CN117237590A - Photovoltaic module hot spot identification method and system based on image identification - Google Patents
Photovoltaic module hot spot identification method and system based on image identification Download PDFInfo
- Publication number
- CN117237590A CN117237590A CN202311492717.8A CN202311492717A CN117237590A CN 117237590 A CN117237590 A CN 117237590A CN 202311492717 A CN202311492717 A CN 202311492717A CN 117237590 A CN117237590 A CN 117237590A
- Authority
- CN
- China
- Prior art keywords
- hot spot
- photovoltaic panel
- image
- data set
- identification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000001514 detection method Methods 0.000 claims abstract description 22
- 238000012549 training Methods 0.000 claims abstract description 14
- 238000005286 illumination Methods 0.000 claims abstract description 9
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 239000000523 sample Substances 0.000 claims description 20
- 238000012545 processing Methods 0.000 claims description 10
- 238000011176 pooling Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000002372 labelling Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 5
- 230000014509 gene expression Effects 0.000 claims description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 241001492331 Pelargonium leaf curl virus Species 0.000 claims description 3
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000010248 power generation Methods 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- NJPPVKZQTLUDBO-UHFFFAOYSA-N novaluron Chemical compound C1=C(Cl)C(OC(F)(F)C(OC(F)(F)F)F)=CC=C1NC(=O)NC(=O)C1=C(F)C=CC=C1F NJPPVKZQTLUDBO-UHFFFAOYSA-N 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000712 assembly Effects 0.000 description 1
- 238000000429 assembly Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- 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
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/50—Photovoltaic [PV] energy
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of photovoltaic hot spot detection, in particular to a method and a system for identifying hot spots of a photovoltaic module based on image identification, comprising the following steps: acquiring a photovoltaic panel image information data set and a target information data set by using an identification device, and preprocessing, wherein the photovoltaic panel image information data set at least comprises image data of photovoltaic panels with different illumination conditions, different angles and different sizes; the target information data set at least comprises temperature data and current data of the photovoltaic panel; constructing a hot spot recognition model, and training the hot spot recognition model by utilizing the preprocessed photovoltaic panel image information data set and the preprocessed target information data set; and identifying the real-time image of the photovoltaic panel by using the hot spot identification model to obtain a hot spot detection result of the photovoltaic panel. The method is used for realizing hot spot identification on the real-time image of the photovoltaic panel, manual intervention is not needed, automatic detection is realized, and labor cost is reduced.
Description
Technical Field
The invention relates to the technical field of photovoltaic hot spot detection, in particular to a method and a system for identifying hot spots of a photovoltaic module based on image identification.
Background
With the deep global energy revolution, the renewable energy market continues to expand, and photovoltaic power generation becomes an important component of renewable energy, and has a large development space; at present, with the increase of construction of photovoltaic equipment, the problems of faults of the photovoltaic equipment are increased, wherein the hot spot defect of the photovoltaic module is a main cause of faults of a photovoltaic power generation system, the real-time monitoring and maintenance requirements of the photovoltaic module cannot be met by manpower inspection, and the problems of misjudgment, building and low detection efficiency exist in manual detection; on the other hand, the existing commonly used photovoltaic module hot spot detection device is usually in a handheld type and a pedestal type, in the detection process, the handheld type equipment is easily affected by external influence to cause shaking of an infrared probe to influence the detection precision, the pedestal type equipment has the problems that the probe is inconvenient to move and the effective detection range is limited, and therefore, the photovoltaic module hot spot identification method and system based on image identification are provided.
Disclosure of Invention
The invention aims to provide a method and a system for identifying hot spots of a photovoltaic module based on image identification, which are used for identifying the hot spots of a real-time image of a photovoltaic panel, do not need manual intervention, realize automatic detection and reduce labor cost.
The technical scheme of the first aspect of the invention provides a photovoltaic module hot spot identification method based on image identification, which comprises the following steps:
acquiring a photovoltaic panel image information data set and a target information data set by using an identification device, and preprocessing, wherein the photovoltaic panel image information data set at least comprises image data of photovoltaic panels with different illumination conditions, different angles and different sizes; the target information data set at least comprises temperature data and current data of the photovoltaic panel;
constructing a hot spot recognition model, and training the hot spot recognition model by utilizing the preprocessed photovoltaic panel image information data set and the preprocessed target information data set;
and identifying the real-time image of the photovoltaic panel by using the hot spot identification model to obtain a hot spot detection result of the photovoltaic panel.
