A photovoltaic power station monitoring method and system based on video sensing and operation data fusion

Abstract

This invention discloses a photovoltaic power plant monitoring method and system based on video perception and operational data fusion, belonging to the field of photovoltaic monitoring technology. The method includes the following steps: identifying shading areas on the component based on video images and determining the projection position of the shading area on the target component; determining the sub-string crossing order corresponding to the shading boundary based on the positional relationship between the projection position and the sub-string boundary line inside the component; constructing a bypass response reference sequence for the target string based on the sub-string crossing order and the sub-strings and bypass diodes; extracting actual voltage step events from the voltage change process of the target string, and outputting the shading electrical response monitoring result of the target string based on the correspondence between the actual voltage step events and the voltage step reference sequence. This invention solves the problem of difficult monitoring of sub-string-level bypass diode operation by identifying the shading crossing order through video and constructing a voltage step reference sequence, which is then compared with the measured voltage.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic monitoring technology, and more specifically, to a photovoltaic power plant monitoring method and system based on the fusion of video perception and operational data. Background Technology

[0002] In current photovoltaic power plant monitoring, video and operational data are independent: video can only show changes in the shading position on the module surface, failing to reflect the geometric relationship between sub-string boundaries and shading within the module; the operational side uses the string as the smallest unit for deviation judgment, making it difficult to distinguish subtle nonlinear electrical changes caused when shading boundaries cross sub-strings. Therefore, when shading dynamically crosses the sub-string boundary line, existing systems cannot effectively identify whether the corresponding bypass diodes are turned on or off in the correct order, creating a monitoring blind spot.

[0003] Each sub-string within a photovoltaic module is connected in parallel with a bypass diode. When shading causes current mismatch in the sub-string, the diode should automatically bypass to maintain the circuit and protect the cells. Even a slight drift at the shading boundary is enough to cause the diode to switch states, instantly altering the string characteristics. If the diode malfunctions, it not only exacerbates power loss but may also cause safety hazards such as hot spots.

[0004] The above-disclosed technical solutions have at least the following technical problems: In the existing photovoltaic power plant monitoring, the video can only see the change in the position of the shading on the surface of the module. The operation side usually makes anomaly judgments based only on the output deviation of string power, current and other parameters. It is difficult to identify whether the corresponding bypass diodes are switched in or out in the correct order after the shading boundary crosses the sub-string boundary line of the module. Therefore, it is impossible to effectively monitor the actual electrical response caused by sub-string shading. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a photovoltaic power plant monitoring method and system based on the fusion of video perception and operational data. The method identifies the projected position of the shading area through video images, determines the sub-string crossing order based on the positional changes of the sub-string boundary line, and then constructs a reference sequence containing the timing of bypass events and voltage steps by combining the correspondence between the sub-strings and bypass diodes. This reference sequence is then compared and verified with actual voltage step events to solve the problem of electrical response blind spots in existing monitoring systems that cannot identify whether bypass diodes operate correctly under sub-string-level shading.

[0006] To achieve the above objectives, the present invention provides the following technical solution: On one hand, a photovoltaic power plant monitoring method based on video perception and operational data fusion includes the following steps: identifying shading areas on the component based on video images and determining the projection position of the shading area on the target component; determining the sub-string crossing order corresponding to the shading boundary based on the positional relationship between the projection position and the sub-string boundary line inside the component; constructing a bypass response reference sequence for the target string based on the sub-string crossing order and the sub-string and bypass diode, wherein the bypass response reference sequence includes a bypass event timing sequence and a corresponding voltage step reference sequence; extracting actual voltage step events from the voltage change process of the target string, and outputting the shading electrical response monitoring result of the target string based on the correspondence between the actual voltage step events and the voltage step reference sequence.

[0007] In a preferred embodiment, the step of identifying the occlusion area on the video image recognition component and determining the projection position of the occlusion area on the target component includes: locating the image region corresponding to the target component from the video image, and identifying the occlusion area on the component surface based on the brightness difference, texture difference, and edge change between the image region and a preset unoccluded reference image; establishing a mapping relationship between the position of the occlusion area in the video image and the position on the surface of the target component based on camera calibration parameters, the boundary position of the target component in the video image, and the actual installation position of the target component, and determining the projection position of the occlusion area on the target component.

