CN114782880A - Monitoring system for off-grid photovoltaic power generation system - Google Patents
Monitoring system for off-grid photovoltaic power generation system Download PDFInfo
- Publication number
- CN114782880A CN114782880A CN202210708919.0A CN202210708919A CN114782880A CN 114782880 A CN114782880 A CN 114782880A CN 202210708919 A CN202210708919 A CN 202210708919A CN 114782880 A CN114782880 A CN 114782880A
- Authority
- CN
- China
- Prior art keywords
- gaussian distribution
- background
- suspected
- gaussian
- pixel
- 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
- 238000010248 power generation Methods 0.000 title claims abstract description 27
- 238000012544 monitoring process Methods 0.000 title claims abstract description 22
- 238000009826 distribution Methods 0.000 claims abstract description 248
- 238000001514 detection method Methods 0.000 claims abstract description 13
- 238000012545 processing Methods 0.000 claims abstract description 8
- 238000000034 method Methods 0.000 claims description 19
- 239000011159 matrix material Substances 0.000 claims description 12
- 239000006185 dispersion Substances 0.000 claims description 8
- 239000000203 mixture Substances 0.000 claims description 6
- 230000000694 effects Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 239000000428 dust Substances 0.000 description 3
- 238000007689 inspection Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000003466 welding Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Photovoltaic Devices (AREA)
Abstract
The invention relates to the technical field of photovoltaic power generation, in particular to a monitoring system for an off-grid photovoltaic power generation system. The system can be used in the solar industry and comprises a suspected background Gaussian distribution acquisition module, a category identification module, an initial background image acquisition module and a detection alarm module; the suspected background Gaussian distribution acquisition module is used for acquiring a video image of the photovoltaic cell panel and suspected background Gaussian distribution corresponding to each pixel point in the video image; the category identification module is used for classifying all the pixel points to obtain a plurality of categories and identifying the photovoltaic cell category and the grating category; the initial background image acquisition module is used for updating the pixel value of each pixel point to obtain an initial background image; the detection alarm module is used for correcting the initial background image again to obtain a background image, and obtaining a shading area according to the background image to perform alarm processing; the monitoring timeliness of the photovoltaic system is improved, and the shading area is more accurately detected.
Description
Technical Field
The invention relates to the technical field of photovoltaic power generation, in particular to a monitoring system for an off-grid photovoltaic power generation system.
Background
The off-grid photovoltaic power generation system is an independent photovoltaic power generation system, and the purpose of supplying power is achieved by directly or indirectly converting solar radiation energy into electric energy through a photoelectric effect or a photochemical effect by absorbing sunlight; the off-grid photovoltaic power generation system is widely applied to places such as remote mountainous areas, non-electricity areas, islands, communication base stations, street lamps and the like.
During use, the off-grid photovoltaic power generation system is often influenced by the environment, so that shading such as tree shadows, dust, fallen leaves and the like is formed. When the battery assembly of the off-grid photovoltaic power generation system has a local shading condition, part of batteries can be heated, and a hot spot phenomenon is generated. If the local shading exists for a long time, when the hot spot effect reaches a certain degree, welding spots on the assembly can be melted and the grid lines can be damaged, so that the whole photovoltaic cell assembly is scrapped.
At present, the local shading condition of a battery assembly of an off-grid photovoltaic power generation system is monitored and detected usually through manual inspection or an image identification method. However, the manual inspection is high in labor cost and low in efficiency, the method for recognizing the images by shooting the images is easily interfered by moving objects such as flying birds and swaying branches and leaves, and the accuracy rate of detection results is low.
Disclosure of Invention
In order to solve the above technical problem, the present invention provides a monitoring system for an off-grid photovoltaic power generation system, which includes the following modules:
the suspected background Gaussian distribution acquisition module is used for acquiring a video image of the photovoltaic cell panel within a period of time; constructing a plurality of Gaussian models of each pixel point in the video image, and performing weighted summation on different Gaussian models corresponding to each pixel point to obtain corresponding Gaussian mixture background modeling, wherein when the Gaussian mixture background modeling is larger than a background model threshold value, the Gaussian models corresponding to the Gaussian mixture background modeling are in suspected background Gaussian distribution;
the category identification module is used for acquiring the suspected background Gaussian distribution with the largest weight of each pixel point as the maximum characteristic Gaussian distribution, calculating the difference of the maximum characteristic Gaussian distribution between any two pixel points, and dividing all the pixel points into a plurality of categories based on the difference between all the pixel points; identifying the photovoltaic cell type and the grid type in the types, and acquiring a first Gaussian distribution of the photovoltaic cell type and a second Gaussian distribution of the grid type;
an initial background image obtaining module, configured to obtain, according to the first gaussian distribution and the second gaussian distribution, a probability that each of the suspected background gaussian distributions is a background, obtain an initial pixel value of each pixel point based on the probability that all of the suspected background gaussian distributions corresponding to each pixel point are the background, where the initial pixel values of all the pixel points form an initial background image;
the detection alarm module is used for judging whether each pixel point in the initial background image meets first Gaussian distribution or second Gaussian distribution, when the first Gaussian distribution and the second Gaussian distribution are not met, the background model threshold is corrected and calculated to obtain a new background image, and when the pixel points which do not meet the first Gaussian distribution and the second Gaussian distribution exist in the background image, the pixel points are pixel points in a shading area, and alarm processing is timely carried out.