Further, the preprocessing process of the photovoltaic panel image information data set and the target information data set comprises the following steps:
denoising and normalizing the temperature data and the current data;
labeling the photovoltaic panel image information data set, labeling the serial number information of the photovoltaic panel label and the position and type information of the hot spots, and converting the serial number information and the type information into a VOC format;
fusing the preprocessed temperature data and current data with the photovoltaic panel image data, wherein the temperature data and the current data are used as channel information of the photovoltaic panel image data;
processing an image in the photovoltaic panel image information dataset based on gamma conversion;
and updating the serial number information, the hot spot position, the category information, the temperature data and the current data of the photovoltaic panel label according to the image after the gamma conversion processing, and dividing the image into a training set and a data set according to a preset proportion.
Further, constructing the hot spot recognition model specifically includes:
based on a convolutional neural network, inputting a preprocessed photovoltaic panel image information data set and a preprocessed target information data set;
based on a CBAM attention mechanism module, outputting two feature graphs through global maximum pooling and global average pooling, and carrying out feature fusion to obtain fused features;
and inputting the fused characteristics into a hot spot classifier, judging whether hot spots exist or not, and outputting category information of the hot spots.
Further, constructing the hot spot recognition model further includes:
respectively calculating predicted loss values of hot spot positions, hot spot types, temperature data and current data, and weighting according to preset weights to obtain a target loss function;
the Adam optimizer is based on minimizing the target loss function and training the hot spot recognition model.
Further, the expression of the objective loss function is:
in the above-mentioned method, the step of,representing the target loss function->、/>、/>、/>Weights respectively representing hot spot position prediction, hot spot category prediction, temperature data prediction and current data prediction; />、/>、/>、/>Loss functions respectively representing hot spot position prediction, hot spot category prediction, temperature data prediction and current data prediction; />、/>、/>、、/>、/>、/>、/>The real position of the hot spot, the predicted position of the hot spot, the real category of the hot spot, the predicted category of the hot spot, the real temperature data, the predicted temperature data, the real current data and the predicted current data are respectively represented.
Further, the method for identifying the real-time image of the photovoltaic panel by using the hot spot identification model further comprises the following steps before the specific hot spot detection result of the photovoltaic panel is obtained:
predicting the real-time image of the photovoltaic panel by using the trained hot spot recognition model, and comparing the real-time image with a real tag frame;
and calculating average accuracy, precision and recall rate to optimize the hot spot identification model.
Further, the obtaining of the photovoltaic panel image information dataset and the target information dataset with the identification means further comprises screening the photovoltaic panel image information dataset: and setting a brightness threshold value interval, an angle threshold value interval, a size threshold value interval, a temperature threshold value interval and a current threshold value interval, and eliminating photovoltaic panel images outside the brightness threshold value interval, the angle threshold value interval, the size threshold value interval, the temperature threshold value interval and the current threshold value interval.
The technical scheme of the second aspect of the invention provides a photovoltaic module hot spot recognition system based on image recognition, which comprises the following components:
the identification device is configured to acquire a photovoltaic panel image information data set, wherein the photovoltaic panel image information data set at least comprises image data of photovoltaic panels with different illumination conditions, different angles and different sizes;
the data acquisition device is configured to acquire a target information data set, wherein the target information data set at least comprises temperature data and current data of the photovoltaic panel;
the data processing module is configured to preprocess the photovoltaic panel image information data set and the target information data set;
a hot spot identification module: the method comprises the steps of configuring a hot spot identification model by utilizing a preprocessed photovoltaic panel image information data set and a preprocessed target information data set;
and an identification result module: the method comprises the steps of identifying a real-time image of a photovoltaic panel by using a trained hot spot identification model, so as to obtain a hot spot detection result of the photovoltaic panel;
and the control module is configured to be electrically connected with the identification device, and is used for controlling the identification device to acquire the image information of the photovoltaic panel and adjusting the position of the identification device through the PLCV.
Further, recognition device includes slip table, logical groove, slider, swivel ball, axostylus axostyle, power module and probe, the inside logical groove that is equipped with of slip table, slider sliding connection is in logical inslot, swivel ball inlays and locates in the slider and rotate with between the slider to be connected, the axostylus axostyle runs through swivel ball, the probe is located the axostylus axostyle bottom, power module locates in the axostylus axostyle and is connected with the electricity between the probe.
