A photovoltaic power generation fault detection method and system

Abstract

The present application relates to the field of photovoltaic technology, especially to a photovoltaic power generation fault detection method and system, comprising: acquiring visible light images and infrared thermal imaging images collected synchronously; extracting local shadow geometric contour according to pixel color difference data inside the visible light images; projecting the local shadow geometric contour to the infrared thermal imaging images according to the space mapping ratio to determine the shadow mapping area; generating two-dimensional thermal distribution atlas according to thermal radiation intensity data inside the shadow mapping area, and extracting the peripheral geometric morphology of the discrete thermal spot area; comparing the spatial coincidence degree of the peripheral geometric morphology and the local shadow geometric contour; determining the abnormal micro-warming node according to the comparison result showing the spatial non-coincidence state and outputting the micro-fault early warning information. The present application solves the technical problem of high micro-fault omission rate caused by local shadow covering abnormal data.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic technology, and in particular to a method and system for detecting photovoltaic power generation faults. Background Technology

[0002] With the continuous development of photovoltaic power generation technology and the expanding scale of photovoltaic power plant construction, routine inspections have become a crucial aspect of maintaining the normal operation of the power plant system. Drone inspection technology is widely used in troubleshooting photovoltaic power plants. Drones equipped with thermal imaging equipment fly above photovoltaic modules, collecting infrared images of the module surfaces. Technicians analyze the temperature distribution presented in the infrared images and determine whether there are any faults in the photovoltaic modules based on temperature difference data. The application of drone inspections improves overall work efficiency and reduces manual maintenance costs.

[0003] Drone inspections primarily utilize temperature differences to identify abnormal conditions in photovoltaic (PV) modules. In natural operating environments, cloud cover or foliage can create localized shadows on the PV module surface, causing a significant temperature drop in the shaded area. Microcracks and poor soldering are minor fault types, typically manifesting only as slight temperature increases. When a localized shadow covers an area with a minor fault, the cooling effect of the shadow neutralizes the temperature rise caused by the minor fault, preventing infrared cameras from detecting abnormally high-temperature hotspots. This localized shadow completely obscures PV module anomaly data, preventing the inspection system from extracting minute anomaly features. Existing technologies suffer from a high rate of missed detection of minor faults due to localized shadows obscuring PV module anomaly data. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a photovoltaic power generation fault detection method and system. This invention solves the technical problem of high missed detection rate of minor faults caused by local shading obscuring abnormal data of photovoltaic modules.

[0005] To solve the above-mentioned technical problems, the specific contents of the present invention are as follows:

[0006] In a first aspect, the present invention provides a photovoltaic power generation fault detection method, comprising:

[0007] Acquire visible light images and infrared thermal images of the photovoltaic module surface simultaneously collected by a drone; extract pixel color difference data within the visible light images, and extract the local shadow geometric contours of the photovoltaic module surface based on the pixel color difference data;

[0008] Obtain the spatial mapping ratio between the visible light image and the infrared thermal imaging image, project the local shadow geometric contour onto the infrared thermal imaging image according to the spatial mapping ratio, and determine the corresponding shadow mapping region inside the infrared thermal imaging image.

[0009] Extract the thermal radiation intensity data within the shadowed region, and generate a two-dimensional thermal distribution map within the shadowed region based on the thermal radiation intensity data; extract the isotherms within the two-dimensional thermal distribution map, and extract the heat dissipation spot region within the two-dimensional thermal distribution map based on the distribution direction of the isotherms;

[0010] Extract the outer geometry of the heat dissipation spot area, compare the spatial overlap between the outer geometry and the local shadow geometry, and generate a comparison result.

[0011] Based on the comparison results showing a spatial non-overlapping state, the abnormal micro-heating nodes and their physical coordinates within the shadow mapping area are determined, and micro-fault warning information including the physical coordinates of the abnormal micro-heating nodes is output.

