A photovoltaic panel flatness detection method and system in photovoltaic construction

Abstract

The present invention relates to the technical field of flatness detection of solar panels, and specifically to a method and system for detecting the flatness of photovoltaic panels in photovoltaic construction. Point cloud data of the surface of photovoltaic panels is obtained to characterize its surface morphology. By analyzing the height characteristics of each data point in the point cloud data, abnormal data points that do not conform to the normal surface morphology are initially screened out. Then, cluster analysis is used to cluster the abnormal data points, and all raised and recessed deformation areas on the surface of the photovoltaic panels are distinguished in combination with the height information. Furthermore, the height change fluctuations and positional relationships of the abnormal data points in each raised deformation area are comprehensively considered, the probability of it being a defective area is calculated, and the real raised defective area is screened out based on this. Finally, the area of ​​all defects and recessed areas is comprehensively considered to determine the flatness of the surface of the photovoltaic panel, thereby improving the credibility of the detection results.

Description

Technical Field

[0001] The present invention relates to the technical field of flatness detection of solar panels, and in particular to a method and system for detecting flatness of photovoltaic panels in photovoltaic construction. Background Art

[0002] Photovoltaic panels, also known as solar panels or photovoltaic modules, are core devices that convert light energy into electrical energy. During the photovoltaic construction process, various mechanical operations (such as cutting, grinding, etc.) may exert uneven forces on the photovoltaic panels, resulting in mechanical stress. This stress may cause depressions or protrusions on the surface of the photovoltaic panels, and the unevenness of the surface depressions or protrusions will further affect the photoelectric conversion efficiency of the photovoltaic panels, thereby reducing the performance of the photovoltaic panels. Therefore, it is necessary to detect the flatness of the surface of the photovoltaic panels.

[0003] The existing technology usually obtains the three-dimensional point cloud data of the surface of the photovoltaic panel through laser scanning technology, and then reflects the flatness of the surface of the photovoltaic panel through the difference characteristics between the height information and the mean height in each point cloud data. However, since the surface of the photovoltaic panel needs to be chemically corroded and textured during the construction process to increase light capture and reduce reflection, this operation will cause many pyramid-shaped protrusions to form on the surface of the photovoltaic panel, making the surface of the photovoltaic panel normal uneven. Therefore, it is impossible to effectively distinguish the normal protrusions caused by the texture operation from the protrusions in the defective area based solely on the difference in height values, and the credibility of the final flatness detection result will also be reduced. Summary of the invention

[0004] In order to solve the technical problem that the normal protrusions caused by the texturing operation cannot be effectively distinguished from the protrusions in the defective area based only on the height value differences at various positions on the surface of the photovoltaic panel, which will lead to a reduction in the credibility of the final flatness detection result, the purpose of the present invention is to provide a photovoltaic panel flatness detection method and system in photovoltaic construction, and the technical scheme adopted is as follows:

[0005] Obtain point cloud data of the photovoltaic panel surface;

[0006] According to the height value characteristics of all data points in the point cloud data, abnormal data points on the surface of the photovoltaic panel are screened out; all abnormal data points are clustered and analyzed, and all convex deformation areas and concave deformation areas on the surface of the photovoltaic panel are determined according to the height information of the data points in each cluster;

[0007] In each convex deformation area, based on the fluctuation of the height values ​​between the corresponding abnormal data points and the positional relationship between the abnormal data points, the probability of each convex deformation area being a defect area is determined and the defect areas are screened out from all convex deformation areas;

[0008] The flatness of the photovoltaic panel surface is determined by comprehensively considering the area of ​​all defective areas and concave deformation areas on the photovoltaic panel surface, thereby performing flatness detection.

[0009] Furthermore, the method for obtaining abnormal data points includes:

[0010] Constructing a three-dimensional height watershed model for the point cloud data on the surface of the photovoltaic panel, and constructing a flooding volume-water injection volume curve based on the three-dimensional height watershed model;

[0011] Obtaining the tangent slope value at each water injection volume on the flooding volume-water injection volume curve, and determining the change demarcation point according to the difference between the tangent slope values ​​at different water injection volumes;

[0012] On the flooding amount-water injection amount curve, the data points flooded after the change boundary point are regarded as abnormal data points on the surface of the photovoltaic panel.

[0013] Furthermore, the three-dimensional height watershed model is constructed based on the point cloud data of the surface of the photovoltaic panel, and the inundation volume-water injection volume curve is constructed based on the three-dimensional height watershed model, including:

[0014] The average of the height values ​​of all data points in the point cloud data of the surface of the photovoltaic panel is used as the height mean, the difference between the height value of each data point and the height mean is used as the z-axis coordinate corresponding to each data point, and a three-dimensional coordinate system is constructed in combination with the position information of the data points;

[0015] Mapping all data points in the point cloud data of the surface of the photovoltaic panel to the three-dimensional coordinate system, and establishing a three-dimensional height watershed model based on a gridding method;

[0016] In the three-dimensional height watershed model, the water injection process is simulated simultaneously from the xoy plane along the z-axis to both sides of the xoy plane, and the flooding amount of data points under different water injection amounts is calculated, thereby obtaining the flooding amount-water injection amount curve.

[0017] Furthermore, the method for obtaining the change demarcation point includes:

[0018] On the flooding volume-injection volume curve, other injection volumes except the first injection volume and the last injection volume are taken as target points;

[0019] On the submerged volume-injection volume curve, for any target point, the normalized value of the difference between the tangent slope value at the next injection volume adjacent to the target point and the tangent slope value at the previous injection volume adjacent to the target point is used as the change degree value of the target point;

[0020] Among all target points, the target point corresponding to the maximum value of the change degree is taken as the change demarcation point.