Further, threaded through holes are respectively formed in the inner side and the outer side of the sliding table, threaded rods are inserted into the threaded through holes, the two ends of each threaded rod extend to the outer portion of the sliding table, and nuts are sleeved on the outer sides of the threaded rods.
The technical scheme of the embodiment of the invention has at least the following advantages and beneficial effects:
according to the invention, the image information data set and the target information data set of the photovoltaic panel are obtained through the identification device and preprocessed, and the hot spot identification of the image data of the photovoltaic panel and the temperature data and the current data of the photovoltaic panel under different illumination conditions, different angle conditions and different sizes is realized through collecting the image data of the photovoltaic panel and the temperature data and the current data of the photovoltaic panel under different illumination conditions, different angle conditions and different sizes, so that the identification precision can be effectively improved through diversified data, and the method has wider applicability; by constructing the hot spot recognition model, the real-time image of the photovoltaic panel is recognized by utilizing the hot spot recognition model, the hot spot detection result of the photovoltaic panel is obtained, the real-time image of the photovoltaic panel is recognized, manual intervention is not needed, automatic detection is realized, the labor cost is reduced, and the safety and the stable operation of a photovoltaic power generation system are guaranteed.
Drawings
Fig. 1 is a schematic flow chart of a method for identifying hot spots of a photovoltaic module based on image identification according to an embodiment of the present invention;
fig. 2 is a schematic front view of an identification device according to an embodiment of the present invention;
fig. 3 is a schematic top view of an identification device according to an embodiment of the present invention;
icon: the device comprises a sliding table 1, a through groove 2, a sliding block 3, a rotary ball 4, a control module 5, a power module 6, a shaft lever 7, a probe 8, a threaded rod 9 and a threaded through hole 10.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
It will be appreciated that "system," "apparatus," "unit" and/or "module" as used herein is one method for distinguishing between different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.
As used in this specification and the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
A flowchart is used in this specification to describe the operations performed by the system according to embodiments of the present specification. It should be appreciated that the preceding or following operations are not necessarily performed in order precisely. Rather, the steps may be processed in reverse order or simultaneously. Also, other operations may be added to or removed from these processes.
Referring to fig. 1, the technical solution of the first aspect of the present invention provides a method for identifying hot spots of a photovoltaic module based on image identification, which includes the following steps:
step S100: acquiring a photovoltaic panel image information data set and a target information data set by using an identification device, and preprocessing, wherein the photovoltaic panel image information data set at least comprises image data of photovoltaic panels with different illumination conditions, different angles and different sizes; the target information data set at least comprises temperature data and current data of the photovoltaic panel; specifically, temperature data and current data of the photovoltaic panel can be acquired in real time by installing a temperature sensor and a current sensor on the photovoltaic panel;
in step S100, the preprocessing process of the photovoltaic panel image information dataset and the target information dataset includes:
step S110: denoising and normalizing the temperature data and the current data;
step S120: labeling the photovoltaic panel image information data set, labeling the serial number information of the photovoltaic panel label and the position and type information of the hot spots, and converting the serial number information and the type information into a VOC format;
step S130: fusing the preprocessed temperature data and current data with the photovoltaic panel image data, wherein the temperature data and the current data are used as channel information of the photovoltaic panel image data; by introducing temperature data and current data of the photovoltaic panel and performing fusion and joint processing on the temperature data and the current data of the photovoltaic panel and the image data of the photovoltaic panel, various data information can be fully utilized, and the accuracy and the reliability of hot spot identification are improved; meanwhile, by combining real-time identification and decision, the real-time monitoring and maintenance of the hot spots of the photovoltaic panel can be realized, and the efficiency and reliability of the photovoltaic system are improved;
step S140: processing an image in the photovoltaic panel image information dataset based on gamma conversion; the contrast and the definition of the image are enhanced by using gamma transformation, so that the subsequent feature extraction and recognition effects can be improved;
step S150: and updating the serial number information, the hot spot position, the category information, the temperature data and the current data of the photovoltaic panel label according to the image after the gamma conversion processing, and dividing the image into a training set and a data set according to a preset proportion.
In step S100, acquiring the photovoltaic panel image information dataset and the target information dataset with the identification means further comprises screening the photovoltaic panel image information dataset: setting a brightness threshold value interval, an angle threshold value interval, a size threshold value interval, a temperature threshold value interval and a current threshold value interval, and eliminating photovoltaic panel images outside the brightness threshold value interval, the angle threshold value interval, the size threshold value interval, the temperature threshold value interval and the current threshold value interval; by setting the threshold interval and eliminating the photovoltaic panel images which do not meet the requirements, noise and interference can be reduced, and the accuracy and reliability of hot spot identification can be improved.