[0012] Furthermore, in the photovoltaic power generation fault detection method of the present invention, the acquisition of visible light images and infrared thermal imaging images of the photovoltaic module surface synchronously collected by the UAV includes:

[0013] Receive the synchronization trigger level signal sent by the UAV;

[0014] Based on the synchronous trigger level signal, the visible light image acquisition channel and the infrared thermal imaging acquisition channel carried by the UAV are simultaneously activated;

[0015] Acquire the visible light image of the photovoltaic module surface captured by the visible light image acquisition channel;

[0016] Extract the thermal radiation distribution state on the surface of the photovoltaic module synchronously recorded by the infrared thermal imaging acquisition channel, and generate the infrared thermal imaging image based on the thermal radiation distribution state;

[0017] The visible light image and the infrared thermal image are associated with the same time signature and spatial physical coordinate reference.

[0018] Furthermore, in the photovoltaic power generation fault detection method of the present invention, the step of extracting pixel color difference data within the visible light image and extracting the local shadow geometric contour of the photovoltaic module surface based on the pixel color difference data includes:

[0019] Extract the grayscale gradient change boundary of the pixels inside the visible light image, and use the grayscale gradient change boundary as the pixel color difference data;

[0020] Based on the grayscale gradient change boundary, the local shadow geometric contour of the photovoltaic module surface is extracted;

[0021] Extract the set of coordinate points of the outer boundary of the local shadow geometric contour.

[0022] Furthermore, in the photovoltaic power generation fault detection method of the present invention, the step of obtaining the spatial mapping ratio between the visible light image and the infrared thermal imaging image, projecting the local shadow geometric contour onto the infrared thermal imaging image according to the spatial mapping ratio, and determining the corresponding shadow mapping region inside the infrared thermal imaging image includes:

[0023] The set of coordinate points of the outer boundary corresponding to the local shadow geometric contour is projected into the infrared thermal image according to the spatial mapping ratio;

[0024] Based on the set of coordinate points of the outer boundary after projection, the corresponding shadow mapping region is determined within the infrared thermal imaging image.

[0025] Furthermore, in the photovoltaic power generation fault detection method of the present invention, the step of extracting thermal radiation intensity data within the shadow mapping region and generating a two-dimensional thermal distribution map within the shadow mapping region based on the thermal radiation intensity data includes:

[0026] Acquire the thermal radiation intensity data of each physical location within the shadow mapping area, and use it as the thermal radiation intensity data;

[0027] Based on the thermal radiation intensity data of each physical location within the shadow mapping region, a corresponding two-dimensional thermal distribution map is generated.

[0028] Furthermore, in the photovoltaic power generation fault detection method of the present invention, the step of extracting isotherms within the two-dimensional thermal distribution map and extracting the heat dissipation spot region within the two-dimensional thermal distribution map based on the distribution direction of the isotherms includes:

[0029] Extract the local heating nodes within the two-dimensional heat distribution map;

[0030] Track the distribution of the isotherms around the localized heating node;

[0031] Based on the distribution of the isotherms, the heat dissipation spot region, including the local heating node, is extracted.

[0032] Furthermore, the photovoltaic power generation fault detection method of the present invention, wherein extracting the peripheral geometry of the area away from the heat dissipation spot, comparing the spatial overlap between the peripheral geometry and the local shadow geometry, and generating a comparison result includes:

[0033] The outer boundary of the heat dissipation spot area is extracted as the peripheral geometry.

[0034] Construct a two-dimensional spatial coordinate system, and input the outer geometric shape and the local shadow geometric outline into the two-dimensional spatial coordinate system;

[0035] Within the two-dimensional coordinate system, a spatial intersection comparison is performed to compare whether the outer geometric shape and the local shadow geometric outline spatially overlap, generating the comparison result, which includes a spatial overlap state or a spatial non-overlap state.