[0021] Furthermore, the cluster analysis is performed on all abnormal data points, and all convex deformation areas and concave deformation areas on the surface of the photovoltaic panel are determined according to the height information of the data points in each cluster, including:

[0022] All abnormal data points are clustered based on the DBSCAN density clustering algorithm to obtain all clusters, where the minimum sample size and neighborhood radius are both preset values;

[0023] The area formed by the data points in each cluster on the surface of the photovoltaic panel is used as the initial deformation area of ​​the photovoltaic panel surface;

[0024] The numerical mean of all data points in each initial deformation area on the z-axis is taken as the height characteristic value corresponding to each initial deformation area, the initial deformation area with positive height characteristic value is taken as the convex deformation area, and the initial deformation area with negative height characteristic value is taken as the concave deformation area.

[0025] Furthermore, the method for obtaining the defective area includes:

[0026] In each convex deformation area, the abnormal data points in the corresponding point cloud data are smoothly connected in the three-dimensional coordinate system to obtain multiple undulating curve segments. Among all the undulating curve segments, the target curve segment is screened out based on the position information of the undulating curve segment;

[0027] Analyze the positional relationship between abnormal data points on each target curve segment and the shape characteristics of the target curve segment to determine the clutter value of each target curve segment;

[0028] The difference between the maximum height value and the minimum height value of the abnormal data points on each target curve segment is taken as the abnormal factor, and the product of the clutter value of each target curve segment and the abnormal factor is taken as the abnormal degree value of each target curve;

[0029] In each convex deformation area, the normalized value of the average of the abnormal degree values ​​of all corresponding target curve segments in the three-dimensional coordinate system is used as the probability that each convex deformation area is a defect area;

[0030] When the probability that the convex deformation region is a defect region is greater than a preset defect threshold, the convex deformation region is determined to be a defect region.

[0031] Furthermore, the method for acquiring the target curve segment includes:

[0032] The geometric center of each convex deformation area is used as a reference point;

[0033] For any undulating curve segment corresponding to each convex deformation area, when the projection of the undulating curve segment on the xoy plane passes through the corresponding reference point and the projection is a straight line, the undulating curve segment is used as the target curve segment.

[0034] Furthermore, the method for obtaining the clutter level value includes:

[0035] On each target curve segment, obtain all extreme value points and take the curve segment between two adjacent minimum value points as a target segment;

[0036] In each target segment, the maximum point is taken as the starting point and each endpoint is taken as the end point, the angle corresponding to each target segment is determined, and the unit vector in the direction of the angle bisector of the angle is taken as the comparison vector;

[0037] Based on the cosine similarity between the comparison vector and the unit vector in the negative direction of the z-axis, a vertical offset corresponding to each target segment is determined, and the vertical offset is negatively correlated with the cosine similarity;

[0038] On each target curve segment, the average of the vertical offsets of the first number of target segments preset before and the first number of target segments preset after is used as the first deviation factor; the average of the vertical offsets of the target segments where the second number of maximum value points closest to the midpoint of the target curve segment are located is used as the second deviation factor;

[0039] A value obtained by normalizing the difference between the first deviation factor and the second deviation factor is used as the clutter level value of each target curve segment.

[0040] Furthermore, the area of ​​all defective areas and concave deformation areas on the surface of the photovoltaic panel is comprehensively determined to determine the flatness of the surface of the photovoltaic panel so as to perform flatness detection, including:

[0041] On the surface of the photovoltaic cell panel, the ratio of the sum of the areas of all the concave deformation areas and all the defect areas to the total area of ​​the photovoltaic cell panel surface is negatively correlated and normalized to obtain a value as the flatness of the photovoltaic cell panel surface;

[0042] When the flatness is less than a preset flatness threshold, the flatness of the photovoltaic panel is judged to be unqualified; when the flatness is greater than or equal to the preset flatness threshold, the flatness of the photovoltaic panel is judged to be qualified.

[0043] The present invention also proposes a photovoltaic panel flatness detection system in photovoltaic construction, the system comprising:

[0044] A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of any one of the methods for detecting the flatness of photovoltaic panels in photovoltaic construction are implemented.

[0045] The present invention has the following beneficial effects:

[0046] First, the point cloud data of the photovoltaic panel surface is obtained. The point cloud data can characterize the morphology of the photovoltaic panel surface, which is helpful for the subsequent identification of protrusions and depressions and provides a data basis for flatness detection. Since the height of the data points in the raised area or the depressed area is relatively large compared to the height of the normal area, the height characteristics of each data point in the point cloud data are analyzed, and the abnormal data points that do not conform to the normal surface morphology are initially screened out to provide key information for the subsequent identification of deformation areas. Since cluster analysis can group data points with similar characteristics into one category, all abnormal data points can be clustered. At the same time, combined with the height information of the data points, the corresponding areas of the clusters on the photovoltaic panel surface are distinguished, and all the raised deformation areas and depressed deformation areas on the photovoltaic panel surface are obtained. Furthermore, since the shapes of the protrusions caused by the texturing operation will have a certain regularity in height, and the shapes of the protrusions caused by defects will have a worse degree of regularity in height, the fluctuations and positional relationships of the height values ​​between the abnormal data points in each protruding deformation area are comprehensively considered to characterize the shape regularity of each protruding deformation area, thereby calculating the probability that each protruding deformation area is a defect area, and based on this indicator, the protruding deformation areas that are truly defective areas are screened out. Finally, the areas of all defective areas and concave deformation areas on the surface of the photovoltaic panel are comprehensively considered to determine the flatness of the surface of the photovoltaic panel and then perform a flatness test. The present invention improves the credibility of the final flatness test result by effectively distinguishing between normal protrusions caused by the texturing operation and protrusions in defective areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A method flow chart of a method for detecting flatness of photovoltaic panels in photovoltaic construction provided by one embodiment of the present invention;