Preferably, the method for calculating the difference of the maximum feature gaussian distribution between any two pixel points in the category identification module includes:
and obtaining an expected difference value of the maximum characteristic Gaussian distribution corresponding to the two pixel points and a difference value of the coverage area, and obtaining the difference of the maximum characteristic Gaussian distribution between the two pixel points based on the expected difference value and the difference value of the coverage area.
Preferably, the method for identifying the photovoltaic cell category and the grid category in the category identification module includes:
counting the number of pixel points in each category, wherein the category corresponding to the maximum number of the pixel points is the category of the photovoltaic cell;
and obtaining the remaining categories except the categories of the photovoltaic battery pieces, calculating the sum of the distances between all pixel points in each remaining category, wherein the sum result is the dispersion degree of the remaining categories, and the remaining category with the maximum dispersion degree in all the remaining categories is the grid category.
Preferably, the method for acquiring a first gaussian distribution of the photovoltaic cell category and a second gaussian distribution of the grid category in the category identification module includes:
acquiring an expected mean value and a mean value of covariance matrices corresponding to all pixel points in the photovoltaic cell category, and obtaining a first Gaussian distribution of the photovoltaic cell category based on the expected mean value and the mean value of the covariance matrices;
and obtaining an expected mean value and a mean value of covariance matrixes corresponding to all pixel points in the grid category, and obtaining a second Gaussian distribution of the grid category based on the expected mean value and the mean value of the covariance matrixes.
Preferably, the method for acquiring, by the initial background image acquisition module, the probability that each suspected background gaussian distribution is a background according to the first gaussian distribution and the second gaussian distribution includes:
acquiring a background rate of each suspected background Gaussian distribution corresponding to any pixel point, wherein the ratio of the background rate of each suspected background Gaussian distribution to the sum of the background rates of all the suspected background Gaussian distributions corresponding to the pixel points is the probability that the suspected background Gaussian distribution is a background;
the formula for obtaining the background rate of the Gaussian distribution of each suspected background is as follows:
wherein,is shown asThe first of each pixel pointThe background rate of the suspected background Gaussian distribution;is shown asThe first of each pixel pointA difference between the respective suspected background gaussian distribution and the first gaussian distribution;is shown asThe first of each pixel pointA difference between the suspected background gaussian distribution and the second gaussian distribution;is shown asThe first of each pixel pointThe suspected background of Gaussian distribution andthe first of each pixel pointThe difference between the gaussian distributions of the individual suspected backgrounds;is shown asThe number of suspected background Gaussian distributions of the pixel points;is shown asThe first of each pixel pointThe suspected background of Gaussian distribution andthe mean value of the difference of other suspected background Gaussian distributions of the pixel points;represents a preset threshold;representing an exponential function;represents the minimum function.
Preferably, the method for obtaining the initial pixel value of the pixel point based on the probability that the gaussian distribution of all the suspected backgrounds corresponding to each pixel point is taken as the background includes:
and acquiring the maximum value of the probability that all the suspected background Gaussian distributions corresponding to each pixel point are backgrounds, wherein the expectation of the suspected background Gaussian distribution corresponding to the maximum value is the initial pixel value of the pixel point.
Preferably, the method for determining whether each pixel point in the initial background image satisfies a first gaussian distribution in the detection alarm module includes:
calculating a difference between a pixel value of each of the pixel points and an expectation of the first gaussian distribution, the pixel points satisfying the first gaussian distribution when the difference is not greater than a threshold; the threshold value is 2.5,The standard deviation of the first gaussian distribution is indicated.
The invention has the following beneficial effects: the invention provides a monitoring system for an off-grid photovoltaic power generation system, which can be used in the solar industry, wherein the monitoring system detects and analyzes a shading part through Gaussian distribution of each pixel point in a video image corresponding to a photovoltaic cell panel, can eliminate shading interference caused by conditions of birds, swaying branches and leaves and the like, enables monitoring of the photovoltaic system to be more timely, enables a shading area to be more accurate through detection, and can perform alarm reminding and timely processing on workers.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a system block diagram of monitoring for an off-grid photovoltaic power generation system according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a difference of coverage areas of gaussian distributions according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description, the structure, the features and the effects of a monitoring system for an off-grid photovoltaic power generation system according to the present invention will be provided with reference to the accompanying drawings and the preferred embodiments. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The method is suitable for analyzing whether the shading area exists on the surface of the photovoltaic cell panel in the off-grid photovoltaic power generation system, in order to solve the problem of low detection efficiency in the prior art, in the embodiment of the invention, a plurality of Gaussian models corresponding to each pixel point are obtained through video images of a photovoltaic cell panel within a period of time, suspected background Gaussian distribution corresponding to each pixel point is obtained, obtaining the initial pixel value of each pixel point according to the probability that all suspected background Gaussian distributions corresponding to each pixel point belong to the background so as to form an initial background image, analyzing and correcting each pixel point in the initial background image to obtain a final background image, analyzing whether a shading area exists on the photovoltaic cell panel or not based on the final background image, when the shading area is detected, the system can timely alarm to the staff for processing, and the analysis accuracy and the detection efficiency are effectively improved.