[0008] In a preferred embodiment, determining the projection position of the occluded region on the target component includes: determining the boundary position of the target component in the video image, and establishing a mapping relationship from image coordinates to component surface coordinates based on the correspondence between the boundary position of the target component in the video image and the reference coordinates of the target component surface; extracting the boundary point set of the occluded region, and converting the image coordinates of the boundary point set into target component surface coordinates through the mapping relationship, and using the area range corresponding to the surface coordinates as the projection position of the occluded region on the target component.

[0009] In a preferred embodiment, determining the substring crossing order corresponding to the occlusion boundary based on the positional relationship between the projection position and the substring boundary line within the component includes: determining the position of each substring boundary line in the target component surface coordinate system according to the substring division parameters corresponding to the target component model; extracting the position change trajectory of the occlusion area boundary in the target component surface coordinate system in consecutive image frames; determining whether a crossing occurs between the position change trajectory and each substring boundary line, and recording the corresponding substring crossing event when a crossing occurs; and generating the substring crossing order corresponding to the occlusion boundary based on the time sequence of each substring crossing event.

[0010] In a preferred embodiment, determining whether a crossing occurs between the position change trajectory and the substring boundary line includes: extracting the set of position points corresponding to the occlusion boundary at consecutive time intervals, and determining the side of each position point relative to the substring boundary line; determining that a crossing occurs between the position change trajectory and the substring boundary line when the corresponding position point changes from one side of the same substring boundary line to the other side in two consecutive time intervals, or when the set of position points at the same time interval is simultaneously distributed on both sides of the substring boundary line; and recording the corresponding substring crossing event when the crossing state continues to reach a preset number of frames or a preset time threshold.

[0011] In a preferred embodiment, generating the substring crossing order corresponding to the occlusion boundary based on the time sequence of each substring crossing event includes: recording the moment when each substring crossing event first meets the crossing judgment condition as the event moment of the corresponding substring crossing event, and establishing the association between the event moment and the corresponding substring boundary marker; sorting the substring crossing events from front to back according to the event moments to obtain the crossing sequence of the occlusion boundary to each substring boundary; and generating the substring crossing order corresponding to the occlusion boundary based on the crossing sequence.

[0012] In a preferred embodiment, constructing a bypass response reference sequence for the target string based on the substring crossover order and the substrings and bypass diodes includes: dividing the change process of the shading boundary on the target component into multiple time segments according to the substring crossover order; for each time segment, determining the corresponding bypass response reference combination based on the set of target substrings within the shading range and the correspondence between each substring and the bypass diode; determining the corresponding bypass entry event and bypass exit event based on the change of the bypass response reference combination between adjacent time segments; and generating a corresponding voltage step reference sequence according to the order of each bypass entry event and bypass exit event to obtain the bypass response reference sequence for the target string.

[0013] In a preferred embodiment, the step of dividing the change process of the occlusion boundary on the target component into multiple time segments according to the substring crossing order includes: recording the event times corresponding to each substring entry event and each substring exit event; using the time interval between two adjacent event times, the time interval between the start time and the first event time, and the time interval between the last event time and the end time as continuous time segments; determining the set of target substrings within the occlusion range in each continuous time segment; when the set of target substrings corresponding to adjacent time segments is increased, determining the next time segment as the time segment where the substring enters the occlusion; when the set of target substrings corresponding to adjacent time segments is decreased, determining the next time segment as the time segment where the substring exits the occlusion.

[0014] In a preferred embodiment, the step of outputting the electrical response monitoring result of the target string's obstruction includes: sorting the extracted actual voltage step events according to their event times to form an actual voltage step sequence; comparing the actual voltage step sequence with a voltage step reference sequence to determine whether the order of occurrence of each step event in the actual voltage step sequence is consistent with the voltage step reference sequence; when the actual voltage step sequence is consistent with the voltage step reference sequence, outputting a monitoring result indicating that the electrical response of the target string's obstruction is normal; when the actual voltage step sequence is inconsistent with the voltage step reference sequence, outputting a monitoring result indicating that the electrical response of the target string's obstruction is abnormal.