Step S200: constructing a hot spot recognition model, and training the hot spot recognition model by utilizing the preprocessed photovoltaic panel image information data set and the preprocessed target information data set;
in step S210, the constructing a hot spot recognition model specifically includes:
step S220: based on a convolutional neural network, inputting a preprocessed photovoltaic panel image information data set and a preprocessed target information data set;
step S230: based on a CBAM attention mechanism module, outputting two feature graphs through global maximum pooling and global average pooling, and carrying out feature fusion to obtain fused features; specifically, the CBAM attention mechanism module is configured to enhance attention to a hot spot area, and the CBAM module may obtain two feature maps of 1×1×c through global max pooling and global average pooling operations, and then send the two feature maps into a two-layer neural network (MLP) to perform feature fusion, and obtain fused features through element-wise addition and operation; performing sigmoid activation operation on the fused features, and then performing 7×7 convolution operation to reduce the channel number to 1 channel, so as to obtain the finally generated features; the CBAM module can adaptively weight important features in the image, so that the expression capacity and generalization capacity of the model are improved;
step S240: inputting the fused characteristics into a hot spot classifier, judging whether hot spots exist or not, and outputting category information of the hot spots;
in step S200, constructing the hot spot recognition model further includes:
step S241: respectively calculating predicted loss values of hot spot positions, hot spot types, temperature data and current data, and weighting according to preset weights to obtain a target loss function;
the expression of the objective loss function is:
in the above-mentioned method, the step of,representing the target loss function->、/>、/>、/>Weights respectively representing hot spot position prediction, hot spot category prediction, temperature data prediction and current data prediction; />、/>、/>、/>Loss functions respectively representing hot spot position prediction, hot spot category prediction, temperature data prediction and current data prediction; />、/>、/>、、/>、/>、/>、/>Respectively representing a hot spot real position, a hot spot predicted position, a hot spot real type, a hot spot predicted type, real temperature data, predicted temperature data, real current data and predicted current data;
step S242: based on an Adam optimizer, minimizing a target loss function and training a hot spot recognition model, specifically using a training data set to train the model, in each training iteration, sending input data into the model, calculating the output of the model, and then comparing the output with a real label to calculate the value of the loss function; then, an optimizer is used for minimizing the loss function, parameters of the model are updated according to the gradient, the process is repeated until the preset training round number is reached or the training stopping condition is reached, the Adam optimizer is used for minimizing the multi-task target loss function, the parameters of the model can be effectively updated, the convergence speed and stability of the model are improved, and better performance of the model on multiple tasks such as position prediction, category prediction, temperature data prediction and current data prediction is achieved.
Step S300: and identifying the real-time image of the photovoltaic panel by using the hot spot identification model to obtain a hot spot detection result of the photovoltaic panel.
Preferably, the method for identifying the real-time image of the photovoltaic panel by using the hot spot identification model further comprises the following steps before the specific hot spot detection result of the photovoltaic panel is obtained:
step S250: predicting the real-time image of the photovoltaic panel by using the trained hot spot recognition model, and comparing the real-time image with a real tag frame;
step S260: calculating average accuracy, precision and recall rate to optimize the hot spot recognition model; in the identification of the hot spots of the photovoltaic panel, the accuracy and the recall rate can be calculated by presetting a plurality of cross ratio thresholds, and the average value is taken as the average accuracy; for example, the cross ratio threshold is set to 0.5,0.6,0.7, the accuracy and recall at each threshold are calculated, and then the average is taken as the average accuracy; in the photovoltaic panel hot spot recognition, the intersection ratio of the predicted hot spot label frame and the real label frame can be calculated, then whether the hot spot label frame is a positive sample is judged according to a set threshold value, and the accuracy can be calculated by counting the proportion that the intersection ratio of the hot spot label frame predicted as the positive sample and the real label frame is larger than the threshold value; in the photovoltaic panel hot spot identification, the recall rate can be calculated by counting the proportion that the intersection ratio of the hot spot label frame which is true and is a positive sample and the predicted label frame is larger than a threshold value; the hot spot recognition result of the real-time data of the photovoltaic panel can be evaluated and analyzed by calculating the average accuracy, precision and recall, and the evaluation indexes can help judge the performance of the model and guide further improvement and optimization.