[0049] Figure 2 A method flow chart of a method for obtaining abnormal data points provided by an embodiment of the present invention;

[0050] Figure 3 A schematic diagram of a flooding volume-water injection volume curve provided by an embodiment of the present invention;

[0051] Figure 4 A side view of the shape of a raised deformed area with height variation caused by a texturing operation provided by one embodiment of the present invention;

[0052] Figure 5 A side view of the shape of a convex deformation area caused by a defect in a height change provided by an embodiment of the present invention;

[0053] Figure 6 A method flow chart of a method for obtaining a defect area provided by an embodiment of the present invention;

[0054] Figure 7 A system block diagram of a photovoltaic panel flatness detection system in photovoltaic construction provided by one embodiment of the present invention;

[0055] Figure 8 A schematic diagram of the system structure of a photovoltaic panel flatness detection system in photovoltaic construction provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the photovoltaic panel flatness detection method and system in photovoltaic construction proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0057] 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.

[0058] The specific scheme of a photovoltaic panel flatness detection method and system in photovoltaic construction provided by the present invention is described in detail below with reference to the accompanying drawings.

[0059] See also Figure 1 , which shows a method flow chart of a method for detecting flatness of photovoltaic panels in photovoltaic construction provided by an embodiment of the present invention, the method comprising the following steps:

[0060] Step S1: Acquire point cloud data of the surface of the photovoltaic panel.

[0061] During the photovoltaic construction process, the surface of the photovoltaic panel is prone to uneven areas such as depressions or protrusions due to uneven mechanical stress. These uneven areas will reduce the photoelectric conversion efficiency of the photovoltaic panel, thereby affecting the performance of the photovoltaic panel. Therefore, it is necessary to extract deformation areas such as depressions or protrusions on the surface of the photovoltaic panel, and then detect the flatness of the surface of the photovoltaic panel.

[0062] First, the point cloud information at each position on the surface of the photovoltaic panel can be collected through laser scanning technology. The process generally includes: the laser transmitter emits laser at each position on the surface of the photovoltaic panel. When the laser encounters an object, it is diffusely reflected and returned to the laser receiver. The laser processing system calculates the point cloud data of each reflected laser, and obtains the distance between the transmitter and each position on the surface of the photovoltaic panel according to the time interval between sending and receiving signals, that is, the height value; thus, the point cloud data of the surface of the photovoltaic panel can be obtained. It should be noted that each data point in the point cloud data on the surface of the photovoltaic panel contains three-dimensional coordinates, which can characterize the position information and height information of each data point on the surface of the photovoltaic panel.

[0063] At this point, the point cloud data of the photovoltaic panel surface can be obtained.

[0064] Step S2: According to the height value characteristics of all data points in the point cloud data, the abnormal data points on the surface of the photovoltaic panel are screened out; all the abnormal data points are clustered and analyzed, and all the convex deformation areas and concave deformation areas on the surface of the photovoltaic panel are determined based on the height information of the data points in each cluster.

[0065] Since the height values ​​of the data points in the raised and recessed areas of the photovoltaic panel have more obvious mutations than the height values ​​of the data points in the normal area, the height value characteristics of all data points in the point cloud data can be analyzed to screen out abnormal data points on the surface of the photovoltaic panel. At the same time, since the position distribution of the data points in each raised area and each recessed area is relatively concentrated, cluster analysis can classify data points with similar characteristics (such as height value, position, etc.) into one category to form clusters. At this time, the area corresponding to each cluster on the surface of the photovoltaic panel is the area where deformation occurs. Then, in each cluster, the height value information of the data points can be integrated to determine the raised deformation area and recessed deformation area on the surface of the photovoltaic panel.

[0066] Considering that the watershed model can simulate the water injection process, and as the amount of water injection increases, certain areas in the model (i.e., areas with lower heights) will be submerged by water, applying it to the surface of photovoltaic panels can reflect the concave and convex conditions of the surface of photovoltaic panels. Therefore, in this embodiment of the present invention, based on the height values ​​of all data points in the point cloud data, the watershed model is used to screen out abnormal data points on the surface of photovoltaic panels.

[0067] Preferably, in one embodiment of the present invention, the method for obtaining abnormal data points includes:

[0068] See also Figure 2 , which shows a method flow chart of a method for obtaining abnormal data points in one embodiment of the present invention, the method comprising the following steps:

[0069] Step S201: construct a three-dimensional height watershed model for the point cloud data on the surface of the photovoltaic panel, and construct a flooding volume-water injection volume curve based on the three-dimensional height watershed model.

[0070] Since most of the surfaces of photovoltaic panels are normal areas, the mean of the height values ​​of all data points in the point cloud data on the surface of the photovoltaic panels can be calculated and used as the height mean. The height mean can characterize the general level of the height values ​​of all data points on the surface of the photovoltaic panels. Then, the difference between the height value of each data point on the surface of the photovoltaic panels and the height mean is used as the z-axis coordinate corresponding to each data point, thereby constructing a three-dimensional coordinate system based on the position information of each data point.