[0036] Furthermore, in the photovoltaic power generation fault detection method of the present invention, determining the abnormally small heating nodes and their physical coordinates within the shadow mapping region based on the comparison results showing a spatial non-overlapping state includes:

[0037] When the comparison result indicates that the space is not overlapping, it is determined that the area away from the heat dissipation spot is completely included within the shadow mapping area.

[0038] Based on the determination result that the heat dissipation spot area is completely included within the shadow mapping area, it is confirmed that the abnormal micro-heating node exists within the shadow mapping area, and the physical coordinate position of the abnormal micro-heating node is extracted.

[0039] Furthermore, in the photovoltaic power generation fault detection method of the present invention, the output of minor fault early warning information including the physical coordinate position of the abnormally small temperature rise node includes:

[0040] Based on the physical coordinates of the abnormally small temperature rise node, the minor fault warning information is generated and output.

[0041] Secondly, the present invention provides a photovoltaic power generation fault detection system, applied to the photovoltaic power generation fault detection method as described above, comprising:

[0042] The image acquisition module is used to acquire visible light images and infrared thermal images of the photovoltaic module surface simultaneously collected by the UAV;

[0043] The contour extraction module is used to extract pixel color difference data inside the visible light image and extract the local shadow geometric contour of the photovoltaic module surface based on the pixel color difference data.

[0044] The region determination module is used to obtain the spatial mapping ratio between the visible light image and the infrared thermal imaging image, project the local shadow geometric contour onto the infrared thermal imaging image according to the spatial mapping ratio, and determine the corresponding shadow mapping region inside the infrared thermal imaging image.

[0045] The map generation module is used to extract thermal radiation intensity data within the shadow mapping region and generate a two-dimensional thermal distribution map within the shadow mapping region based on the thermal radiation intensity data.

[0046] The hot spot extraction module is used to extract isotherms inside the two-dimensional heat distribution map and extract the heat dissipation spot area inside the two-dimensional heat distribution map according to the distribution direction of the isotherms.

[0047] The geometric comparison module is used to extract the outer geometric shape of the area away from the heat dissipation spot, compare the spatial overlap between the outer geometric shape and the local shadow geometric contour, and generate a comparison result.

[0048] The fault warning module is used to determine the abnormal micro-heating nodes and their physical coordinates within the shadow mapping area based on the comparison results showing a spatial non-overlapping state, and output micro-fault warning information including the physical coordinates of the abnormal micro-heating nodes.

[0049] Beneficial effects of this invention:

[0050] This invention establishes an accurate spatial mapping relationship between visible light and infrared thermal imaging images by simultaneously acquiring visible light and infrared thermal imaging images. The processing logic extracts local shadow geometric contours from the visible light image and projects these contours onto the infrared thermal imaging image, thereby defining the corresponding shadow mapping region within the infrared thermal imaging image. The system generates a two-dimensional heat distribution map within the shadow mapping region, and then extracts the peripheral geometry of the heat dissipation spot region. By comparing the spatial overlap between the peripheral geometry and the local shadow geometric contour, when the comparison result shows a spatial non-overlap, the system can accurately identify abnormally small heating nodes under shadow cover. This invention effectively eliminates the overall cooling interference caused by cloud or vegetation obstruction, solves the technical problem of high missed detection rate of minor faults due to local shadow obscuring abnormal data of photovoltaic modules, and improves the accuracy of identification and early warning of minor faults such as microcracks and poor soldering in photovoltaic modules. Attached Figure Description

[0051] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the drawings without creative effort.

[0052] Figure 1 This is a flowchart illustrating a photovoltaic power generation fault detection method according to the present invention. Detailed Implementation

[0053] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Firstly, please refer to Figure 1 The present invention provides a photovoltaic power generation fault detection method, comprising:

[0055] Step 1: Acquire visible light images and infrared thermal images of the photovoltaic module surface simultaneously collected by the drone; extract pixel color difference data within the visible light images, and extract the local shadow geometric contours of the photovoltaic module surface based on the pixel color difference data;

[0056] Step 2: Obtain the spatial mapping ratio between the visible light image and the infrared thermal imaging image, project the local shadow geometric contour onto the infrared thermal imaging image according to the spatial mapping ratio, and determine the corresponding shadow mapping area inside the infrared thermal imaging image.

