Method and system for inspecting photovoltaic modules through image analysis

Abstract

The invention relates to a method for inspecting a module of a photovoltaic power plant equipped with a plurality of modules each consisting of a plurality of cells. This method comprises: - processing (SIGNcar) an image (IM) of the module under inspection in order to determine a characteristic signature of the module under inspection; - comparing (COMP) the characteristic signature with reference signatures stored in a database (BdD) that associates a reference signature and a serial number with each of a plurality of reference modules; - when the characteristic signature matches one of the reference signatures, assigning the module under inspection the serial number (IDref) of the reference module whose reference signature matches the specific signature.

Description

[0001] Method and system for inspecting photovoltaic modules by image analysis

[0002] TECHNICAL FIELD

[0003] The field of the invention is that of the inspection of photovoltaic modules by image analysis.

[0004] PRIOR ART

[0005] One technique for characterizing a photovoltaic module involves observing light emitted by electroluminescence from the various cells that make up the module. To do this, a current is injected into the module and a near-infrared image of the module is acquired using a camera. The resulting image shows the quality of the cells and information about any defects within the module.

[0006] Generally speaking, the operator of a photovoltaic solar power plant seeks to source photovoltaic modules from suppliers who offer the best value for money. To this end, they conduct quality audits of suppliers. For the purpose of evaluating samples, they collect electroluminescence images that are taken during the quality control carried out by suppliers on the modules produced.

[0007] Then, when the operator purchases modules, all the images acquired during quality control are transmitted to him. He thus has an electroluminescence image of each of the modules purchased, associated with the serial number or identifier of the module.

[0008] During the construction of the power plant, the modules are deployed without recording in which position a particular module is installed, i.e. a module corresponding to a given serial number. Therefore, during the operation phase of the power plant, it is not possible to know the serial number of a module deployed in a given position, except by going on site to read this number on the back of the module, which is impractical and very time-consuming because a power plant can have more than 100,000 modules.

[0009] To locate the modules in a power plant, one solution would be to use the electronic and communicating devices placed on the back of the modules, whose main function is to optimize the electrical power produced. This solution would require programming each of these devices "by hand" to memorize the serial number of the corresponding module. However, on the one hand, these devices are expensive, particularly in terms of labor. And, on the other hand, this manual reading is a strategy that is far too time-consuming in a power plant that can have more than 100,000 modules.

[0010] It follows that there is currently no solution to ensure the traceability of a module from its production to its operation and thus allow, for example, monitoring of its state of health and, where appropriate, identifying whether defects affecting it are defects linked to its production, its transport, its installation, or its use (aging).

[0011] STATEMENT OF THE INVENTION

[0012] An objective of the invention is to improve the analysis and monitoring of the performance of a photovoltaic solar power plant. To this end, the invention proposes a method for inspecting a module of a photovoltaic power plant provided with a plurality of modules each consisting of a plurality of cells, comprising:

[0013] - processing an image of the inspected module to determine a characteristic signature of the inspected module;

[0014] - comparing the characteristic signature with reference signatures stored in a database which associates with each of a plurality of reference modules a reference signature and an identifier;

[0015] - when the characteristic signature coincides with one of the reference signatures, the assignment to the inspected module of the identifier of the reference module whose reference signature coincides with the specific signature.

[0016] Some preferred but non-limiting aspects of this method are as follows:

[0017] - the characteristic signature of the inspected module is a mapping of characteristics of the image of the inspected module;

[0018] - the processing of the image of the inspected module comprises a segmentation in the image of each of the cells of the inspected module and, for each of the segmented cells, a characteristic mapping, the characteristic signature of the inspected module consisting of the union of the characteristic mappings of the segmented cells;

[0019] - the feature mapping of the image of the inspected module is obtained using an object detection model that has previously been machine-learned using annotated module images;

[0020] - the feature mapping of the image of the inspected module includes a fault mapping of the module;