[0071] At this point, each data point in the point cloud data on the surface of the photovoltaic panel has a position information (x-axis coordinate and y-axis coordinate) and a z-axis coordinate. All data points in the point cloud data on the surface of the photovoltaic panel can be mapped to a three-dimensional coordinate system, and a three-dimensional height watershed model can be established based on the grid method. Since the z-axis coordinate at this time represents the offset of the height value of each data point compared to the mean height, the constructed three-dimensional height watershed model can also more intuitively represent the high and low positions of each data point on the surface of the photovoltaic panel.

[0072] Because the difference between the height value of each data point and the height mean may be positive or negative, there will be data points distributed on both sides of the xoy plane, and the data points distributed on the positive half axis of the z axis are regarded as data points above the average height level, and the data points distributed on the negative half axis of the z axis are regarded as data points below the average height level. Therefore, in the three-dimensional height watershed model, the water injection process is simulated from the xoy plane along the z axis to both sides of the xoy plane, and the flooding amount of the data points under different water injection amounts is calculated. As time goes by, the water injection amount continues to increase, and then the flooded data points in the three-dimensional height watershed model will increase, so that the flooding amount-water injection amount curve can be obtained, where the water injection amount is the horizontal axis and the flooding amount is the vertical axis. The flooding amount refers to the total number of flooded data points corresponding to the corresponding water injection amount, and the data points that are flooded first are the data points with smaller absolute values ​​on the z axis, that is, the data points with closer vertical distance to the xoy plane.

[0073] It should be noted that establishing a watershed model based on a gridding method is a well-known technology, and the specific process will not be described here.

[0074] Step S202: Obtain the tangent slope value at each water injection volume on the flooding volume-water injection volume curve, and determine the change demarcation point according to the difference between the tangent slope values ​​at different water injection volumes.

[0075] Since most of the data points are at normal heights, the flooding volume will increase rapidly and then slow down as the water injection volume increases. For details, please refer to Figure 3 , which shows a schematic diagram of the flooding volume-injection volume curve (this schematic diagram is for reference only and is used to characterize the trend change of the flooding volume during the water injection process); so on the flooding volume-injection volume curve, when the change trend of the curve changes significantly at a certain injection volume point, it can be regarded as that before this injection volume, the normal data points have been completely flooded, and after this injection volume, the abnormal data points (data points with sudden changes in height values) begin to be flooded, then this injection volume can be used as a change demarcation point to distinguish normal data points from abnormal data points.

[0076] The changing trend can be represented by the tangent slope value at the data point, so on the flooding volume-injection volume curve, the other injection volumes except the first injection volume and the last injection volume are taken as target points.

[0077] Then, for any target point, the difference between the absolute value of the tangent slope value at the next injection volume adjacent to the target point and the tangent slope value at the previous injection volume adjacent to the target point is normalized and used as the change degree value of the target point. The larger the change degree value, the greater the change trend of the curve before and after the target point, and the target point is likely to be a change boundary point. The normalization method here can be used Normalization function. The specific normalization method is not limited here.

[0078] Finally, among all the target points, the target point corresponding to the maximum value of the change degree value is taken as the change demarcation point. It should be noted that if there are multiple equal maximum values, the minimum water injection amount among them is taken as the change demarcation point.

[0079] Step S203: on the flooding amount-water injection amount curve, determine the abnormal data points on the surface of the photovoltaic panel based on the change demarcation point.

[0080] Based on the analysis in step S202, it can be seen that on the flooding amount-water injection amount curve, the change demarcation point is regarded as a point where the change trend of the curve before and after the point has changed significantly, which can be used to distinguish normal data points from abnormal data points. Therefore, the data points flooded after the change demarcation point are selected as abnormal data points on the surface of the photovoltaic panel.

[0081] Since the deformation area on the surface of the photovoltaic panel is divided into convex deformation and concave deformation, and clustering analysis can divide the data points in the data set into multiple clusters, so that the similarity between the data points in the same cluster is higher, clustering results can be obtained by clustering; at the same time, in each cluster, the deformation area is divided into convex deformation area and concave deformation area based on the height information of the data point, providing a basis for the subsequent analysis process.

[0082] Preferably, in one embodiment of the present invention, cluster analysis is performed on all abnormal data points, and all convex deformation areas and concave deformation areas on the surface of the photovoltaic panel are determined based on the height information of the data points in each cluster, including:

[0083] Because the distribution of data points belonging to the same deformation area will be relatively concentrated, in this embodiment of the present invention, all abnormal data points are clustered based on the DBSCAN density clustering algorithm to obtain all cluster clusters, wherein the minimum sample number and neighborhood radius are both preset values.

[0084] Since the data points in each cluster are highly similar and together represent a deformation area on the surface of the photovoltaic panel, the area formed by the data points in each cluster on the surface of the photovoltaic panel is used as the initial deformation area of ​​the surface of the photovoltaic panel. That is, the initial deformation area is determined in the image corresponding to the surface of the photovoltaic panel according to the position information (x-coordinate and y-coordinate) of the data points in each cluster.

[0085] In view of the fact that convex deformation will lead to an increase in height value and concave deformation will lead to a decrease in height value, the numerical mean of all data points in each initial deformation area on the z-axis is taken as the height characteristic value corresponding to each initial deformation area, because the z-axis coordinate represents the offset of the height value of each data point compared to the height mean (the average level of the height of all data points on the surface of the photovoltaic panel), and it may be positive or negative, so the initial deformation area corresponding to the positive height characteristic value is taken as the convex deformation area, and the initial deformation area corresponding to the negative height characteristic value is taken as the concave deformation area.