The technical scheme of the second aspect of the invention provides a photovoltaic module hot spot recognition system based on image recognition, which comprises the following components:
the identification device is configured to acquire a photovoltaic panel image information data set, wherein the photovoltaic panel image information data set at least comprises image data of photovoltaic panels with different illumination conditions, different angles and different sizes;
the data acquisition device is configured to acquire a target information data set, wherein the target information data set at least comprises temperature data and current data of the photovoltaic panel;
the data processing module is configured to preprocess the photovoltaic panel image information data set and the target information data set;
a hot spot identification module: the method comprises the steps of configuring a hot spot identification model by utilizing a preprocessed photovoltaic panel image information data set and a preprocessed target information data set;
and an identification result module: the method comprises the steps of identifying a real-time image of a photovoltaic panel by using a trained hot spot identification model, so as to obtain a hot spot detection result of the photovoltaic panel;
and the control module is configured to be electrically connected with the identification device, and is used for controlling the identification device to acquire the image information of the photovoltaic panel and adjusting the position of the identification device through the PLCV.
Referring to fig. 2 and 3, in this embodiment, the identifying device includes a sliding table 1, a through slot 2, a sliding block 3, a rotating ball 4, a shaft rod 7, a power module 6 and a probe 8, the sliding table 1 is internally provided with the through slot 2, the sliding block 3 is slidably connected in the through slot 2, the rotating ball 4 is embedded in the sliding block 3 and is rotationally connected with the sliding block 3, the shaft rod 7 penetrates through the rotating ball 4, the probe 8 is arranged at the bottom of the shaft rod 7, the power module 6 is arranged in the shaft rod 7 and is electrically connected with the probe 8, threaded through holes 10 are respectively formed in two inner sides of the sliding table 1, threaded rods 9 are inserted in the threaded through holes 10, two ends of each threaded rod 9 extend to the outside of the sliding table 1, and nuts are sleeved outside the threaded rods 9; during practical application, the probe 8 is used for acquiring image information of a photovoltaic panel, the horizontal position of the probe 8 can be adjusted through the sliding block 3, the relative position of the rotating ball 4 and the sliding block 3 can be changed through the rotating ball 4, the angle position of the probe 8 can be adjusted, the control module 5 is further configured on the shaft lever 7, the control module 5 can adopt the existing PLC controller, the switch and the position adjustment of the probe 8 can be realized by combining with a motor, the accurate positioning of the probe 8 is realized, the accuracy of the photovoltaic panel image is ensured, the photovoltaic panel with different sizes and shapes is adapted, and the adaptability of identification is improved.
Claims (10)
1. The method for identifying the hot spots of the photovoltaic module based on the image identification is characterized by comprising the following steps of:
acquiring a photovoltaic panel image information data set and a target information data set by using an identification device, and preprocessing, wherein the photovoltaic panel image information data set at least comprises image data of photovoltaic panels with different illumination conditions, different angles and different sizes; the target information data set at least comprises temperature data and current data of the photovoltaic panel;
constructing a hot spot recognition model, and training the hot spot recognition model by utilizing the preprocessed photovoltaic panel image information data set and the preprocessed target information data set;
and identifying the real-time image of the photovoltaic panel by using the hot spot identification model to obtain a hot spot detection result of the photovoltaic panel.
2. The method for identifying hot spots of a photovoltaic module based on image identification according to claim 1, wherein the preprocessing process of the photovoltaic panel image information data set and the target information data set comprises the following steps:
denoising and normalizing the temperature data and the current data;
labeling the photovoltaic panel image information data set, labeling the serial number information of the photovoltaic panel label and the position and type information of the hot spots, and converting the serial number information and the type information into a VOC format;
fusing the preprocessed temperature data and current data with the photovoltaic panel image data, wherein the temperature data and the current data are used as channel information of the photovoltaic panel image data;
processing an image in the photovoltaic panel image information dataset based on gamma conversion;
and updating the serial number information, the hot spot position, the category information, the temperature data and the current data of the photovoltaic panel label according to the image after the gamma conversion processing, and dividing the image into a training set and a data set according to a preset proportion.
3. The method for identifying the hot spots of the photovoltaic module based on the image identification according to claim 1, wherein the constructing the hot spot identification model specifically comprises the following steps:
based on a convolutional neural network, inputting a preprocessed photovoltaic panel image information data set and a preprocessed target information data set;
based on a CBAM attention mechanism module, outputting two feature graphs through global maximum pooling and global average pooling, and carrying out feature fusion to obtain fused features;
and inputting the fused characteristics into a hot spot classifier, judging whether hot spots exist or not, and outputting category information of the hot spots.