[0021] - the mapping of characteristics of the image of the inspected module further comprises a mapping of brightness variations in the image of the inspected module;

[0022] - image processing includes pre-processing comprising the successive operations of isolating the inspected module, resizing the inspected module and transforming the inspected module into a rectangle;

[0023] - it also includes the issuance of an alert when the characteristic signature does not coincide with any of the reference signatures;

[0024] - it includes the repetition of the processing step from an image of the inspected module acquired at a later date and the comparison of the characteristic signatures of the inspected module determined at each of the iterations of the processing step;

[0025] - it includes a preliminary step consisting, for each of the reference modules, of processing an image of the reference module to determine the reference signature of the reference module;

[0026] - the image of the inspected module is an electroluminescence image;

[0027] - the image of the inspected module is an image acquired by a drone.

[0028] BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Other aspects, aims, advantages and characteristics of the invention will appear better on reading the following detailed description of preferred embodiments thereof, given by way of non-limiting example, and made with reference to the appended drawings in which: - Figure 1 is a diagram representing different steps implemented in a possible embodiment of the method according to the invention;

[0030] - Figure 2 represents an electroluminescence image acquired by a drone flying over a power plant at different zoom levels;

[0031] - Figure 3 represents different defects likely to be observed by means of an electroluminescence image of a cell;

[0032] - Figure 4 illustrates brightness variations in an electroluminescence image of a module;

[0033] - Figure 5 is a diagram representing a mapping of characteristics imaged by an image of a module.

[0034] DETAILED DESCRIPTION OF SPECIFIC EMBODIMENTS

[0035] The invention relates to a method for inspecting a photovoltaic module of a photovoltaic solar power plant having a plurality of photovoltaic modules each consisting of a plurality of photovoltaic cells. For example, the power plant may comprise on the order of 100,000 to 200,000 modules and each module may consist of 144 cells in a 6*24 cell matrix arrangement.

[0036] In the following, an inspection based on electroluminescence images of modules will be taken as a preferred example. The invention is however not limited to this inspection method and in fact extends to any module imaging method from which it is possible to extract differentiating characteristics of an imaged module and thus to determine a characteristic signature of the imaged module.

[0037] The invention can thus also be implemented by using infrared thermal images, photoluminescence images, synchronous detection thermography images (“Lock-In Thermography” in English), active thermography images, UV fluorescence images or even optical images.

[0038] With reference to Figure 1, the inspection method is preceded by a phase T0 of constituting a reference database BdD from electroluminescence images of reference modules having been previously acquired during an ACQini step, each of these images being associated with an identifier IDref of the corresponding reference module (for example a serial number). The constitution of the reference database BdD comprises processing, during a SIGNref step, of the electroluminescence images to determine a reference signature of each of the reference modules. The database BdD associates, for each reference module, the identifier IRref of the reference module with the reference signature of the reference module.

[0039] The electroluminescence images of the reference modules are typically images provided to the plant operator by a third party. For example, as previously indicated, when a plant operator purchases modules from a supplier, all the electroluminescence images produced during the supplier's quality control are transmitted to it. It thus has an electroluminescence image of each of the modules delivered to it by the supplier, associated with a unique identifier of the module. For example, the name of the file containing the image includes the serial number of the module. The ACQini step is therefore implemented here by the module supplier.

[0040] Alternatively, the ACQini step can be implemented by the operator, for example in the laboratory after receipt of the modules and before their deployment on site.

[0041] In the context of the invention, the inspected module is a module that has been delivered to the operator of the power plant. With reference to Figure 1, an electroluminescence image of the inspected module is acquired by the operator during an ACQinsp step, for example in a laboratory (in particular when the ACQini step is implemented by the third party) or on site in a dark room or outdoors (at night or during the day). For example, the image may be an image representing one or more modules deployed on site acquired by means of a tripod arranged on the supporting structures of the modules or by means of a flying drone. Figure 2 illustrates in this regard an electroluminescence image acquired by a drone flying over a power plant according to different zoom levels.On the IM image zoomed to the scale of a module, we observe cracks probably caused by poor treatment of the module during its manufacture or poor handling of the module during its transport or during its installation on site.