[0057] Step 3: Extract the thermal radiation intensity data inside the shadowed area, and generate a two-dimensional thermal distribution map inside the shadowed area based on the thermal radiation intensity data; extract the isotherms inside the two-dimensional thermal distribution map, and extract the heat dissipation spot area inside the two-dimensional thermal distribution map based on the distribution direction of the isotherms;

[0058] Step 4: Extract the outer geometry of the heat dissipation spot area, compare the spatial overlap between the outer geometry and the local shadow geometry, and generate a comparison result;

[0059] Step 5: Based on the comparison result showing a spatial non-overlapping state, determine the abnormal micro-heating nodes and their physical coordinates within the shadow mapping area, and output micro-fault warning information including the physical coordinates of the abnormal micro-heating nodes.

[0060] The system simultaneously acquires visible light and infrared thermal imaging images and establishes a spatial binding relationship between them. The processing logic extracts pixel color difference data from the visible light image and generates local shadow geometric contours based on this data. These local shadow geometric contours are then projected onto the infrared thermal imaging image according to a spatial mapping ratio, thereby determining the corresponding shadow mapping region within the infrared thermal imaging image. Within the boundaries defined by the shadow mapping region, the system extracts thermal radiation intensity data and generates a two-dimensional thermal distribution map. Subsequently, the extraction logic traces the isotherm distribution within the two-dimensional thermal distribution map and extracts the area away from the heat dissipation spot. The verification module acquires the outer geometry of the area away from the heat dissipation spot and compares the spatial overlap between the outer geometry and the local shadow geometric contour. When the comparison result shows a spatial non-overlap, the system determines that a hidden fault has occurred, further identifying the abnormally small heating nodes and their physical coordinates within the shadow mapping region. Finally, it outputs a minor fault warning message including the physical coordinates of the abnormally small heating nodes.

[0061] Image source acquisition and association are crucial steps in obtaining synchronously acquired images from the UAV. The receiving node acquires the synchronization trigger level signal sent by the UAV; when the synchronization trigger level signal undergoes a state transition, the visible light image acquisition channel and the infrared thermal imaging acquisition channel onboard the UAV are simultaneously activated. The visible light image acquisition channel captures the visible light image of the photovoltaic module surface; simultaneously, the infrared thermal imaging acquisition channel records the thermal radiation distribution state of the photovoltaic module surface and generates an infrared thermal image based on the thermal radiation distribution state. The data alignment logic reads the system clock parameters and UAV spatial positioning parameters at the moment of sensor triggering, and then associates the visible light image and the infrared thermal image with the same time stamp and spatial physical coordinate reference, thereby eliminating spatial misalignment caused by acquisition delay or UAV displacement.

[0062] The feature separation logic is used to extract pixel color difference data within the visible light image and extract local shadow geometric contours based on this data. The processing module scans the visible light image line by line, comparing the grayscale differences between adjacent pixels. When the grayscale difference between adjacent pixels crosses a preset contrast limit, the system extracts the grayscale gradient change boundary of the pixels within the visible light image and uses this boundary as pixel color difference data. Based on the closed boundary contour formed by the pixel color difference data, the system extracts the local shadow geometric contour of the photovoltaic module surface within the visible light image. The edge extraction node performs coordinate point extraction along the outer edge of the local shadow geometric contour, ultimately obtaining the set of coordinate points for the outer boundary of the local shadow geometric contour.