[0086] It should be noted that, in cluster analysis, the minimum sample size is set to 8, and the neighborhood radius is set to 4. The specific values ​​can be adjusted according to the implementation scenario and are not limited here. The DBSCAN density clustering algorithm is a well-known technology and will not be described in detail here.

[0087] At this point, the concave deformation area and the convex deformation area on the surface of the photovoltaic panel can be obtained.

[0088] Step S3: In each convex deformation area, based on the fluctuation of the height values ​​between the corresponding abnormal data points and the positional relationship between the abnormal data points, determine the probability that each convex deformation area is a defect area and screen out the defect areas in all convex deformation areas.

[0089] During the construction of photovoltaic panels, the surface is prone to depressions or protrusions caused by uneven mechanical stress, thus affecting the flatness of the entire photovoltaic panel surface. However, during its production process, in order to reduce the reflectivity of the surface and increase the absorption of light, thereby improving the photoelectric conversion rate of the photovoltaic panel surface, the surface of the photovoltaic panel will be chemically corroded, also known as a velvet operation, to form numerous pyramid-shaped peak packages on the surface of the photovoltaic panel, also known as velvet structures. These velvet structures can increase the roughness of the surface of the photovoltaic panel, change the incident angle and propagation path of light, and improve the photoelectric conversion rate. After the aforementioned steps, the peak packages caused by the velvet structure will also be identified as convex deformation areas on the surface of the photovoltaic panel. However, the unevenness caused by this convex deformation area is necessary, which is beneficial to improving the performance of the panel. Therefore, it is necessary to distinguish the peak packages caused by the velvet structure from the convexities formed by the defective areas caused by mechanical stress in all convex deformation areas to avoid the problem of large deviations in the subsequent flatness test results.

[0090] Therefore, for the convex deformation area obtained in step S2, it is necessary to further analyze it to determine the source of its unevenness. Figure 4 , which shows a side view of the shape of the protrusion deformation area in height variation caused by the texturing operation in one embodiment of the present invention; and the protrusion caused by the defect will have a worse shape regularity in height, please refer to Figure 5 , which shows a side view of the shape of a convex deformation area caused by a defect in one embodiment of the present invention. Therefore, the change fluctuation and position relationship of the height value between the abnormal data points in each convex deformation area are comprehensively considered to characterize the shape regularity of each convex deformation area in height, so as to calculate the probability of each convex deformation area being a defect area, and based on this indicator, the convex deformation area that is truly a defect area is screened out.

[0091] Preferably, in one embodiment of the present invention, the method for obtaining the defect area includes:

[0092] See also Figure 6 , which shows a method flow chart of a method for obtaining a defect area in one embodiment of the present invention, the method comprising the following steps:

[0093] Step S301: In each convex deformation area, the abnormal data points in the corresponding point cloud data are smoothly connected in a three-dimensional coordinate system to obtain multiple undulating curve segments, and the target curve segment is screened out from all the undulating curve segments based on the position information of the undulating curve segments.

[0094] When determining the undulating curve segments in each convex deformation region, any abnormal data points can be smoothly connected in the three-dimensional coordinate system to obtain multiple undulating curve segments, but it is necessary to ensure that there is at least one maximum point of height on each undulating curve segment, in order to ensure that the shape of the undulating curve segment can be displayed. It should be noted that the number of undulating curve segments can be set to a preset number, such as 20, or each abnormal data point can be used as the starting point and any abnormal data point as the end point, and at least one maximum point of height is obtained on the obtained curve segment, thereby obtaining all non-repeating undulating curve segments in each convex deformation region.

[0095] Because the convex deformation area is often expanded around or close to the geometric center, the geometric center of each convex deformation area is used as a reference point, representing the central position of each convex deformation area.

[0096] Then, for any undulating curve segment corresponding to each convex deformation area, when the projection of the undulating curve segment on the xoy plane passes through the corresponding reference point, that is, through the geometric center, and the projection is still a straight line, it is considered that this undulating curve segment is likely to represent the main boundary line of the height change of the convex deformation area, which is of great significance for the subsequent analysis of the shape formed by the convex deformation area in height, so the undulating curve segment is used as the target curve segment.

[0097] In other embodiments of the present invention, the projection of the geometric center of the three-dimensional figure formed by all abnormal data points in each convex deformation area in the three-dimensional coordinate system on the xoy plane can be used as the reference point corresponding to each convex deformation area.

[0098] Step S302: Analyze the positional relationship between abnormal data points on each target curve segment and the shape characteristics of the target curve segment to determine the clutter level value of each target curve segment.

[0099] In order to analyze the shape characteristics of the target curve segment, we can first obtain all extreme value points on each target curve segment, and take the curve segment between two adjacent minimum value points as a target segment. At this time, the shape of each target segment is a convex peak package.

[0100] Depend on Figure 4 and Figure 5 As shown in the figure, the target segments in the target curve segment caused by the texturing operation are arranged more regularly, and the line passing through the maximum point in the vertical direction can bisect the angle formed by the maximum point and the minimum points on both sides; while the target segments in the target curve segment caused by the defect are arranged more disorderly, and the possibility that the line passing through the maximum point in the vertical direction can bisect the angle formed by the maximum point and the minimum points on both sides is low. Therefore, based on this feature, the disorder degree value of each target curve segment can be determined.

[0101] In each target segment, the maximum point is used as the starting point and each endpoint is used as the end point, so as to determine the angle corresponding to each target segment, and the unit vector pointing downward in the direction of the angle bisector of the angle is used as the comparison vector. It should be noted that the three-dimensional coordinates of the starting point and the two end points are known, and the comparison vector can be obtained by mathematical calculation. The specific calculation process is a well-known technology and will not be described in detail here.