[0042] The module inspection method is implemented during a phase T1 subsequent to the phase T0 of constituting the database BdD, after the modules have been delivered to the operator and before or after their deployment on site. This method comprises the implementation by a processor of a data processing device of steps of obtaining the electroluminescence image representing the inspected module acquired during the ACQinsp step and of processing this image during a SIGNcar step to determine a characteristic signature of the inspected module.

[0043] The process continues with a COMP step of comparing the characteristic signature with the reference signatures stored in the BdD database which associates each of the reference modules with its reference signature and its identifier.

[0044] When the comparison step COMP concludes that the characteristic signature coincides with one of the reference signatures (i.e. these signatures have a coincidence rate greater than a threshold, for example 80%), the method comprises assigning to the inspected module the identifier of the reference module whose reference signature coincides with the specific signature.

[0045] The traceability of the inspected module is therefore ensured, allowing, for example, during the analysis and monitoring of the plant's performance, to go back to the manufacturing batch of a module at the origin of the plant's underperformance. Or, as described below, to monitor the health status of a module.

[0046] The process can further continue by using the identifier thus assigned to the inspected module to identify the inspected module in a map of the modules deployed in the power plant. By repeating the process previously described for each of the modules deployed in the power plant, it then becomes possible to locate any given module using its identifier. In particular, inspecting the modules by drone makes it possible to obtain such a map of the position of the modules with their identifier.

[0047] When the COMP comparison step concludes that the characteristic signature of the inspected module does not match any of the reference signatures, the method includes issuing an alert informing of a risk of fraud. This alert can in particular be issued during a quality control of the modules implemented by the operator after receipt of the modules. The operator can thus ensure that all the inspected modules correspond to the expected modules. Otherwise, the operator may be faced with fraud (for example, the manufacturer provided an image of a module without defects while the inspected module actually has one or more defects) or not (the manufacturer made a mistake by not transmitting the image of the correct module with the correct serial number).To remove any doubt following the issue of the alert, an operator can go to the plant site to note the identifier of the inspected module whose characteristic signature does not coincide with any of the reference signatures and compare the image of the inspected module with the image of the reference module in question (retrieved from the reference database with the identifier of the module noted).

[0048] The method according to the invention may further comprise a new inspection of the module during a phase T2 subsequent to phase T1, for example one or two years later. This new inspection comprises obtaining an image of the inspected module acquired during an ACQinsp* step at the later date, the SIGNcar* processing of this new image to determine a new characteristic signature of the inspected module and then the COMP* comparison of the characteristic signatures determined at phases T1 and T2 of the inspected module determined at each of the iterations of the processing step. If the characteristic signatures determined at phases T1 and T2 do not coincide, the method may comprise the emission of an information alert of a possible deterioration of the module. It should be noted that the assignment to phase T1 of its identifier to the module makes such monitoring over time of the health status of the module possible.This T2 phase can also be repeated over time, for example every year.

[0049] The following details a possible implementation of the determination of the reference signatures of the reference modules and the characteristic signatures of the inspected modules. According to this implementation, a signature of a module is a mapping of characteristics of the module image, specific to the module.

[0050] This feature mapping may include (if applicable consist of) a mapping of defects of the module, these defects being apparent in the image of the module. As illustrated in Figure 3, these defects are for example defects revealed by electroluminescence such as cracks or microcracks or cross-shaped defects C (on the left in Figure 3), black spots T (in the center of Figure 3), ring-type defects R (on the left in Figure 3). The feature mapping may also include a mapping of brightness variations in the image of the inspected module. As shown in Figure 4, this mapping of brightness variations may indicate high-brightness cells CH, low-brightness cells CL or include black edge regions BN at the periphery of cells.