The following describes a specific scheme of the monitoring system for the off-grid photovoltaic power generation system provided by the invention in detail with reference to the accompanying drawings.
Referring to fig. 1, a block diagram of a system for monitoring an off-grid photovoltaic power generation system according to an embodiment of the present invention is shown, the system including the following modules:
a suspected background Gaussian distribution acquisition module 10, configured to acquire a video image of the photovoltaic cell panel within a period of time; and constructing a plurality of Gaussian models of each pixel point in the video image, carrying out weighted summation on different Gaussian models corresponding to each pixel point to obtain corresponding mixed Gaussian background modeling, and when the mixed Gaussian background modeling is larger than a background model threshold value, taking the plurality of Gaussian models corresponding to the mixed Gaussian background modeling as suspected background Gaussian distribution.
Specifically, carry on the camera on unmanned aerial vehicle, stop unmanned aerial vehicle at the place ahead of photovoltaic cell board and shoot the video image of this photovoltaic cell board in a period, the video is shot regional and is only contained the photovoltaic cell board region. In order to analyze and detect the accuracy of the photovoltaic cell panel, the shooting time of the video image is set to be 1 minute, and video image acquisition is performed once in the morning, at noon and in the afternoon of a day.
For a photovoltaic cell panel, dust, fallen leaves on the surface of the photovoltaic cell panel, tree shadows projected on the photovoltaic cell panel and the like belong to shading, and shadows caused by birds, branches, leaves and the like near the photovoltaic cell panel shake do not belong to shading; however, factors such as birds, branches, leaves and the like may exist in the video image captured by the camera to interfere with the surface image of the photovoltaic cell panel, and therefore the captured video image of the photovoltaic cell panel needs to be analyzed and identified to eliminate interference of the factors such as the birds, the branches, the leaves and the like.
Because the video image is actually formed by a plurality of frames of video images, the pixel points of the photovoltaic cell panel contained in each frame of video image are the same, but the pixel values of the pixel points may have different changes; for each pixel point in the video image, the change of the pixel point in the video image is random, so that a corresponding Gaussian model can be constructed according to the change of the pixel value of the pixel point in the whole video image, namely the corresponding Gaussian model is constructed according to the change of the pixel value of the same pixel point in continuous video images; in order to increase the reliability of data, in the embodiment of the present invention, 5 gaussian models are constructed for each position pixel, and in other embodiments, an implementer may set the models according to the situation.
Further, obtaining a mixed Gaussian background modeling of each position pixel point in the video image, wherein the mixed Gaussian background modeling is obtained by overlapping Gaussian models corresponding to the position pixel point in the video image through different weights, namely the mixed Gaussian background modeling of each position pixel point is obtained by overlapping 5 Gaussian models of the position pixel point under different weights; the weight value of each Gaussian model is correspondingly obtained when the Gaussian model is constructed; and preliminarily dividing the classes of the pixel points based on the mixed Gaussian background modeling corresponding to each pixel point.
In the embodiment of the invention, the photovoltaic cell panel and the shading part are recorded as a background, and the non-shading interference part is recorded as a foreground; when the mixed Gaussian background modeling of any pixel point is larger than the background model threshold, marking the Gaussian distribution of the pixel point as the suspected background Gaussian distribution, and acquiring the suspected background Gaussian distribution by the specific method comprises the following steps:
in the embodiment of the invention, the threshold value of a background model is set to be 0.6, the superposition of 5 Gaussian models corresponding to each pixel point is calculated, the Gaussian models with different quantities are superposed according to weights to obtain different modeling results of mixed Gaussian backgrounds, and when the weight of any 1 Gaussian model in the 5 Gaussian models is more than 0.6, the 1 Gaussian model of the pixel point is in suspected background Gaussian distribution; when the sum of the weights corresponding to the two Gaussian models is greater than 0.6, the two Gaussian models of the pixel point are in suspected background Gaussian distribution; that is, when the weight superposition of the minimum number of gaussian models in the gaussian models corresponding to each pixel point is greater than 0.6, the gaussian models are the suspected background gaussian distribution corresponding to the pixel point.
The category identification module 20 is configured to obtain a suspected background gaussian distribution with the largest weight of each pixel point as a maximum feature gaussian distribution, calculate a difference between the maximum feature gaussian distributions of any two pixel points, and divide all the pixel points into a plurality of categories based on the difference between all the pixel points; and identifying the photovoltaic cell type and the grid type in the types, and acquiring a first Gaussian distribution of the photovoltaic cell type and a second Gaussian distribution of the grid type.
When the color of the pixel point at the same position in the video image is unchanged, namely the photovoltaic cell panel is not interfered by any moving target, the pixel point is corresponding to only one suspected background Gaussian distribution; when the color of the pixel point at the same position in the video image changes, for example, the pixel point shows the color of a branch at some time, the color of a leaf at some time and the color of a photovoltaic cell panel at some time due to the shaking of the branch and the leaf, the pixel point may correspond to a plurality of suspected background Gaussian distributions; therefore, it is necessary to find an accurate background from a gaussian distribution of a plurality of suspected backgrounds corresponding to each pixel point, so as to keep the shading characteristics of dust, tree shadows, fallen leaves, and the like while eliminating the interference of non-shading objects such as birds, swaying branches and leaves, and the like.