4. The method for identifying hot spots of a photovoltaic module based on image identification according to claim 3, wherein constructing a hot spot identification model further comprises:
respectively calculating predicted loss values of hot spot positions, hot spot types, temperature data and current data, and weighting according to preset weights to obtain a target loss function;
the Adam optimizer is based on minimizing the target loss function and training the hot spot recognition model.
5. The method for identifying hot spots of a photovoltaic module based on image identification according to claim 4, wherein the expression of the objective loss function is:
in the above-mentioned method, the step of,representing the target loss function->、/>、/>、/>Weights respectively representing hot spot position prediction, hot spot category prediction, temperature data prediction and current data prediction; />、/>、/>、/>Loss functions respectively representing hot spot position prediction, hot spot category prediction, temperature data prediction and current data prediction; />、/>、/>、、/>、/>、/>、/>The real position of the hot spot, the predicted position of the hot spot, the real category of the hot spot, the predicted category of the hot spot, the real temperature data, the predicted temperature data, the real current data and the predicted current data are respectively represented.
6. The method for identifying hot spots of a photovoltaic module based on image identification according to claim 5, wherein the method for identifying the real-time image of the photovoltaic panel by using the hot spot identification model further comprises the following steps before the specific hot spot detection result of the photovoltaic panel is obtained:
predicting the real-time image of the photovoltaic panel by using the trained hot spot recognition model, and comparing the real-time image with a real tag frame;
and calculating average accuracy, precision and recall rate to optimize the hot spot identification model.
7. The method of any one of claims 1 to 6, wherein obtaining the photovoltaic panel image information dataset and the target information dataset with the identification device further comprises screening the photovoltaic panel image information dataset: and setting a brightness threshold value interval, an angle threshold value interval, a size threshold value interval, a temperature threshold value interval and a current threshold value interval, and eliminating photovoltaic panel images outside the brightness threshold value interval, the angle threshold value interval, the size threshold value interval, the temperature threshold value interval and the current threshold value interval.
8. Photovoltaic module hot spot recognition system based on image recognition, its characterized in that includes:
the identification device is configured to acquire a photovoltaic panel image information data set, wherein the photovoltaic panel image information data set at least comprises image data of photovoltaic panels with different illumination conditions, different angles and different sizes;
the data acquisition device is configured to acquire a target information data set, wherein the target information data set at least comprises temperature data and current data of the photovoltaic panel;
the data processing module is configured to preprocess the photovoltaic panel image information data set and the target information data set;
a hot spot identification module: the method comprises the steps of configuring a hot spot identification model by utilizing a preprocessed photovoltaic panel image information data set and a preprocessed target information data set;
and an identification result module: the method comprises the steps of identifying a real-time image of a photovoltaic panel by using a trained hot spot identification model, so as to obtain a hot spot detection result of the photovoltaic panel;
and the control module is configured to be electrically connected with the identification device, and is used for controlling the identification device to acquire the image information of the photovoltaic panel and adjusting the position of the identification device through the PLCV.
9. The system of claim 8, wherein the identifying device comprises a sliding table, a through groove, a sliding block, a rotating ball, a shaft lever, a power module and a probe, the sliding table is internally provided with the through groove, the sliding block is slidably connected in the through groove, the rotating ball is embedded in the sliding block and is rotationally connected with the sliding block, the shaft lever penetrates through the rotating ball, the probe is arranged at the bottom of the shaft lever, and the power module is arranged in the shaft lever and is electrically connected with the probe.