[0015] On the other hand, a photovoltaic power plant monitoring system based on video perception and operational data fusion includes the following modules: a shading projection determination module, used to identify shading areas on the component based on video images and determine the projection position of the shading area on the target component; a cross-boundary sequence generation module, used to determine the cross-boundary sequence of sub-strings corresponding to the shading boundary based on the positional relationship between the projection position and the sub-string boundary line inside the component; a bypass response reference construction module, used to construct a bypass response reference sequence of the target string based on the sub-string cross-boundary sequence and the sub-strings and bypass diodes, wherein the bypass response reference sequence includes a bypass event timing sequence and its corresponding voltage step reference sequence; and a response determination module, used to extract actual voltage step events from the voltage change process of the target string and output the shading electrical response monitoring result of the target string based on the correspondence between the actual voltage step events and the voltage step reference sequence.

[0016] The technical effects and advantages of the photovoltaic power plant monitoring method and system based on video perception and operational data fusion of the present invention are as follows: This invention, through deep fusion of video perception and operational data, maps the geometric position of component surface shading to the substring cross-boundary sequence, and constructs an expected electrical fingerprint containing bypass event timing and voltage steps. This fingerprint is then dynamically compared and verified with step events extracted from the actual operating voltage to accurately determine whether the bypass diodes are correctly switched in or out in sequence. This effectively solves the blind spot problem of traditional decoupling monitoring being unable to identify the electrical response of substring-level shading, significantly improving the accuracy and timeliness of shading fault diagnosis. It is of great value for preventing local hot spots, reducing mismatch losses, and ensuring the safe and economical operation of power plants. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a photovoltaic power plant monitoring method based on the fusion of video perception and operational data according to the present invention. Figure 2 This is a schematic diagram of the structure of a photovoltaic power plant monitoring system based on the fusion of video perception and operational data according to the present invention; Figure 3A schematic diagram of the projection of the occluded area onto the surface coordinate system of the target component. Figure 4 A timing diagram illustrating the occlusion of boundaries crossing substring boundaries; Figure 5 A schematic diagram for generating the bypass response reference sequence. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, Figure 1 This invention presents a photovoltaic power plant monitoring method based on the fusion of video perception and operational data, comprising the following steps: S1, determine the projection position of the occlusion area on the target component based on the occlusion area on the video image recognition component; The video images are visible light video images of the array surface obtained through continuous monitoring of the target photovoltaic array. These video images can be acquired by a video acquisition device pre-installed on the side of the photovoltaic array channel, in front of the array, above the support, or other locations capable of covering the target component area. The video acquisition device is preferably a fixed industrial camera, a PTZ camera, or an in-station monitoring camera. In some embodiments, inspection drones or mobile inspection terminals may also be used to supplement the acquisition, but a fixed video acquisition device is preferred to continuously obtain a video sequence of the component surface from the same viewpoint.

[0020] In this embodiment, determining the projection position of the occlusion area on the target component based on the occlusion area on the video image recognition component includes: The image region corresponding to the target component is located in the video image. The image region is then denoised and its brightness is normalized. Based on the brightness difference, texture difference, and edge change between the image region and the preset unobstructed reference image, the occlusion area on the surface of the component is identified. The preset unobstructed reference image can be a reference image pre-collected by the target component under unobstructed conditions, or an unobstructed image frame selected from historical video sequences. Based on the camera calibration parameters, the boundary position of the target component in the video image, and the actual installation position of the target component, a mapping relationship is established between the position of the occluded area in the video image and the surface position of the target component, and the projection position of the occluded area on the target component is determined according to the mapping relationship.

[0021] Further, determining the projection position of the occluded area on the target component based on the mapping relationship includes: Determine the boundary position of the target component in the video image, and establish a mapping relationship from image coordinates to component surface coordinates based on the correspondence between the boundary position of the target component in the video image and the reference coordinates of the target component surface; Extract the boundary point set of the occluded area, and convert the image coordinates of the boundary point set into the surface coordinates of the target component through the mapping relationship. Use the area range corresponding to the surface coordinates as the projection position of the occluded area on the target component.

[0022] like Figure 3 As shown, the rectangular outline represents the physical boundary of the target component in the surface coordinate system. The horizontal axis represents the length direction of the component surface, and the vertical axis represents the width direction. The component's interior is divided into substring 1, substring 2, and substring 3 along its length, with dashed lines between different substrings indicating substring boundaries within the component. The shaded polygonal area in the figure represents the projection of the occluded area obtained from video image recognition onto the target component surface. By mapping the occluded area in the video image to the target component surface coordinate system, the positional relationship of the occluded area relative to the substring boundaries can be clearly defined, thus providing a spatial positioning basis for subsequent extraction of occlusion boundary crossing events and determination of substring entry and exit order.