Firstly, obtaining the number of the suspected background Gaussian distribution corresponding to any pixel point and recording the number asWhen it comes toWhen indicates thatThe probability that the suspected background Gaussian distribution of the pixel points is taken as the background is 1; when the temperature is higher than the set temperatureCalculating the probability that each suspected background Gaussian distribution of the pixel point is a background; will be firstThe first of each pixel pointThe gaussian distribution of each suspected background is given as:
wherein,is shown asThe first of each pixel pointA gaussian distribution of individual suspected backgrounds;denotes the firstFirst pixel pointCovariance of Gaussian distribution of individual suspected backgroundsA matrix;representing a transposed symbol;representing the circumferential ratio;is shown asThe first of each pixel point(ii) an expectation of a gaussian distribution of the suspected background;representing an exponential function.
It should be noted that each frame of video image in the video image is an RGB image, and each suspected background gaussian distribution is a three-dimensional gaussian distribution.
Then, the suspected background Gaussian distribution with the largest weight in all the suspected background Gaussian distributions corresponding to each pixel point is obtained and recorded as the maximum characteristic Gaussian distribution of the pixel point, the difference between the maximum characteristic Gaussian distributions between every two pixel points is calculated and used as the difference distance between every two corresponding pixel points, and all the pixel points are subjected to spectral clustering based on the difference distance between every two pixel points, so that a plurality of categories are obtained.
Preferably, in the embodiment of the present invention, all the pixel points are divided into 10 categories.
And obtaining an expected difference value of the maximum characteristic Gaussian distribution corresponding to the two pixel points and a difference value of the coverage area, and obtaining the difference of the maximum characteristic Gaussian distribution between the two pixel points based on the expected difference value and the difference value of the coverage area. Calculating the difference between the maximum characteristic Gaussian distributions between every two pixel points as follows:
wherein,denotes the firstMaximum characteristic Gaussian distribution and number of pixel pointsThe difference between the maximum characteristic gaussian distributions of the individual pixel points;denotes the firstExpectation of maximum characteristic Gaussian distribution of each pixel point;denotes the firstExpectation of maximum characteristic Gaussian distribution of each pixel point;representing a normalized coefficient;is shown asMaximum characteristic Gaussian distribution of each pixel point;denotes the firstMaximum characteristic Gaussian distribution of the pixel points;representAnd withA difference in coverage area;an L2 paradigm representing the expected difference of the two maximum characteristic Gaussian distributions;a hyperbolic tangent function is represented for normalization.
Further, take a one-dimensional gaussian distribution as an example; referring to fig. 2, a schematic diagram of the difference of coverage areas with gaussian distributions is shown, wherein the black part is the difference of coverage areas with two gaussian distributions, and when the difference of coverage areas is larger, the difference of distribution of two gaussian distributions is larger; the difference in normalized coverage areas is closer to 1 as the two gaussian distributions are farther apart. But the effect of the difference between the two largest characteristic gaussian distributions measured only in terms of the distance covering the difference is limited, so that in combination with the expected difference between the two largest characteristic gaussian distributions, the greater the expected difference, the greater the difference between the two largest characteristic gaussian distributions is indicated.
Counting the number of pixel points in each category, wherein the category corresponding to the maximum number of the pixel points is the category of the photovoltaic cell; and obtaining the remaining categories except the categories of the photovoltaic cells, calculating the sum of the distances between all pixel points in each remaining category, wherein the sum result is the dispersion degree of the remaining categories, and the remaining category with the maximum dispersion degree in all the remaining categories is the grid category. Acquiring an expected mean value and a mean value of covariance matrixes corresponding to all pixel points in the photovoltaic cell category, and obtaining a first Gaussian distribution of the photovoltaic cell category based on the expected mean value and the mean value of the covariance matrixes; and obtaining an expected mean value and a mean value of the covariance matrix corresponding to all pixel points in the grid category, and obtaining a second Gaussian distribution of the grid category based on the expected mean value and the mean value of the covariance matrix.
Specifically, because the area of the photovoltaic cell on the photovoltaic cell panel is large, and the shading is only a local area on the photovoltaic cell panel, the category corresponding to the maximum number of the pixel points in the 10 categories is identified as the category of the photovoltaic cell, the expected mean value of the maximum characteristic gaussian distribution corresponding to all the pixel points in the category of the photovoltaic cell and the mean value of the covariance matrix are obtained, and the first gaussian distribution of the category of the photovoltaic cell obtained based on the expected mean value of all the pixel points and the mean value of the covariance matrix is:
wherein,representing a first Gaussian distribution of the photovoltaic cell;representing the mean value of covariance matrixes of all pixel points in the photovoltaic cell category;representing expected mean values of all pixel points in the category of the photovoltaic cell;representing a transposed symbol;representing an exponential function.