10. The photovoltaic module hot spot recognition system based on image recognition according to claim 9, wherein threaded through holes are respectively formed in the inner side and the outer side of the sliding table, threaded rods are inserted in the threaded through holes, the two ends of each threaded rod extend to the outer portion of the sliding table, and nuts are sleeved on the outer sides of the threaded rods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311492717.8A CN117237590B (en) | 2023-11-10 | 2023-11-10 | Photovoltaic module hot spot identification method and system based on image identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311492717.8A CN117237590B (en) | 2023-11-10 | 2023-11-10 | Photovoltaic module hot spot identification method and system based on image identification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117237590A true CN117237590A (en) | 2023-12-15 |
| CN117237590B CN117237590B (en) | 2024-04-02 |
Family
ID=89088381
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311492717.8A Active CN117237590B (en) | 2023-11-10 | 2023-11-10 | Photovoltaic module hot spot identification method and system based on image identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117237590B (en) |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017184201A1 (en) * | 2016-04-22 | 2017-10-26 | Entropia Llc | Methods for thermal breast cancer detection |
| CN210327503U (en) * | 2019-09-17 | 2020-04-14 | 无锡润俊精密机械有限公司 | Hot spot detection device for photovoltaic module |
| US20200166593A1 (en) * | 2018-11-28 | 2020-05-28 | Shahar RINOTT | Systems and methods for correcting measurement artifacts in mr thermometry |
| CN112163018A (en) * | 2020-09-27 | 2021-01-01 | 国家电网有限公司 | Method, device and system for determining life cycle of photovoltaic module |
| CN112465738A (en) * | 2020-12-21 | 2021-03-09 | 国网山东省电力公司电力科学研究院 | Photovoltaic power station online operation and maintenance method and system based on infrared and visible light images |
| CN113076816A (en) * | 2021-03-17 | 2021-07-06 | 上海电力大学 | Solar photovoltaic module hot spot identification method based on infrared and visible light images |
| CN113378459A (en) * | 2021-06-02 | 2021-09-10 | 兰州交通大学 | Photovoltaic power station ultra-short-term power prediction method based on satellite and internet of things information |
| CN114299033A (en) * | 2021-12-29 | 2022-04-08 | 中国科学技术大学 | YOLOv 5-based photovoltaic panel infrared image hot spot detection method and system |
| CN114419360A (en) * | 2021-11-23 | 2022-04-29 | 东北电力大学 | Photovoltaic panel infrared thermal image classification and hot spot positioning method |
| CN114627044A (en) * | 2021-11-24 | 2022-06-14 | 上海电力大学 | Solar photovoltaic module hot spot detection method based on deep learning |
| EP4075112A1 (en) * | 2021-04-14 | 2022-10-19 | ABB Schweiz AG | A fault detection system |
| CN115546670A (en) * | 2022-10-21 | 2022-12-30 | 国网山东省电力公司威海供电公司 | Photovoltaic panel infrared image hot spot detection method based on improved BETR model |
| CN115712873A (en) * | 2022-11-18 | 2023-02-24 | 国网江苏省电力有限公司徐州供电分公司 | Photovoltaic grid-connected operation abnormity detection system and method based on data analysis and infrared image information fusion |
| US20230064675A1 (en) * | 2021-08-31 | 2023-03-02 | Digital Path, Inc. | Pt/pt-z camera command, control & visualization system and method |
| CN116317943A (en) * | 2023-03-10 | 2023-06-23 | 国网浙江省电力有限公司电力科学研究院 | Photovoltaic array hot spot detection method |
| CN116503354A (en) * | 2023-04-27 | 2023-07-28 | 清华大学 | Method and device for detecting and evaluating hot spots of photovoltaic cells based on multi-mode fusion |
-
2023
- 2023-11-10 CN CN202311492717.8A patent/CN117237590B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017184201A1 (en) * | 2016-04-22 | 2017-10-26 | Entropia Llc | Methods for thermal breast cancer detection |
| US20200166593A1 (en) * | 2018-11-28 | 2020-05-28 | Shahar RINOTT | Systems and methods for correcting measurement artifacts in mr thermometry |
| CN210327503U (en) * | 2019-09-17 | 2020-04-14 | 无锡润俊精密机械有限公司 | Hot spot detection device for photovoltaic module |
| CN112163018A (en) * | 2020-09-27 | 2021-01-01 | 国家电网有限公司 | Method, device and system for determining life cycle of photovoltaic module |
| WO2022063282A1 (en) * | 2020-09-27 | 2022-03-31 | 国家电网有限公司 | Method and device for determining life cycle of photovoltaic module |
| CN112465738A (en) * | 2020-12-21 | 2021-03-09 | 国网山东省电力公司电力科学研究院 | Photovoltaic power station online operation and maintenance method and system based on infrared and visible light images |
| CN113076816A (en) * | 2021-03-17 | 2021-07-06 | 上海电力大学 | Solar photovoltaic module hot spot identification method based on infrared and visible light images |