[0023] S2, determine the substring crossing order corresponding to the occlusion boundary based on the positional relationship between the projection position and the substring boundary line inside the component; The sub-string boundary line within the module refers to the boundary line between the coverage areas of adjacent sub-strings, determined in the coordinate system of the target module surface, based on the cell arrangement structure, sub-string series grouping relationship, and the correspondence between each sub-string and the bypass diode. This internal sub-string boundary line is not necessarily a physical line visible on the module surface; rather, it is a structural reference line used to characterize the surface coverage of different electrical sub-strings. Specifically, the coverage area of ​​each sub-string in the module surface coordinate system can be determined based on the target module's model parameters, size parameters, and internal sub-string division parameters, and the boundary position of adjacent coverage areas is determined as the internal sub-string boundary line.

[0024] The substring crossing order includes the substring entering order and the substring exit order corresponding to the occlusion boundary.

[0025] In this embodiment, determining the substring crossing order corresponding to the occlusion boundary based on the positional relationship between the projection position and the substring boundary line inside the component includes: Based on the substring division parameters corresponding to the target component model, determine the position of each substring boundary line in the coordinate system on the surface of the target component; Extract the position change trajectory of the occlusion region boundary in the target component surface coordinate system from consecutive image frames; Determine whether the position change trajectory crosses the boundary lines of each substring, and record the corresponding substring boundary crossing event when a cross occurs; Based on the time sequence of each substring's boundary crossing event, the order of substring boundary crossings corresponding to the occlusion boundary is generated.

[0026] The determination of whether the position change trajectory crosses the boundary lines of each substring includes: Extract the set of position points corresponding to the occlusion boundary at consecutive time steps, and determine the side of each position point relative to the boundary line of each substring; When the corresponding position point changes from one side of the same substring boundary line to the other side in two consecutive moments, or when the set of position points at the same moment is simultaneously distributed on both sides of the substring boundary line, it is determined that the position change trajectory crosses the substring boundary line. When the crossing state continues to reach a preset number of frames or a preset time threshold, the corresponding substring crossing event is recorded.

[0027] Furthermore, the step of generating the substring crossing order corresponding to the occlusion boundary based on the temporal sequence of each substring crossing event includes: Record the moment when each substring cross-boundary event first meets the crossing judgment condition, and use it as the event moment of the corresponding substring cross-boundary event, and establish the association between the event moment and the corresponding substring boundary marker; The events of each substring crossing the boundary are sorted from front to back according to their event times to obtain the sequence of crossing the boundary of each substring from the occlusion boundary; among them, the substring entry order is generated based on the sequence of crossing the boundary during the advancement phase of the occlusion boundary, and the substring exit order is generated based on the sequence of crossing the boundary during the exit phase of the occlusion boundary. The order of substring crossings corresponding to the occlusion boundary is generated based on the crossing sequence.

[0028] In existing technologies, video surveillance can typically only identify the occlusion area or size on the component surface, making it difficult to determine the order in which different substrings within the component enter the system due to occlusion. Therefore, it cannot convert the surface occlusion process in the video into a basis for determining the switching order of bypass diodes. To address this, this embodiment does not directly rely on the occlusion area or center position. Instead, it extracts substring boundary-crossing events based on the positional changes of the occlusion boundary relative to the substring boundary lines within the component, and determines the substring boundary-crossing order according to the sequence of these events. This transforms the continuous spatial changes on the video side into sequential information that can be used to construct a bypass response reference sequence.

[0029] like Figure 4As shown, the horizontal axis represents time, and the vertical axis represents the position of the occlusion boundary along the length of the component. The two horizontal dashed lines represent substring boundaries L1 and L2, respectively, and the solid line represents the trajectory of the occlusion boundary position changing over time. The t1, t2, t3, and t4 marked in the figure correspond to the event times when the occlusion boundary crosses different substring boundaries. t1 and t2 belong to the advancement phase, corresponding to the substring entry sequence; t3 and t4 belong to the exit phase, corresponding to the substring exit sequence. This figure visually demonstrates that this invention does not rely on static occlusion area or occlusion center position for judgment, but rather extracts substring crossing events based on the dynamic crossing relationship between the occlusion boundary and the substring boundaries, and generates the substring crossing sequence according to the order of these events, thereby converting continuous spatial changes on the video side into sequential information with electrical structural significance.