Further, residual categories except the photovoltaic cell category are analyzed, the distance between every two pixel points in each residual category is calculated and summed, the obtained result is used for expressing the dispersion degree of all the pixel points in the residual category in the video image, and the larger the distance sum is, the wider the distribution of the pixel points in the residual category is; the grids of the photovoltaic cell panel are distributed over the whole photovoltaic cell panel area, and the distribution range is wide; the shaded or shaken object is only in a local area of the photovoltaic cell panel, and the distribution range is small; therefore, the remaining category with the largest dispersion degree among the remaining categories except the photovoltaic cell sheet category is identified as the grid category of the photovoltaic cell panel. Correspondingly, obtaining the expected mean value of the maximum characteristic gaussian distribution corresponding to all the pixel points in the grid category and the mean value of the covariance matrix, and obtaining the second gaussian distribution of the grid category of the photovoltaic cell panel according to the expected mean value of all the pixel points in the grid category and the mean value of the covariance matrix is as follows:
wherein,a second gaussian distribution representing a grid;representing the mean of the covariance matrices of all the pixels in the grid class;representing an expected mean of all pixel points in the grid category;representing a transposed symbol;representing an exponential function.
The initial background image obtaining module 30 is configured to obtain a probability that each suspected background gaussian distribution is a background according to the first gaussian distribution and the second gaussian distribution, obtain initial pixel values of the pixel points based on the probabilities that all the suspected background gaussian distributions corresponding to each pixel point are the background, and form an initial background image with the initial pixel values of all the pixel points.
The category identification module 20 obtains the first gaussian distribution of the photovoltaic cell and the second gaussian distribution of the grid, and calculates the probability that each suspected background gaussian distribution is the background based on the first gaussian distribution of the photovoltaic cell and the second gaussian distribution of the grid as follows:
wherein,denotes the firstThe first of each pixel pointProbability that the suspected background Gaussian distribution is the background;denotes the firstThe first of each pixel pointBackground rate of each suspected background gaussian distribution;denotes the firstThe first of each pixel pointBackground rate of each suspected background gaussian distribution;denotes the firstThe number of suspected background gaussian distributions for each pixel.
The calculation of the background rate of the gaussian distribution of the specific suspected background is:
wherein,is shown asThe first of each pixel pointBackground rate of each suspected background gaussian distribution;is shown asThe first of each pixel pointThe difference between the suspected background Gaussian distribution and the first Gaussian distribution of the photovoltaic cell slice;is shown asThe first of each pixel pointA difference between the gaussian distribution of each suspected background and a second gaussian distribution of the grid;is shown asThe first of each pixel pointThe suspected background of Gaussian distribution andthe first of each pixel pointThe difference between the gaussian distributions of the individual suspected backgrounds;is shown asThe number of suspected background Gaussian distributions of the pixel points;is shown asThe first of each pixel pointThe suspected background of Gaussian distribution andthe mean value of the difference of other suspected background Gaussian distributions of the pixel points;represents a preset threshold;representing an exponential function;represents the minimum function.
When it comes toThe first of each pixel pointThe suspected background of Gaussian distribution andthe average value of the difference of other suspected background Gaussian distributions of the pixel points is larger than a preset threshold valueWhen it comes toThe first of each pixel pointThe suspected background Gaussian distribution may contain the characteristics of swaying leaves, branches and photovoltaic panels, namelyThe smaller the value of (A), theThe first of each pixel pointThe difference between the suspected background gaussian distribution and the first gaussian distribution of the photovoltaic cell or the second gaussian distribution of the grid is small,the larger, the corresponding, theThe first of each pixel pointBackground rate of Gaussian distribution of suspected backgroundsThe larger, theThe first of each pixel pointProbability of background being Gaussian distribution of suspected backgroundThe larger.
When it comes toThe first of each pixel pointGaussian distribution of suspected background andthe average value of the difference of other suspected background Gaussian distributions of each pixel point is less than a preset threshold valueWhen it comes toThe first of each pixel pointIndividual suspected backgroundThe Gaussian distribution may include the characteristics of a swaying tree shadow and a photovoltaic cell panel, the tree shadow is a shadow part superposed on the photovoltaic cell panel, and compared with the photovoltaic cell panel, the tree shadow is darker in color and is closer to the characteristics of the photovoltaic cell panel; when the surface of the photovoltaic cell panel has the tree shadow, the tree shadow area belongs to the background, and the photovoltaic cell panel area covered by the tree shadow does not belong to the background; thus at the time of judgingThe first of each pixel pointWhen the probability that the Gaussian distribution of each suspected background belongs to the background is determined, the probability is actually the firstThe first of each pixel pointProbability that each suspected background Gaussian distribution belongs to a tree shadow; when it comes toThe first of each pixel pointWhen the suspected background Gaussian distribution belongs to the tree shadow, the difference between the suspected background Gaussian distribution and the first Gaussian distribution of the photovoltaic cell and the difference between the suspected background Gaussian distribution and the second Gaussian distribution of the grid are larger, and at the moment, the difference is largerIs also larger, correspondinglyThe first of each pixel pointBackground rate of Gaussian distribution of individual suspected backgroundsThe larger, theThe first of each pixel pointProbability of background being Gaussian distribution of suspected backgroundThe larger.
By analogy, obtain the firstAnd the background rate of each suspected background Gaussian distribution of each pixel point is obtained, so that the probability that each suspected background Gaussian distribution belongs to the background is obtained according to the background rate.