[0102] The vertical line passing through the maximum point can be represented by a unit vector in the negative direction of the z-axis. Therefore, the vertical offset corresponding to each target segment is determined based on the cosine similarity between the comparison vector and the unit vector in the negative direction of the z-axis, and the vertical offset is negatively correlated with the cosine similarity. The formula model of the vertical offset can be, for example:

[0103]

[0104] in, represents the vertical offset corresponding to each target segment; θ represents the cosine similarity between the contrast vector of each target segment and the unit vector in the negative direction of the z-axis; Represents an exponential function with the natural constant e as the base.

[0105] In the formula model of vertical offset, when the cosine similarity between the contrast vector of each target segment and the unit vector in the negative direction of the z-axis is larger, it means that the directions of the two are more similar, then the offset of the contrast vector from the vertical direction is smaller, and the bulge corresponding to the target segment is more regular, which can be regarded as the bulge caused by the velveting operation; conversely, the smaller the cosine similarity value is, the greater the directional deviation between the two, that is, the larger the offset of the contrast vector from the vertical direction, the more chaotic the bulge corresponding to the target segment, which is more likely to be the bulge caused by the defect. Because the value range of cosine similarity is [-1, 1], first adjust its value, and use , the value range is [0, 2], and the closer it is to 2, the smaller the offset between the contrast vector and the vertical direction is. Then Negative correlation mapping and normalization processing are performed to achieve logical relationship correction, so as to obtain the vertical offset corresponding to each target segment. At this time, the smaller the vertical offset, the more regular the bulge corresponding to the target segment is, and it can be regarded as more likely to be a bulge caused by the texturing operation. The larger the vertical offset, the more disorderly the bulge corresponding to the target segment is, and it is more likely to be a bulge caused by a defect.

[0106] For the protrusions caused by defects, the vertical offset of the target segments near the two end points of the target curve segment will be larger, and the vertical offset of the target segments near the middle position will be smaller. Therefore, when traversing from one end point on the target curve segment to the other end point, the vertical offset of the target segment will show a trend of changing from large to small and then increasing; while for the protrusions caused by the texturing operation, the vertical offsets of all target segments on the target curve segment are relatively similar.

[0107] Based on the above characteristics, on each target curve segment, the average of the vertical offsets of the first preset number of target segments and the first preset number of target segments is taken as the first deviation factor, which characterizes the average level of the vertical offsets of the target segments close to the end points on both sides of each target curve segment; the average of the vertical offsets of the target segments where the second preset number of maximum points closest to the midpoint of the target curve segment are located is taken as the second deviation factor, which characterizes the average level of the vertical offsets of the target segments close to the middle position of the target curve segment.

[0108] Then compare the difference between the first deviation factor and the second deviation factor, and calculate the difference between the first deviation factor and the second deviation factor. When the difference is greater than 0, and the larger the difference is, the change in the vertical offset of the target segment on the target curve segment is consistent with the bulge caused by the defect. Because the difference may be positive or negative, it is used here. The function normalizes the difference and uses the normalized value as the clutter value of each target curve segment. At this time, the larger the clutter value, the more likely the target curve segment is to be in the convex deformation area caused by the defect.

[0109] It should be noted that the preset first number is set to 3; the preset second number is set to 3, and the specific values ​​can be adjusted according to the implementation scenario and are not limited here.

[0110] Step S303: In each convex deformation region, the probability of each convex deformation region being a defect region is determined by combining the clutter degree values ​​of all target curve segments and the height value characteristics of abnormal data points on the target curve segments.

[0111] Because the main purpose of the pyramid velvet surface obtained by the velveting operation is to increase the surface area and reduce the reflection loss, its height difference is relatively small and evenly distributed; while the height difference of the raised deformation area caused by the defect is usually larger; therefore, the difference between the maximum height value and the minimum height value of the abnormal data point on each target curve segment is taken as the abnormal factor, and the product of the clutter value of each target curve segment and the abnormal factor is taken as the abnormal degree value of each target curve, because the larger the abnormal factor, the larger the height value of the data point on each curve segment, and the more likely it is to exist in the raised deformation area caused by the defect. At the same time, the larger the clutter value, the more likely the target curve segment is to be in the raised deformation area caused by the defect. Therefore, the larger the final abnormal degree value, the more likely the target curve segment is to be the curve segment corresponding to the raised deformation area caused by the defect.

[0112] Finally, in each convex deformation area, the normalized value of the mean of the abnormality values ​​of all target curve segments is used as the probability that each convex deformation area is a defect area. The larger the probability value, the more likely the convex deformation area is to be a defect area.

[0113] It should be noted that the normalization operation here can be used Normalization function, the specific normalization operation is not limited or elaborated here.

[0114] Step S304: based on the probability that each convex deformation region is a defective region, screening defective regions from among all convex deformation regions.

[0115] The probability that each convex deformation region is a defect region is compared with a preset defect threshold, and when the probability that the convex deformation region is a defect region is greater than the preset defect threshold, the convex deformation region is determined to be a defect region.

[0116] It should be noted that the preset defect threshold is set to 0.65, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0117] At this point, in all raised deformation areas, normal raised areas caused by texturing operations can be excluded, and abnormal raised areas caused by defects can be screened out, effectively improving the credibility of subsequent flatness detection.

[0118] Step S4: Comprehensively consider the areas of all defective regions and concave deformation regions on the surface of the photovoltaic panel to determine the flatness of the surface of the photovoltaic panel and perform flatness detection.