[0063] The cross-layer projection processing path is responsible for calculating the spatial mapping ratio and determining the shadow mapping area. The control node reads the resolution specifications of the visible light image acquisition channel and the infrared thermal imaging acquisition channel, and establishes the spatial mapping ratio based on the difference in the number of pixels in the horizontal and vertical directions according to the resolution specifications. The system performs a proportional scaling and translation operation on the outer boundary coordinate point set corresponding to the local shadow geometric contour according to the spatial mapping ratio, thereby accurately projecting the outer boundary coordinate point set into the infrared thermal imaging image. Based on the projected outer boundary coordinate point set, the system generates a mask boundary with closed properties within the infrared thermal imaging image, thus accurately determining the corresponding shadow mapping area within the infrared thermal imaging image.

[0064] The construction logic of the thermal mapping structure realizes the generation process of the two-dimensional thermal distribution map. Data extraction nodes traverse each physical location within the shadowed mapping region, sequentially reading the thermal radiation intensity data recorded by the thermal imaging sensor at each physical location. The system then fills the blank two-dimensional matrix structure with the thermal radiation intensity data from each physical location within the shadowed mapping region, following the original spatial arrangement. The height distribution of data nodes within the two-dimensional matrix structure directly corresponds to the changes in the intensity of the thermal radiation data. Based on the filled two-dimensional matrix structure, the system generates the corresponding two-dimensional thermal distribution map, completing the structured data extraction of local temperature anomaly trends.

[0065] Hotspot isolation and clustering logic are used to extract heat-dissipating spot regions within a 2D thermal distribution map. Search nodes perform extreme value range matching within the 2D thermal distribution map to identify continuous pixel regions with temperature amplitudes higher than the surrounding background, extracting locally heated nodes within the map. The system uses these locally heated nodes as the spatial origin and performs a surrounding search, connecting data points with equal thermal radiation intensity around each node to form isotherm distribution patterns. Following the isotherm distribution patterns, the system performs boundary stripping on the regions within continuous and closed isotherms, extracting heat-dissipating spot regions including locally heated nodes, thereby isolating minute anomalous heat sources from the normal shaded cooling background.

[0066] The spatial topology verification mechanism compares the spatial overlap between the outer geometric shape and the local shadow geometric contour through geometric analysis logic. The processing module extracts the outermost boundary of the heat dissipation spot region as the outer geometric shape and constructs a unified two-dimensional spatial coordinate system. The vertex coordinate set of the outer geometric shape and the vertex coordinate set of the local shadow geometric contour are input into the two-dimensional spatial coordinate system. The system performs spatial intersection comparison within the two-dimensional spatial coordinate system, comparing each coordinate point of the outer geometric shape according to boundary coordinate matching logic to see if it falls within the boundary coordinate point set corresponding to the local shadow geometric contour, or falls outside the local shadow geometric contour's outer region. By comparing whether the outer geometric shape and the local shadow geometric contour spatially overlap, the system generates a comparison result; the comparison result outputs a fixed spatial overlap state or a spatial non-overlap state.

[0067] The conditional derivation and localization logic determines the physical coordinates based on the spatial non-overlapping state. The state judgment branch receives the comparison results output by the spatial intersection comparison node; when the comparison result is a spatial non-overlapping state, the system determines that the outer boundary of the heat dissipation spot area is completely within the boundary of the local shadow geometric contour, and further determines that the heat dissipation spot area is completely included within the shadow mapping area. Based on the determination that the heat dissipation spot area is completely included within the shadow mapping area, the system confirms the existence of an abnormally small heating node within the shadow mapping area. The above nested state indicates that the local shadow masks the minor fault heating phenomenon; the system then locates the index position of the abnormally small heating node within the image matrix and extracts the physical coordinates of the abnormally small heating node.

[0068] The data encapsulation and distribution mechanism executes the output of fault warning information. The communication node extracts the physical coordinates of the abnormally small temperature rise nodes within the data queue. Based on the physical coordinate data, the system concatenates the physical coordinate information with the fault type identifier field to generate a standardized data format for the minor fault warning information. The control terminal outputs the minor fault warning information through the transmission link and sends a warning signal to the front-end display device to execute alarm prompts.