[0030] S3. Based on the substring crossover order and the substring and bypass diode, construct the bypass response reference sequence of the target string. The bypass response reference sequence includes the bypass event timing sequence and the corresponding voltage step reference sequence. The correspondence between the substring and the bypass diode refers to the preset connection correspondence between each substring inside the target component and the bypass diode used to bypass and protect the substring.

[0031] The preset connection relationships are derived from the internal electrical structure of the target component and can be predetermined based on component model parameters, junction box configuration, and sub-string division method. Typically, each sub-string corresponds to one bypass diode; when a sub-string experiences an abnormal operating state due to obstruction, the bypass diode corresponding to that sub-string enters a conducting state, providing bypass protection for that sub-string.

[0032] In this embodiment, the construction of the bypass response reference sequence of the target string based on the substring cross-boundary order and the substring and bypass diode includes: Based on the order of substring crossing, the process of the occlusion boundary changing on the target component is divided into multiple temporal segments in which substrings enter the occlusion sequentially and exit the occlusion sequentially. For each time sequence segment, based on the set of target substrings within the blocking range in each consecutive time sequence segment and the correspondence between each substring and the bypass diode, the bypass response reference combination corresponding to the target string in that time sequence segment is determined; the correspondence between the substring and the bypass diode refers to the preset connection correspondence between each substring and each bypass diode determined based on the junction box configuration of the target component, the substring series grouping method, and the bypass protection connection method; the bypass response reference combination refers to the target bypass diode combination corresponding to that time sequence segment, determined based on the set of target substrings within the blocking range and the correspondence between each substring and the bypass diode. Based on the changes in the bypass response reference combination between adjacent time segments, the corresponding bypass entry event and bypass exit event are determined; specifically, whether the bypass response reference combination between the preceding and following time segments is added or reduced is used as the criterion for judgment. Based on the chronological order of bypass entry and exit events, a corresponding voltage step reference sequence is generated, resulting in the bypass response reference sequence for the target string. The voltage step reference sequence refers to a reference step sequence determined based on the bypass event timing sequence, formed by arranging voltage drop steps and voltage rise steps in chronological order of occurrence.

[0033] The process of dividing the occlusion boundary change on the target component into multiple temporal segments where substrings sequentially enter and exit occlusion, based on the substring crossing order, includes: Record the event times corresponding to the entry and exit events of each substring; The time interval between two adjacent event moments, the time interval between the start moment and the first event moment, and the time interval between the last event moment and the end moment are taken as continuous time sequence segments; Determine the set of target substrings within the occlusion range in each consecutive time segment; When the set of target substrings corresponding to adjacent time segments is increased, the next time segment is determined as the time segment where the substring enters the occlusion; when the set of target substrings corresponding to adjacent time segments is decreased, the next time segment is determined as the time segment where the substring exits the occlusion.

[0034] Further, determining the bypass response reference combination corresponding to the target string in this time segment includes: Based on the internal electrical structure of the target component, a one-to-one correspondence is established between each substring and each bypass diode, wherein each substring corresponds to a bypass diode used for bypass protection of that substring. Based on the set of target substrings within the blocking range in each time segment, determine the set of target bypass diodes corresponding to each substring in the set of target substrings; The target bypass diode set is determined as the bypass response reference combination for the corresponding timing segment.

[0035] The step of generating a corresponding voltage step reference sequence based on the chronological order of each bypass entry and exit event, to obtain the bypass response reference sequence of the target string, includes: The bypass event time sequence is obtained by sorting the events corresponding to each bypass entry event and each bypass exit event. Based on the bypass event timing sequence, determine the voltage drop step reference relationship corresponding to each bypass entry event, and the voltage recovery step reference relationship corresponding to each bypass exit event; Based on the order in which the reference relationships of each voltage drop step and each voltage rise step appear, a voltage step reference sequence for the target string is generated. By associating the bypass event timing sequence with the voltage step reference sequence, a bypass response reference sequence for the target string is obtained.