Further, corresponding on a per pixel basisObtaining initial background image according to probability of the Gaussian distribution of the suspected background belonging to the background, and selecting the point corresponding to each pixel pointThe suspected background Gaussian distribution belongs to the maximum value of the probability of the background, and the expectation of the suspected background Gaussian distribution corresponding to the maximum value is used as the initial pixel value of the pixel point; and obtaining initial pixel values of all the pixel points to form an initial background image in the same way.
And the detection alarm module 40 is used for judging whether each pixel point in the initial background image meets the first Gaussian distribution or the second Gaussian distribution, when the first Gaussian distribution and the second Gaussian distribution are not met, correcting and calculating the threshold value of the background model to obtain a new background image, and when the pixel points which do not meet the first Gaussian distribution and the second Gaussian distribution exist in the background image, the pixel points are pixel points in a shading area, and timely carrying out alarm processing.
When the color of the photovoltaic cell panel is less, the situation that the pixel does not contain the photovoltaic cell panel when suspected background Gaussian distribution is selected may be caused; therefore, the suspected background gaussian distribution corresponding to each pixel point needs to be corrected.
Specifically, it is determined whether the pixel value of each pixel in the initial background image satisfies the first gaussian distribution of the photovoltaic cell or the second gaussian distribution of the grid, that is, the difference between the pixel value of the pixel and the expectation of the first gaussian distribution of the photovoltaic cell is calculated, and when the difference is not greater than 2.5If so, the pixel point meets the first Gaussian distribution of the photovoltaic cell; otherwise, the pixel point does not conform to the first Gaussian distribution of the photovoltaic cell; wherein,standard deviation of a first Gaussian distribution of a photovoltaic cell sheet, covariance matrix of the first Gaussian distributionThe square root of the diagonal elements is taken.
And similarly, calculating the difference between the pixel value of the pixel point and the expectation of the second Gaussian distribution of the grating, and judging whether the pixel point accords with the second Gaussian distribution of the grating.
When any pixel point does not accord with the first Gaussian distribution of the photovoltaic cell piece and does not accord with the second Gaussian distribution of the grating, recording the pixel point as a central point to judge whether the pixel point in the neighborhood meets the first Gaussian distribution of the photovoltaic cell piece or the second Gaussian distribution of the grating; in the embodiment of the invention, the neighborhood pixel points of any pixel point comprise 25 pixel points; when the pixel points in 25 neighborhood pixel points of the pixel point accord with the first Gaussian distribution of the photovoltaic cell or the second Gaussian distribution of the grating, the central point is a suspected background pixel point;
obtaining all suspected background pixel points in the initial background image, correcting background model thresholds of all the suspected background pixel points, obtaining new suspected background Gaussian distribution corresponding to each suspected background pixel point based on the corrected background model threshold, further calculating the probability of belonging to the background of all the suspected background Gaussian distributions to obtain a new pixel value of each suspected background pixel point, and thus obtaining a final background image.
Preferably, in the embodiment of the present invention, the background model threshold is modified to 0.8.
And further, whether each pixel point in the background image belongs to the first Gaussian distribution of the photovoltaic cell or the second Gaussian distribution of the grid is judged, when the pixel points do not satisfy the first Gaussian distribution and do not satisfy the second Gaussian distribution, the pixel points are pixel points of the shading area, all the pixel points which do not satisfy the first Gaussian distribution and do not satisfy the second Gaussian distribution are obtained to obtain the final shading area, the alarm notification is carried out on the staff, and the staff can timely process shading.
To sum up, an embodiment of the present invention provides a monitoring system for an off-grid photovoltaic power generation system, where the monitoring system includes: a suspected background Gaussian distribution acquisition module 10, a category identification module 20, an initial background image acquisition module 30 and a detection alarm module 40; the method comprises the steps of analyzing a video image of a photovoltaic cell panel within a period of time to obtain a plurality of Gaussian models corresponding to each pixel point, further obtaining suspected background Gaussian distribution in the Gaussian models corresponding to each pixel point, calculating the probability that each suspected background Gaussian distribution belongs to a background, obtaining an initial pixel value of each pixel point according to the probability that all the suspected background Gaussian distributions corresponding to each pixel point belong to the background, forming an initial background image by the initial pixel values of all the pixel points, analyzing and correcting each pixel point in the initial background image to obtain a final background image, analyzing whether a shading area exists on the photovoltaic cell panel based on the final background image, timely alarming to a worker for processing when the shading area is detected, and effectively improving the accuracy of analysis and the efficiency of detection.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
All the embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and should not be taken as limiting the scope of the present invention, which is intended to cover any modifications, equivalents, improvements, etc. within the spirit of the present invention.