[0069] Secondly, the present invention provides a photovoltaic power generation fault detection system, applied to the photovoltaic power generation fault detection method as described above, comprising:

[0070] The image acquisition module is used to acquire visible light images and infrared thermal images of the photovoltaic module surface simultaneously collected by the UAV;

[0071] The contour extraction module is used to extract pixel color difference data inside the visible light image and extract the local shadow geometric contour of the photovoltaic module surface based on the pixel color difference data.

[0072] The region determination module is used to obtain the spatial mapping ratio between the visible light image and the infrared thermal imaging image, project the local shadow geometric contour onto the infrared thermal imaging image according to the spatial mapping ratio, and determine the corresponding shadow mapping region inside the infrared thermal imaging image.

[0073] The map generation module is used to extract thermal radiation intensity data within the shadow mapping region and generate a two-dimensional thermal distribution map within the shadow mapping region based on the thermal radiation intensity data.

[0074] The hot spot extraction module is used to extract isotherms inside the two-dimensional heat distribution map and extract the heat dissipation spot area inside the two-dimensional heat distribution map according to the distribution direction of the isotherms.

[0075] The geometric comparison module is used to extract the outer geometric shape of the area away from the heat dissipation spot, compare the spatial overlap between the outer geometric shape and the local shadow geometric contour, and generate a comparison result.

[0076] The fault warning module is used to determine the abnormal micro-heating nodes and their physical coordinates within the shadow mapping area based on the comparison results showing a spatial non-overlapping state, and output micro-fault warning information including the physical coordinates of the abnormal micro-heating nodes.

[0077] In this embodiment of the invention, during a drone inspection of a large-scale mountain photovoltaic power station, when the surface of the photovoltaic module is obscured by surrounding vegetation, the drone sends a synchronization trigger level signal to its onboard image acquisition device. This synchronization trigger level signal includes a high / low level transition command to control the shutter action. The visible light image acquisition channel and the infrared thermal imaging acquisition channel are activated synchronously according to the transition command, capturing visible light and infrared thermal images of the photovoltaic module surface, respectively. The processing system traverses the visible light image and extracts the grayscale gradient change boundaries of pixels as pixel color difference data. Specifically, the pixel color difference data represents the color contrast boundaries where the brightness of the red, green, and blue channels in the pixel array abruptly changes. Based on these color contrast boundaries, the processing system extracts the geometric contour of the local shadow formed by the foliage obstruction, as well as the set of coordinate points of the outer boundary of the local shadow geometric contour.

[0078] Subsequently, the system determines the spatial mapping ratio based on the lens focal length and target resolution specifications of the visible light image acquisition channel and the infrared thermal imaging acquisition channel. The system scales and translates the set of coordinate points corresponding to the outer boundary of the local shadow geometric contour proportionally according to the spatial mapping ratio, projecting it onto the infrared thermal imaging image to delineate the corresponding shadow mapping region within the infrared thermal imaging image. When the slight temperature rise caused by the microcrack fault is masked by the overall cooling effect of the leaf shadow, the system extracts the thermal radiation intensity data of each physical location within the shadow mapping region, generating a two-dimensional thermal distribution map within the shadow mapping region. The processing logic further tracks the distribution trend of isotherms within the two-dimensional thermal distribution map in the shadow cooling background, extracting the region where the uniform low-temperature distribution law of the shadow is broken due to the impedance heating of the microcrack as the heat dissipation spot region.