[0036] like Figure 5 As shown, the upper part presents the different time segments P0, P1, P2, P3, and P4 corresponding to the occlusion boundary change process, as well as the set of target substrings within the occlusion range in each time segment; the middle part presents the bypass response reference combination corresponding to each time segment, for example, gradually changing from an empty set to containing only D1, then changing to containing both D1 and D2, then exiting to containing only D1 and finally returning to an empty set; the lower part presents the reference voltage step sequence generated based on the sequential relationship between bypass entry and bypass exit events, where a downward step represents a reference voltage decrease event, and an upward step represents a reference voltage recovery event.

[0037] S4 extracts actual voltage step events from the voltage change process of the target string, and outputs the shading electrical response monitoring results of the target string based on the correspondence between the actual voltage step events and the voltage step reference sequence.

[0038] In this embodiment, extracting actual voltage step events from the voltage change process of the target string includes: The voltage time-series data of the target string is denoised and smoothed to obtain the voltage change sequence; Calculate the voltage change between adjacent sampling times in the voltage change sequence, and identify candidate change segments with continuous and consistent voltage change directions; When the cumulative voltage change amplitude within a candidate change segment exceeds a preset step threshold, and the duration of the candidate change segment reaches a preset duration threshold, the candidate change segment is determined as an actual voltage step event. Among them, when the cumulative voltage change is decreasing, it is determined as an actual voltage decrease step event; when the cumulative voltage change is increasing, it is determined as an actual voltage recovery step event. Record the event time, duration, and direction of each actual voltage step event.

[0039] The voltage change sequence is specifically as follows:

[0040]

[0041]

[0042] in, The smoothed voltage sequence The width of the sliding window is half its width. To smooth the window length, For a moment The collected target string voltage values, This represents the voltage change between adjacent sampling times. It is a voltage change sequence. This represents the total number of voltage sampling points acquired by the target string within the current monitoring time period.

[0043] Furthermore, the monitoring results of the obstruction electrical response of the output target string include: The extracted actual voltage step events are sorted according to the event time to form an actual voltage step sequence; The actual voltage step sequence is compared with the voltage step reference sequence to determine whether the two are consistent in the number of step events, the step direction at each corresponding position, and the order in which each step event occurs. When the actual voltage step sequence is consistent with the voltage step reference sequence in terms of the number of step events, the step direction at each corresponding position, and the order of occurrence of each step event, the monitoring result of normal electrical response of the target string block is output. When the actual voltage step sequence is inconsistent with the voltage step reference sequence in terms of the number of step events, the step direction at each corresponding position, or the order of occurrence of each step event, the monitoring result of the abnormal electrical response of the target string blockage is output.

[0044] Example 2, Figure 2 This invention presents a photovoltaic power plant monitoring system based on the fusion of video perception and operational data, comprising the following modules: Occlusion projection determination module: used to identify occlusion areas on the component based on video images and determine the projection position of the occlusion area on the target component; Cross-boundary order generation module: used to determine the cross-boundary order of substrings corresponding to the occlusion boundary based on the positional relationship between the projection position and the substring boundary line inside the component; Bypass Response Reference Construction Module: Used to construct a bypass response reference sequence for a target string based on the substring crossover order and the substring and bypass diode. The bypass response reference sequence includes a bypass event timing sequence and its corresponding voltage step reference sequence. Response determination module: used to extract actual voltage step events from the voltage change process of the target string, and output the obstruction electrical response monitoring results of the target string based on the correspondence between the actual voltage step events and the voltage step reference sequence.

[0045] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0046] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0047] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0048] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0049] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0050] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A photovoltaic power plant monitoring method and system based on the fusion of video perception and operational data, characterized in that, Includes the following steps: Based on the occlusion area on the video image recognition component, determine the projection position of the occlusion area on the target component; Based on the positional relationship between the projection position and the boundary line of the substring inside the component, determine the substring crossing order corresponding to the occlusion boundary; Based on the substring crossover order and the substring and bypass diode, a bypass response reference sequence for the target string is constructed. The bypass response reference sequence includes a bypass event timing sequence and a corresponding voltage step reference sequence. The actual voltage step events are extracted from the voltage change process of the target string, and the obstruction electrical response monitoring results of the target string are output based on the correspondence between the actual voltage step events and the voltage step reference sequence.

2. The photovoltaic power plant monitoring method based on video perception and operational data fusion according to claim 1, characterized in that, The step of determining the projection position of the occlusion area on the target component based on the occlusion area on the video image recognition component includes: Locate the image region corresponding to the target component in the video image, and identify the occluded area on the surface of the component based on the brightness difference, texture difference and edge change between the image region and the preset unoccluded reference image; Based on the camera calibration parameters, the boundary position of the target component in the video image, and the actual installation position of the target component, a mapping relationship is established between the position of the occluded area in the video image and the surface position of the target component, and the projection position of the occluded area on the target component is determined.