Claims (7)
1. A monitoring system for an off-grid photovoltaic power generation system, the system comprising the following modules:
the suspected background Gaussian distribution acquisition module is used for acquiring a video image of the photovoltaic cell panel within a period of time; constructing a plurality of Gaussian models of each pixel point in the video image, weighting and summing different Gaussian models corresponding to each pixel point to obtain corresponding Gaussian mixture background modeling, wherein when the Gaussian mixture background modeling is larger than a background model threshold value, the Gaussian models corresponding to the Gaussian mixture background modeling are in suspected background Gaussian distribution;
the category identification module is used for acquiring the suspected background Gaussian distribution with the largest weight of each pixel point as the maximum characteristic Gaussian distribution, calculating the difference of the maximum characteristic Gaussian distribution between any two pixel points, and dividing all the pixel points into a plurality of categories based on the difference between all the pixel points; identifying the photovoltaic cell type and the grid type in the types, and acquiring a first Gaussian distribution of the photovoltaic cell type and a second Gaussian distribution of the grid type;
an initial background image obtaining module, configured to obtain, according to the first gaussian distribution and the second gaussian distribution, a probability that each of the suspected background gaussian distributions is a background, obtain an initial pixel value of each pixel point based on the probability that all of the suspected background gaussian distributions corresponding to each pixel point are the background, where the initial pixel values of all the pixel points form an initial background image;
the detection alarm module is used for judging whether each pixel point in the initial background image meets first Gaussian distribution or second Gaussian distribution, when the first Gaussian distribution and the second Gaussian distribution are not met, the background model threshold is corrected and calculated to obtain a new background image, and when the pixel points which do not meet the first Gaussian distribution and the second Gaussian distribution exist in the background image, the pixel points are pixel points in a shading area, and alarm processing is timely carried out.
2. The monitoring system for the off-grid photovoltaic power generation system according to claim 1, wherein the method for calculating the difference of the maximum characteristic gaussian distribution between any two pixel points in the category identification module comprises:
and obtaining an expected difference value of the maximum characteristic Gaussian distribution corresponding to the two pixel points and a difference value of the coverage area, and obtaining the difference of the maximum characteristic Gaussian distribution between the two pixel points based on the expected difference value and the difference value of the coverage area.
3. The monitoring system for the off-grid photovoltaic power generation system according to claim 1, wherein the method for identifying the photovoltaic cell category and the grid category in the category identification module comprises:
counting the number of pixel points in each category, wherein the category corresponding to the maximum number of the pixel points is the category of the photovoltaic cell;
and obtaining the residual categories except the photovoltaic cell categories, calculating the sum of the distances between all pixel points in each residual category, wherein the sum result is the dispersion degree of the residual categories, and the residual category with the maximum dispersion degree in all the residual categories is the grid category.
4. The monitoring system for the off-grid photovoltaic power generation system according to claim 1, wherein the method for obtaining the first gaussian distribution of the photovoltaic cell category and the second gaussian distribution of the grid category in the category identification module comprises:
acquiring an expected mean value and a mean value of a covariance matrix corresponding to all pixel points in the photovoltaic cell category, and obtaining a first Gaussian distribution of the photovoltaic cell category based on the expected mean value and the mean value of the covariance matrix;
and obtaining an expected mean value and a mean value of the covariance matrix corresponding to all pixel points in the grid category, and obtaining a second Gaussian distribution of the grid category based on the expected mean value and the mean value of the covariance matrix.
5. The monitoring system for the off-grid photovoltaic power generation system according to claim 1, wherein the method for obtaining the probability of each suspected background gaussian distribution as the background according to the first gaussian distribution and the second gaussian distribution in the initial background image obtaining module comprises:
acquiring a background rate of each suspected background Gaussian distribution corresponding to any pixel point, wherein the ratio of the background rate of each suspected background Gaussian distribution to the sum of the background rates of all the suspected background Gaussian distributions corresponding to the pixel points is the probability that the suspected background Gaussian distribution is the background;
the formula for obtaining the background rate of each suspected background Gaussian distribution is as follows:
wherein,denotes the firstThe first of each pixel pointBackground rate of each suspected background gaussian distribution;is shown asThe first of each pixel pointA difference between the respective suspected background gaussian distribution and the first gaussian distribution;is shown asThe first of each pixel pointA difference between the respective suspected background gaussian distribution and the second gaussian distribution;is shown asThe first of each pixel pointThe suspected background of Gaussian distribution andthe first of each pixel pointThe difference between the gaussian distributions of the individual suspected backgrounds;is shown asThe number of suspected background Gaussian distributions of the pixel points;denotes the firstThe first of each pixel pointThe suspected background of Gaussian distribution andthe mean value of the difference of other suspected background Gaussian distributions of the pixel points;represents a preset threshold;representing an exponential function;represents the minimum function.
6. The monitoring system for the off-grid photovoltaic power generation system according to claim 1, wherein the method for obtaining the initial pixel value of the pixel point based on the probability that all suspected background gaussian distributions corresponding to each pixel point are backgrounds comprises:
and acquiring the maximum value of the probability that all the suspected background Gaussian distributions corresponding to each pixel point are backgrounds, wherein the expectation of the suspected background Gaussian distribution corresponding to the maximum value is the initial pixel value of the pixel point.