[0079] The geometric comparison module extracts the outer boundary of the heat dissipation spot area as the peripheral geometry and compares it with the local shadow geometry within a constructed two-dimensional coordinate system. When the microcrack heating node is completely located within the area obscured by leaves, the boundary between the peripheral geometry and the local shadow geometry exhibits a completely non-overlapping spatial state. Based on this non-overlapping spatial state, the system determines that the area away from the heat dissipation spot is not a temperature gradient artifact caused by uneven illumination, but rather an abnormally small heating node completely contained within the shadow mapping area. The system ultimately extracts the true spatial coordinates of the abnormally small heating node and generates a micro-fault warning message including latitude and longitude information and the module array number. The output terminal sends the micro-fault warning message through a data link, driving the ground monitoring and control terminal to execute an alarm prompt, thereby solving the technical problem of missed detection of micro-faults in photovoltaic modules caused by partial shadowing from vegetation or clouds.

Claims

1. A method for detecting photovoltaic power generation faults, characterized in that, include: Acquire visible light images and infrared thermal images of the photovoltaic module surface simultaneously collected by a drone; extract pixel color difference data within the visible light images, and extract the local shadow geometric contours of the photovoltaic module surface based on the pixel color difference data; Obtain the spatial mapping ratio between the visible light image and the infrared thermal imaging image, project the local shadow geometric contour onto the infrared thermal imaging image according to the spatial mapping ratio, and determine the corresponding shadow mapping region inside the infrared thermal imaging image. Extract the thermal radiation intensity data within the shadowed region, and generate a two-dimensional thermal distribution map within the shadowed region based on the thermal radiation intensity data; extract the isotherms within the two-dimensional thermal distribution map, and extract the heat dissipation spot region within the two-dimensional thermal distribution map based on the distribution direction of the isotherms; Extract the outer geometry of the heat dissipation spot area, compare the spatial overlap between the outer geometry and the local shadow geometry, and generate a comparison result. Based on the comparison results showing a spatial non-overlapping state, the abnormal micro-heating nodes and their physical coordinates within the shadow mapping area are determined, and micro-fault warning information including the physical coordinates of the abnormal micro-heating nodes is output.

2. The photovoltaic power generation fault detection method according to claim 1, characterized in that, The acquisition of visible light images and infrared thermal images of the photovoltaic module surface synchronously collected by the UAV includes: Receive the synchronization trigger level signal sent by the UAV; Based on the synchronous trigger level signal, the visible light image acquisition channel and the infrared thermal imaging acquisition channel carried by the UAV are simultaneously activated; Acquire the visible light image of the photovoltaic module surface captured by the visible light image acquisition channel; Extract the thermal radiation distribution state on the surface of the photovoltaic module synchronously recorded by the infrared thermal imaging acquisition channel, and generate the infrared thermal imaging image based on the thermal radiation distribution state; The visible light image and the infrared thermal image are associated with the same time signature and spatial physical coordinate reference.

3. The photovoltaic power generation fault detection method according to claim 2, characterized in that, The step of extracting pixel color difference data within the visible light image and extracting the local shadow geometric contour of the photovoltaic module surface based on the pixel color difference data includes: Extract the grayscale gradient change boundary of the pixels inside the visible light image, and use the grayscale gradient change boundary as the pixel color difference data; Based on the grayscale gradient change boundary, the local shadow geometric contour of the photovoltaic module surface is extracted; Extract the set of coordinate points of the outer boundary of the local shadow geometric contour.

4. The photovoltaic power generation fault detection method according to claim 3, characterized in that, The step of obtaining the spatial mapping ratio between the visible light image and the infrared thermal imaging image, projecting the local shadow geometric contour onto the infrared thermal imaging image according to the spatial mapping ratio, and determining the corresponding shadow mapping region inside the infrared thermal imaging image includes: The set of coordinate points of the outer boundary corresponding to the local shadow geometric contour is projected into the infrared thermal image according to the spatial mapping ratio; Based on the set of coordinate points of the outer boundary after projection, the corresponding shadow mapping region is determined within the infrared thermal imaging image.