[0051] In an implementation using optical images of the modules, the modules may bear a unique mark, deliberately inscribed on the module by the manufacturer (such as a QR code engraved on the glass at the edge of a module) and visible when the image of the modules is acquired. The signature of a module corresponds in this case to a mapping of image patterns that correspond to this unique mark.

[0052] In order to optimize performance in image processing to determine module signatures, these images are divided into cells and the feature mapping is performed cell by cell. Thus, preferably, the processing of the image of a module (reference module or inspected module) comprises a segmentation in the image of each of the cells of the module and, for each of the segmented cells, a feature mapping as previously described. The feature signature of the inspected module or the reference signature of a reference module then consists of the union of the feature maps of the segmented cells.

[0053] In order to carry out this segmentation, the image of a module (reference or inspected) may be subjected to a pre-processing which comprises the successive operations of isolating the inspected module, resizing the image of the inspected module and transforming this image of the inspected module into a rectangle. For example, this pre-processing comprises an extraction of the parts of the image which concern one or more modules. Each whole module in this extraction is then isolated to provide an image of a single module. The non-whole modules in this extraction are deleted. The image of a module, deformed by the shooting into a parallelogram, is then resized and transformed into a rectangle of pre-established size. The image is then divided into cells, for example according to an automatic division taking into account an expected size for the cells or by identifying the contours of the cells in the image.In one possible embodiment, the mapping of imaged features is obtained using an object detection model (or a semantic segmentation model) that has previously been machine-learned using annotated images of modules and / or cells. The object detection model is, for example, the Yolov8 model trained on thousands of electroluminescence images to detect defects on each cell.

[0054] The resulting map can thus consist of a set of rectangles encompassing each of the characteristics identified in the image. Each rectangle thus records the position and surface of a characteristic while being associated with a characteristic label (or category). Figure 5 illustrates in this respect an example of a mapping of imaged characteristics reported on a matrix of 6*24 cells. In this mapping, the label associated with a characteristic category corresponds to the background pattern of the corresponding bounding rectangle which in this example makes it possible to identify a defect type C, T, R or a brightness type BN, CL, CH. This mapping can be saved in a file which constitutes a form of unique fingerprint of the analyzed module. This mapping can also be superimposed on the image of the module (if applicable the image resulting from the pre-processing).

[0055] A possible implementation of the signature comparison (characteristic signature of an inspected module compared to the reference signatures stored by the database or comparison of characteristic signatures of the same module determined at different moments in time) consists of the intersection over union (loU) measurement which describes the level of overlap (or recovery rate) between the maps, typically between the different bounding rectangles mentioned above.

[0056] This loU measurement may in particular consist of determining whether a given percentage X of characteristics extracted from a first image are found in the characteristics extracted from a second image. Here, the first image is for example an image of a reference module and we seek to verify whether the characteristics extracted from an image of an inspected module are more than X%, for example more than 80%, consistent with the characteristics extracted from the image of the reference module. If so, the signatures associated with the first and second images coincide.

[0057] This loU measurement may include calculating the overlap rate of each feature of the first image and if the average of all overlap rates is greater than X%, then the first and second images are images of the same module. Alternatively, each of the overlap rates must be greater than X% to conclude that the first and second images are images of the same module. In one possible embodiment, overlap rates may be calculated for a given type of feature considered as reference features, for example defects that can only appear during manufacturing.

[0058] Furthermore, in the case where the number of characteristics differs between the first image (for example that of a reference module) and the second image (that of an inspected module), for example characteristics present in the first image are not identified in the second image, it is still possible to consider that the two images coincide and represent the same module if more than Z% of characteristics extracted from the first image are present in the set of characteristics extracted from the second image.