7. The monitoring system for the off-grid photovoltaic power generation system according to claim 1, wherein the method for determining whether each pixel point in the initial background image satisfies a first gaussian distribution in the detection alarm module comprises:
calculating a difference between a pixel value of each of the pixel points and an expectation of the first gaussian distribution, the pixel points satisfying the first gaussian distribution when the difference is not greater than a threshold; the threshold value is 2.5,The standard deviation of the first gaussian distribution is indicated.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210708919.0A CN114782880B (en) | 2022-06-22 | 2022-06-22 | Monitoring system for off-grid photovoltaic power generation system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210708919.0A CN114782880B (en) | 2022-06-22 | 2022-06-22 | Monitoring system for off-grid photovoltaic power generation system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114782880A true CN114782880A (en) | 2022-07-22 |
CN114782880B CN114782880B (en) | 2022-11-01 |
Family
ID=82422364
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210708919.0A Active CN114782880B (en) | 2022-06-22 | 2022-06-22 | Monitoring system for off-grid photovoltaic power generation system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114782880B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116797603A (en) * | 2023-08-25 | 2023-09-22 | 江苏惠汕新能源集团有限公司 | Photovoltaic array fault diagnosis and positioning method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102568005A (en) * | 2011-12-28 | 2012-07-11 | 江苏大学 | Moving object detection method based on Gaussian mixture model |
CN112257557A (en) * | 2020-10-20 | 2021-01-22 | 中国电子科技集团公司第五十八研究所 | High-altitude parabolic detection and identification method and system based on machine vision |
CN112802054A (en) * | 2021-02-04 | 2021-05-14 | 重庆大学 | Mixed Gaussian model foreground detection method fusing image segmentation |
CN113506230A (en) * | 2021-09-10 | 2021-10-15 | 南通欧泰机电工具有限公司 | Photovoltaic power station aerial image dodging processing method based on machine vision |
-
2022
- 2022-06-22 CN CN202210708919.0A patent/CN114782880B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102568005A (en) * | 2011-12-28 | 2012-07-11 | 江苏大学 | Moving object detection method based on Gaussian mixture model |
CN112257557A (en) * | 2020-10-20 | 2021-01-22 | 中国电子科技集团公司第五十八研究所 | High-altitude parabolic detection and identification method and system based on machine vision |
CN112802054A (en) * | 2021-02-04 | 2021-05-14 | 重庆大学 | Mixed Gaussian model foreground detection method fusing image segmentation |
CN113506230A (en) * | 2021-09-10 | 2021-10-15 | 南通欧泰机电工具有限公司 | Photovoltaic power station aerial image dodging processing method based on machine vision |
Non-Patent Citations (1)
Title |
---|
刘璐等: "一种基于高斯混合模型的分块背景更新算法研究", 《工业控制计算机》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116797603A (en) * | 2023-08-25 | 2023-09-22 | 江苏惠汕新能源集团有限公司 | Photovoltaic array fault diagnosis and positioning method |
CN116797603B (en) * | 2023-08-25 | 2023-10-24 | 江苏惠汕新能源集团有限公司 | Photovoltaic array fault diagnosis and positioning method |
Also Published As
Publication number | Publication date |
---|---|
CN114782880B (en) | 2022-11-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109416413B (en) | Solar energy forecast | |
Kamadinata et al. | Sky image-based solar irradiance prediction methodologies using artificial neural networks | |
CN110514298B (en) | Solar radiation intensity calculation method based on foundation cloud picture | |
Chu et al. | Short-term probabilistic forecasts for direct normal irradiance | |
Di Tommaso et al. | A multi-stage model based on YOLOv3 for defect detection in PV panels based on IR and visible imaging by unmanned aerial vehicle | |
CN106779130B (en) | Photovoltaic power station radiation prediction method based on all-sky cloud picture | |
Marquez et al. | Intra-hour DNI forecasting based on cloud tracking image analysis | |
CN107133685B (en) | Method and system for predicting power generation capacity of photovoltaic power generation system | |
WO2017193172A1 (en) | "solar power forecasting" | |
Dissawa et al. | Sky Image‐Based Localized, Short‐Term Solar Irradiance Forecasting for Multiple PV Sites via Cloud Motion Tracking | |
CN116191680B (en) | Monitoring management system applied to photovoltaic power generation | |
CN114782880B (en) | Monitoring system for off-grid photovoltaic power generation system | |
CN110633862A (en) | Optical power prediction algorithm based on satellite cloud picture | |
CN114743115A (en) | Shelter recognition method and fault early warning system for photovoltaic electronic equipment | |
CN106780540B (en) | photovoltaic power generation-oriented ground cloud picture cloud layer tracking and early warning method | |
CN116046796A (en) | Photovoltaic module hot spot detection method and system based on unmanned aerial vehicle | |
Dissawa et al. | Cross-correlation based cloud motion estimation for short-term solar irradiation predictions | |
Aakroum et al. | Deep learning for inferring the surface solar irradiance from sky imagery | |
CN116307220A (en) | Roof photovoltaic power prediction method and device and electronic equipment | |
CN115423758B (en) | Full-field refined DNI prediction method | |
CN116029440A (en) | Ultra-short-term power prediction method and device for photovoltaic power station | |
CN113033326B (en) | Photovoltaic power station construction treading assembly monitoring method | |
CN109860742B (en) | Method for identifying electrolyte leakage of communication power supply storage battery of transformer substation | |
CN113657621A (en) | Hidden danger monitoring method and system | |
Dissawa et al. | On-Site Solar Power Forecasting Using Sky-Images |
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 |