[0119] Based on the above steps, all the concave deformation areas and defective areas on the surface of the photovoltaic panel can be obtained. Therefore, the flatness of the photovoltaic panel surface can be determined by comparing the proportion of these two areas in the entire photovoltaic panel surface, and the photovoltaic panel can be tested for flatness based on the flatness.

[0120] Preferably, in one embodiment of the present invention, the process of obtaining flatness and performing flatness detection includes:

[0121] On the surface of the photovoltaic panel, the ratio of the sum of the areas of all concave deformation areas and all defective areas to the total surface area of ​​the photovoltaic panel is calculated. The larger the ratio, the worse the flatness of the photovoltaic panel surface. Therefore, the ratio is negatively correlated and normalized to the value obtained as the flatness of the photovoltaic panel surface. At this time, the larger the flatness, the higher the flatness of the photovoltaic panel surface.

[0122] When the flatness is less than the preset flatness threshold, the flatness of the photovoltaic panel is judged to be unqualified; conversely, when the flatness is greater than or equal to the preset flatness threshold, the flatness of the photovoltaic panel is judged to be qualified.

[0123] It should be noted that the preset smoothing threshold is set to 0.7, and the specific value can be adjusted according to the implementation scenario and is not limited here; the normalization method can be used Normalization function, the specific normalization operation is not limited or elaborated here.

[0124] In summary, firstly, the point cloud data of the surface of the photovoltaic panel is obtained. The point cloud data can characterize the morphology of the surface of the photovoltaic panel, which is helpful for the subsequent identification of protrusions and depressions and provides a data basis for flatness detection. Since the height of the data points in the raised area or the depressed area is relatively large compared to the height of the normal area, the height characteristics of each data point in the point cloud data are analyzed, and the abnormal data points that do not conform to the normal surface morphology are initially screened out to provide key information for the subsequent identification of deformation areas. Since cluster analysis can group data points with similar characteristics into one category, all abnormal data points can be clustered. At the same time, combined with the height information of the data points, the corresponding areas of the clusters on the surface of the photovoltaic panel are distinguished, and all the raised deformation areas and depressed deformation areas on the surface of the photovoltaic panel are obtained. Furthermore, since the shapes of the protrusions caused by the texturing operation will have a certain regularity in height, and the shapes of the protrusions caused by defects will have a worse degree of regularity in height, the fluctuations and positional relationships of the height values ​​between the abnormal data points in each protruding deformation area are comprehensively considered to characterize the shape regularity of each protruding deformation area, thereby calculating the probability that each protruding deformation area is a defect area, and based on this indicator, the protruding deformation areas that are truly defective areas are screened out. Finally, the areas of all defective areas and concave deformation areas on the surface of the photovoltaic panel are comprehensively considered to determine the flatness of the surface of the photovoltaic panel and then perform a flatness test. The present invention improves the credibility of the final flatness test result by effectively distinguishing between normal protrusions caused by the texturing operation and protrusions in defective areas.

[0125] The present invention also proposes a photovoltaic panel flatness detection system in photovoltaic construction, please refer to Figure 7 , which shows a system block diagram, including: a data acquisition module 701, used to implement step S1 in the above method; a deformation area acquisition module 702, used to implement step S2 in the above method; a defect area screening module 703, used to implement step S3 in the above method; and a flatness detection module 704, used to implement step S4 in the above method.

[0126] It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the photovoltaic panel flatness detection system in photovoltaic construction provided in the above embodiment and the photovoltaic panel flatness detection method embodiment in photovoltaic construction belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0127] See also Figure 8 , which shows a system structure diagram of a photovoltaic panel flatness detection system in photovoltaic construction provided by an embodiment of the present invention, including a processor 800, a memory 801, a bus 802 and a communication interface 803, wherein the processor 800, the communication interface 803 and the memory 801 are connected via the bus 802; wherein the memory 801 may include a high-speed random access memory, the bus 802 may be an ISA bus, a PCI bus or an EISA bus, etc., and the processor 800 may be an integrated circuit chip with signal processing capabilities; the memory 801 stores at least one instruction, at least one program, a code set or an instruction set, and when the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor, the steps in a photovoltaic panel flatness detection method in photovoltaic construction are implemented.

[0128] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0129] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for detecting the flatness of photovoltaic panels in photovoltaic construction, characterized in that: The method comprises: Obtain point cloud data of the photovoltaic panel surface; According to the height value characteristics of all data points in the point cloud data, abnormal data points on the surface of the photovoltaic panel are screened out; all abnormal data points are clustered and analyzed, and all convex deformation areas and concave deformation areas on the surface of the photovoltaic panel are determined according to the height information of the data points in each cluster; In each convex deformation area, based on the fluctuation of the height values ​​between the corresponding abnormal data points and the positional relationship between the abnormal data points, the probability of each convex deformation area being a defect area is determined and the defect areas are screened out from all convex deformation areas; The flatness of the photovoltaic panel surface is determined by comprehensively considering the area of ​​all defective areas and concave deformation areas on the photovoltaic panel surface, so as to conduct flatness detection; The method for obtaining abnormal data points includes: Constructing a three-dimensional height watershed model for the point cloud data on the surface of the photovoltaic panel, and constructing a flooding volume-water injection volume curve based on the three-dimensional height watershed model; Obtaining the tangent slope value at each water injection volume on the flooding volume-water injection volume curve, and determining the change demarcation point according to the difference between the tangent slope values ​​at different water injection volumes; On the flooding amount-water injection amount curve, the data points flooded after the change boundary point are regarded as abnormal data points on the surface of the photovoltaic panel; The step of constructing a three-dimensional height watershed model based on the point cloud data on the surface of the photovoltaic panel, and constructing a flooding volume-water injection volume curve based on the three-dimensional height watershed model, includes: The average of the height values ​​of all data points in the point cloud data of the surface of the photovoltaic panel is used as the height mean, the difference between the height value of each data point and the height mean is used as the z-axis coordinate corresponding to each data point, and a three-dimensional coordinate system is constructed in combination with the position information of the data points; Mapping all data points in the point cloud data of the surface of the photovoltaic panel to the three-dimensional coordinate system, and establishing a three-dimensional height watershed model based on a gridding method; In the three-dimensional height watershed model, the water injection process is simulated from the xoy plane along the z axis to both sides of the xoy plane at the same time, and the flooding amount of the data points under different water injection amounts is calculated, so as to obtain the flooding amount-water injection amount curve; The method for obtaining the defect area comprises: In each convex deformation area, the abnormal data points in the corresponding point cloud data are smoothly connected in the three-dimensional coordinate system to obtain multiple undulating curve segments. Among all the undulating curve segments, the target curve segment is screened out based on the position information of the undulating curve segment; Analyze the positional relationship between abnormal data points on each target curve segment and the shape characteristics of the target curve segment to determine the clutter value of each target curve segment; The difference between the maximum height value and the minimum height value of the abnormal data points on each target curve segment is taken as the abnormal factor, and the product of the clutter value of each target curve segment and the abnormal factor is taken as the abnormal degree value of each target curve; In each convex deformation area, the normalized value of the average of the abnormal degree values ​​of all corresponding target curve segments in the three-dimensional coordinate system is used as the probability that each convex deformation area is a defect area; When the probability that the convex deformation area is a defect area is greater than a preset defect threshold, the convex deformation area is judged to be a defect area; The method for obtaining the clutter level value comprises: On each target curve segment, obtain all extreme value points and take the curve segment between two adjacent minimum value points as a target segment; In each target segment, the maximum point is taken as the starting point and each endpoint is taken as the end point, the angle corresponding to each target segment is determined, and the unit vector in the direction of the angle bisector of the angle is taken as the comparison vector; Based on the cosine similarity between the comparison vector and the unit vector in the negative direction of the z-axis, a vertical offset corresponding to each target segment is determined, and the vertical offset is negatively correlated with the cosine similarity; On each target curve segment, the average of the vertical offsets of the first number of target segments preset before and the first number of target segments preset after is used as the first deviation factor; the average of the vertical offsets of the target segments where the second number of maximum value points closest to the midpoint of the target curve segment are located is used as the second deviation factor; A value obtained by normalizing the difference between the first deviation factor and the second deviation factor is used as the clutter level value of each target curve segment.

2. A photovoltaic panel flatness detection method in photovoltaic construction according to claim 1, characterized in that: The method for obtaining the change demarcation point comprises: On the flooding volume-injection volume curve, other injection volumes except the first injection volume and the last injection volume are taken as target points; On the submerged volume-injection volume curve, for any target point, the normalized value of the difference between the tangent slope value at the next injection volume adjacent to the target point and the tangent slope value at the previous injection volume adjacent to the target point is used as the change degree value of the target point; Among all the target points, the target point corresponding to the maximum value of the change degree is taken as the change demarcation point.

3. The method for detecting flatness of photovoltaic panels in photovoltaic construction according to claim 1, characterized in that: The cluster analysis is performed on all abnormal data points, and all convex deformation areas and concave deformation areas on the surface of the photovoltaic panel are determined based on the height information of the data points in each cluster, including: All abnormal data points are clustered based on the DBSCAN density clustering algorithm to obtain all clusters, where the minimum sample size and neighborhood radius are both preset values; The area formed by the data points in each cluster on the surface of the photovoltaic panel is used as the initial deformation area of ​​the photovoltaic panel surface; The numerical mean of all data points in each initial deformation area on the z-axis is taken as the height characteristic value corresponding to each initial deformation area, the initial deformation area with positive height characteristic value is taken as the convex deformation area, and the initial deformation area with negative height characteristic value is taken as the concave deformation area.

4. The method for detecting flatness of photovoltaic panels in photovoltaic construction according to claim 1, characterized in that: The method for obtaining the target curve segment includes: The geometric center of each convex deformation area is used as a reference point; For any undulating curve segment corresponding to each convex deformation area, when the projection of the undulating curve segment on the xoy plane passes through the corresponding reference point and the projection is a straight line, the undulating curve segment is used as the target curve segment.

5. The method for detecting flatness of photovoltaic panels in photovoltaic construction according to claim 1, characterized in that: The method of comprehensively analyzing the areas of all defective regions and concave deformation regions on the surface of the photovoltaic panel to determine the flatness of the surface of the photovoltaic panel and thus perform flatness detection includes: On the surface of the photovoltaic cell panel, the ratio of the sum of the areas of all the concave deformation areas and all the defect areas to the total area of ​​the photovoltaic cell panel surface is negatively correlated and normalized to obtain a value as the flatness of the photovoltaic cell panel surface; When the flatness is less than a preset flatness threshold, the flatness of the photovoltaic panel is judged to be unqualified; when the flatness is greater than or equal to the preset flatness threshold, the flatness of the photovoltaic panel is judged to be qualified.

6. A photovoltaic panel flatness detection system in photovoltaic construction, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of a method for detecting flatness of photovoltaic panels in photovoltaic construction as described in any one of claims 1 to 5 are implemented.

Images