5. A photovoltaic power generation fault detection method according to claim 4, characterized in that, The step of extracting thermal radiation intensity data within the shadowed region and generating a two-dimensional thermal distribution map of the shadowed region based on the thermal radiation intensity data includes: Acquire the thermal radiation intensity data of each physical location within the shadow mapping area, and use it as the thermal radiation intensity data; Based on the thermal radiation intensity data of each physical location within the shadow mapping region, a corresponding two-dimensional thermal distribution map is generated.

6. The photovoltaic power generation fault detection method according to claim 5, characterized in that, The step of extracting isotherms from the two-dimensional heat distribution map and extracting the heat dissipation spot region from the two-dimensional heat distribution map based on the distribution direction of the isotherms includes: Extract the local heating nodes within the two-dimensional heat distribution map; Track the distribution of the isotherms around the localized heating node; Based on the distribution of the isotherms, the heat dissipation spot region, including the local heating node, is extracted.

7. A photovoltaic power generation fault detection method according to claim 6, characterized in that, The step of extracting the peripheral geometry of the heat dissipation spot region and comparing the spatial overlap between the peripheral geometry and the local shadow geometry to generate a comparison result includes: The outer boundary of the heat dissipation spot area is extracted as the peripheral geometry. Construct a two-dimensional spatial coordinate system, and input the outer geometric shape and the local shadow geometric outline into the two-dimensional spatial coordinate system; Within the two-dimensional coordinate system, a spatial intersection comparison is performed to compare whether the outer geometric shape and the local shadow geometric outline spatially overlap, generating the comparison result, which includes a spatial overlap state or a spatial non-overlap state.

8. A photovoltaic power generation fault detection method according to claim 7, characterized in that, The step of determining the abnormally small heating nodes and their physical coordinates within the shadow mapping region based on the comparison results showing a spatial non-overlapping state includes: When the comparison result indicates that the space is not overlapping, it is determined that the area away from the heat dissipation spot is completely included within the shadow mapping area. Based on the determination result that the heat dissipation spot area is completely included within the shadow mapping area, it is confirmed that the abnormal micro-heating node exists within the shadow mapping area, and the physical coordinate position of the abnormal micro-heating node is extracted.

9. A photovoltaic power generation fault detection method according to claim 8, characterized in that, The output includes minor fault warning information regarding the physical coordinates of the abnormally small temperature rise node, including: Based on the physical coordinates of the abnormally small temperature rise node, the minor fault warning information is generated and output.

10. A photovoltaic power generation fault detection system, applied to the photovoltaic power generation fault detection method as described in any one of claims 1 to 9, characterized in that, include: The image acquisition module is used to acquire visible light images and infrared thermal images of the photovoltaic module surface simultaneously collected by the UAV; The contour extraction module is used to extract pixel color difference data inside the visible light image and extract the local shadow geometric contour of the photovoltaic module surface based on the pixel color difference data. The region determination module is used to obtain the spatial mapping ratio between the visible light image and the infrared thermal imaging image, project the local shadow geometric contour onto the infrared thermal imaging image according to the spatial mapping ratio, and determine the corresponding shadow mapping region inside the infrared thermal imaging image. The map generation module is used to extract thermal radiation intensity data within the shadow mapping region and generate a two-dimensional thermal distribution map within the shadow mapping region based on the thermal radiation intensity data. The hot spot extraction module is used to extract isotherms inside the two-dimensional heat distribution map and extract the heat dissipation spot area inside the two-dimensional heat distribution map according to the distribution direction of the isotherms. The geometric comparison module is used to extract the outer geometric shape of the area away from the heat dissipation spot, compare the spatial overlap between the outer geometric shape and the local shadow geometric contour, and generate a comparison result. The fault warning module is used to determine the abnormal micro-heating nodes and their physical coordinates within the shadow mapping area based on the comparison results showing a spatial non-overlapping state, and output micro-fault warning information including the physical coordinates of the abnormal micro-heating nodes.

Images