[0059] Alternatively, this loU measurement may consist of determining whether a given percentage Y of characteristics extracted from a first image are not found in the characteristics extracted from a second image. Here, the two images may be those of the same inspected module, identified as such by the identifier assigned to it in accordance with the invention. The first image may be that of the module inspected at a date later than that (earlier date) of the acquisition of the second image. In this case, it is sought to verify whether the characteristics extracted from an image of the module inspected at the later date differ by more than Y%, for example by more than 20%, from the characteristics extracted from the image of the module inspected at the earlier date. If so, the inspected module may be identified as carrying more defects at the later date than at the earlier date.This degradation of the inspected module is further quantified by the result of the loU measurement. Furthermore, in order to improve identification, it is possible to refine this quantification by type of characteristic, for example by type of defect by retaining for example defects that can only appear during the manufacturing of the module (ring-type defects for example) or during its transport or installation (crack-type defects of a certain formation for example), or defects likely to appear during the operation of the plant (for example defects caused by wind or a strong load such as snow). Alternatively, it is possible to exclude certain characteristics (for example manufacturing defects) and only retain certain others (for example defects likely to appear during the operation of the plant).

[0060] The invention is not limited to the method previously described and extends to a system for inspecting a module of a photovoltaic power plant, comprising a processor configured to implement the method previously described as well as to a computer program product comprising instructions which, when executed by a computer, lead the computer to implement the method previously described.

Claims

CLAIMS 1. Method for inspecting a module of a photovoltaic power plant provided with a plurality of modules each consisting of a plurality of cells, comprising: - processing (SIGNcar) of an image (IM) of the inspected module to determine a characteristic signature of the inspected module; - the comparison (COMP) of the characteristic signature with reference signatures stored in a database (BdD) which associates with each of a plurality of reference modules a reference signature and an identifier; - when the characteristic signature coincides with one of the reference signatures, the assignment to the inspected module of the identifier (IDref) of the reference module whose reference signature coincides with the specific signature.

2. Method according to claim 1, in which the characteristic signature of the inspected module is a mapping of characteristics revealed by the image of the inspected module.

3. Method according to claim 2, in which the processing (SIGNcar) of the image of the inspected module comprises a segmentation in the image of each of the cells of the inspected module and, for each of the segmented cells, a mapping of characteristics, the characteristic signature of the inspected module consisting of the union of the mappings of characteristics of the segmented cells.

4. Method according to one of claims 2 and 3, in which the mapping of characteristics revealed by the image of the inspected module is obtained by means of an object detection model which has previously been the subject of machine learning using annotated images of modules.

5. Method according to one of claims 2 to 4, in which the mapping of characteristics revealed by the image of the inspected module comprises a mapping of defects of the module.

6. The method of claim 5, wherein the mapping of characteristics revealed by the image of the inspected module further comprises a mapping of brightness variations in the image of the inspected module.

7. Method according to one of claims 1 to 6, in which the processing of the image comprises a pre-processing comprising the successive operations of isolating the inspected module, resizing the inspected module and transforming the inspected module into a rectangle.

8. Method according to one of claims 1 to 7, further comprising the emission of an alert when the characteristic signature does not coincide with any of the reference signatures.

9. Method according to one of claims 1 to 8, comprising the repetition of the processing step (SIGNcar*) from an image of the inspected module acquired at a later date and the comparison (COMP*) of the characteristic signatures of the inspected module determined at each of the iterations of the processing step.

10. Method according to one of claims 1 to 9, comprising a prior step consisting, for each of the reference modules, in processing an image of the reference module to determine the reference signature of the reference module.

11. Method according to one of claims 1 to 10, in which the image of the inspected module is an electroluminescence image.

12. Method according to one of claims 1 to 11, in which the image of the inspected module is an image acquired by a drone.

13. System for inspecting a module of a photovoltaic power plant, comprising a processor configured to implement the method according to one of claims 1 to 12.

14. Computer program product comprising instructions which, when executed by a computer, lead the computer to implement the method according to one of claims 1 to 12.

Images