| EP4075112A1 (en) * | 2021-04-14 | 2022-10-19 | ABB Schweiz AG | A fault detection system |
| CN113378459A (en) * | 2021-06-02 | 2021-09-10 | 兰州交通大学 | Photovoltaic power station ultra-short-term power prediction method based on satellite and internet of things information |
| US20230064675A1 (en) * | 2021-08-31 | 2023-03-02 | Digital Path, Inc. | Pt/pt-z camera command, control & visualization system and method |
| CN114419360A (en) * | 2021-11-23 | 2022-04-29 | 东北电力大学 | Photovoltaic panel infrared thermal image classification and hot spot positioning method |
| CN114627044A (en) * | 2021-11-24 | 2022-06-14 | 上海电力大学 | Solar photovoltaic module hot spot detection method based on deep learning |
| CN114299033A (en) * | 2021-12-29 | 2022-04-08 | 中国科学技术大学 | YOLOv 5-based photovoltaic panel infrared image hot spot detection method and system |
| CN115546670A (en) * | 2022-10-21 | 2022-12-30 | 国网山东省电力公司威海供电公司 | Photovoltaic panel infrared image hot spot detection method based on improved BETR model |
| CN115712873A (en) * | 2022-11-18 | 2023-02-24 | 国网江苏省电力有限公司徐州供电分公司 | Photovoltaic grid-connected operation abnormity detection system and method based on data analysis and infrared image information fusion |
| CN116317943A (en) * | 2023-03-10 | 2023-06-23 | 国网浙江省电力有限公司电力科学研究院 | Photovoltaic array hot spot detection method |
| CN116503354A (en) * | 2023-04-27 | 2023-07-28 | 清华大学 | Method and device for detecting and evaluating hot spots of photovoltaic cells based on multi-mode fusion |
Non-Patent Citations (2)
| Title |
|---|
| GOUDELIS, GEORGIOS等: "A review of models for photovoltaic crack and hotspot prediction", 《ENERGIES》, vol. 15, no. 12, 12 June 2022 (2022-06-12), pages 1 - 5 * |
| 吴琼等: "光伏组件智能化故障检测综述", 《上海电力大学学报》, vol. 39, no. 3, 30 June 2023 (2023-06-30), pages 252 - 257 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117237590B (en) | 2024-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110598736B (en) | Power equipment infrared image fault positioning, identifying and predicting method | |
| CN116625438B (en) | Gas pipe network safety on-line monitoring system and method thereof | |
| CN107767374B (en) | Intelligent diagnosis method for local overheating of inner conductor of GIS basin-type insulator | |
| CN113324864B (en) | Pantograph carbon slide plate abrasion detection method based on deep learning target detection | |
| CN112697798B (en) | Infrared image-oriented diagnosis method and device for current-induced thermal defects of power transformation equipment | |
| CN116690613B (en) | Control method and system of photovoltaic intelligent cleaning robot | |
| CN113205039A (en) | Power equipment fault image identification and disaster investigation system and method based on multiple DCNNs | |
| CN115311585A (en) | A discernment detecting system for electric wire netting overhead line foreign matter | |
| CN117237590B (en) | Photovoltaic module hot spot identification method and system based on image identification | |
| CN117315380B (en) | Deep learning-based pneumonia CT image classification method and system | |
| CN113819085A (en) | Fan control system for tracking and adjusting based on intelligent infrared sensing technology | |
| CN111563886A (en) | Tunnel steel rail surface disease detection method and device based on unsupervised feature learning | |
| CN116128879B (en) | Lightweight transmission line defect detection method and device | |
| CN116091976A (en) | Station room defect identification detection method, system, device and storage medium | |
| CN112255141B (en) | Thermal imaging gas monitoring system | |
| CN114298973A (en) | Intelligent heat supply monitoring method based on infrared image segmentation | |
| CN114283367A (en) | Artificial intelligent open fire detection method and system for garden fire early warning | |
| CN112307896A (en) | Method for detecting lewd behavior abnormity of elevator under community monitoring scene | |
| CN116454882B (en) | Photovoltaic power generation prediction method based on machine vision predictor | |
| CN114429148A (en) | Power equipment state detection method based on multi-source data fusion | |
| CN112132088A (en) | Inspection point location missing inspection identification method | |
| CN110992339A (en) | Detection and positioning method and system for roller path line hub based on camera and machine learning | |
| CN116776086B (en) | Signal fault discriminating method and device based on self-attention mechanism self-encoder | |
| CN117409332B (en) | Long wood shaving appearance data detection system and method based on big data processing | |
| Yu et al. | Improved DenseNet-Based Defect Detection System for Photovoltaic Panels |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |