for executing |
CN110619343A (en) * | 2018-06-20 | 2019-12-27 | 亚摩信息技术股份有限公司 | Automatic defect classification method based on machine learning |
CN110619343B (en) * | 2018-06-20 | 2023-06-06 | 亚摩信息技术(上海)有限公司 | Automatic defect classification method based on machine learning |
CN109239075A (en) * | 2018-08-27 | 2019-01-18 | 北京百度网讯科技有限公司 | Battery detection method and device |
CN110426395B (en) * | 2019-07-02 | 2022-02-11 | 广州大学 | Method and device for detecting surface of solar EL battery silicon wafer |
CN110426395A (en) * | 2019-07-02 | 2019-11-08 | 广州大学 | A kind of solar energy EL cell silicon chip surface inspecting method and device |
CN110473806A (en) * | 2019-07-13 | 2019-11-19 | 河北工业大学 | The deep learning identification of photovoltaic cell sorting and control method and device |
CN112396083B (en) * | 2019-08-19 | 2024-02-20 | 阿里巴巴集团控股有限公司 | Image recognition, model training and construction and detection methods, systems and equipment |
CN110736547A (en) * | 2019-10-17 | 2020-01-31 | 华能海南发电股份有限公司 | Photovoltaic panel fault intelligent diagnosis system based on infrared imaging technology |
CN113965163A (en) * | 2021-02-03 | 2022-01-21 | 苏州威华智能装备有限公司 | Battery piece defect detection method |
CN114581362A (en) * | 2021-07-22 | 2022-06-03 | 正泰集团研发中心(上海)有限公司 | Photovoltaic module defect detection method and device, electronic equipment and readable storage medium |
CN114581362B (en) * | 2021-07-22 | 2023-11-07 | 正泰集团研发中心(上海)有限公司 | Photovoltaic module defect detection method and device, electronic equipment and readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107831173A (en) | Photovoltaic component defect detection method and system | |
CN106203331B (en) | A kind of crowd density evaluation method based on convolutional neural networks | |
CN104850836B (en) | Insect automatic distinguishing method for image based on depth convolutional neural networks | |
CN106951870B (en) | Intelligent detection and early warning method for active visual attention of significant events of surveillance video | |
CN108510451B (en) | Method for reconstructing license plate based on double-layer convolutional neural network | |
CN109035260A (en) | A kind of sky areas dividing method, device and convolutional neural networks | |
CN108986050A (en) | A kind of image and video enhancement method based on multiple-limb convolutional neural networks | |
CN113850242B (en) | Storage abnormal target detection method and system based on deep learning algorithm | |
CN107506695A (en) | Video monitoring equipment failure automatic detection method | |
CN107844797A (en) | A kind of method of the milking sow posture automatic identification based on depth image | |
CN107545249A (en) | A kind of population ages' recognition methods and device | |
CN109740721B (en) | Wheat ear counting method and device | |
CN107784305A (en) | Facilities vegetable disease recognition method and device based on convolutional neural networks | |
CN111242127A (en) | Vehicle detection method with granularity level multi-scale characteristics based on asymmetric convolution | |
CN113469074B (en) | Remote sensing image change detection method and system based on twin attention fusion network | |
CN109949224A (en) | A kind of method and device of the connection grade super-resolution rebuilding based on deep learning | |
CN109670509A (en) | Winter wheat seedling stage growing way parameter evaluation method and system based on convolutional neural networks | |
CN111553227A (en) | Lightweight face detection method based on task guidance | |
CN111639668A (en) | Crowd density detection method based on deep learning | |
CN112070720A (en) | Transformer substation equipment defect identification method based on deep learning model | |
CN116863353A (en) | Electric power tower inclination degree detection method based on rotating target detection network | |
CN117218329B (en) | Wellhead valve detection method and device, storage medium and electronic equipment | |
CN112711985B (en) | Fruit identification method and device based on improved SOLO network and robot | |
CN113936299A (en) | Method for detecting dangerous area in construction site | |
CN105095923A (en) | Image processing method and device |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180323 |