3. The photovoltaic power plant monitoring method based on video perception and operational data fusion according to claim 2, characterized in that, Determining the projection position of the occluded area on the target component includes: Determine the boundary position of the target component in the video image, and establish a mapping relationship from image coordinates to component surface coordinates based on the correspondence between the boundary position of the target component in the video image and the reference coordinates of the target component surface; Extract the boundary point set of the occluded area, and convert the image coordinates of the boundary point set into the surface coordinates of the target component through the mapping relationship. Use the area range corresponding to the surface coordinates as the projection position of the occluded area on the target component.

4. The photovoltaic power plant monitoring method based on video perception and operational data fusion according to claim 3, characterized in that, The step of determining the substring crossing order corresponding to the occlusion boundary based on the positional relationship between the projection position and the substring boundary line inside the component includes: Based on the substring division parameters corresponding to the target component model, determine the position of each substring boundary line in the coordinate system on the surface of the target component; Extract the position change trajectory of the occlusion region boundary in the target component surface coordinate system from consecutive image frames; Determine whether the position change trajectory crosses the boundary lines of each substring, and record the corresponding substring boundary crossing event when a cross occurs; Based on the time sequence of each substring's boundary crossing event, the order of substring boundary crossings corresponding to the occlusion boundary is generated.

5. The photovoltaic power plant monitoring method based on video perception and operational data fusion according to claim 4, characterized in that, The determination of whether the position change trajectory crosses the boundary lines of each substring includes: Extract the set of position points corresponding to the occlusion boundary at consecutive time steps, and determine the side of each position point relative to the boundary line of each substring; When the corresponding position point changes from one side of the same substring boundary line to the other side in two consecutive moments, or when the set of position points at the same moment is simultaneously distributed on both sides of the substring boundary line, it is determined that the position change trajectory crosses the substring boundary line. When the crossing state continues to reach a preset number of frames or a preset time threshold, the corresponding substring crossing event is recorded.

6. The photovoltaic power plant monitoring method based on video perception and operational data fusion according to claim 5, characterized in that, The step of generating the substring crossing order corresponding to the occlusion boundary based on the time sequence of each substring crossing event includes: Record the moment when each substring cross-boundary event first meets the cross-boundary judgment condition, and use it as the event moment of the corresponding substring cross-boundary event, and establish the association between the event moment and the corresponding substring boundary marker; The events of each substring crossing the boundary are sorted from front to back according to their event times to obtain the sequence of crossings of the occlusion boundary with respect to the boundary line of each substring. Generate the substring crossing order corresponding to the occlusion boundary based on the sequence of crossings.

7. The photovoltaic power plant monitoring method based on video perception and operational data fusion according to claim 6, characterized in that, The method for constructing a bypass response reference sequence for the target string based on the substring cross-boundary order and the substring and bypass diode includes: Based on the substring crossing order, the process of the change of the occlusion boundary on the target component is divided into multiple time segments; For each time segment, based on the set of target substrings within the blocking range and the correspondence between each substring and the bypass diode, the corresponding bypass response reference combination is determined; Based on the changes in the bypass response reference combination between adjacent time segments, the corresponding bypass entry event and bypass exit event are determined; Based on the order of bypass entry and bypass exit events, a corresponding voltage step reference sequence is generated to obtain the bypass response reference sequence of the target string.

8. The photovoltaic power plant monitoring method based on video perception and operational data fusion according to claim 7, characterized in that, The process of changing the occlusion boundary on the target component is divided into multiple time segments based on the substring crossing order, including: Record the event times corresponding to the entry and exit events of each substring; The time interval between two adjacent event moments, the time interval between the start moment and the first event moment, and the time interval between the last event moment and the end moment are taken as continuous time sequence segments; Determine the set of target substrings within the occlusion range in each consecutive time segment; When a new target substring set is added to the adjacent time sequence segment, the next time sequence segment is determined as the time sequence segment where the substring enters the occlusion. When the set of target substrings corresponding to adjacent time segments decreases, the next time segment is determined as the time segment where the substring exits the occlusion.

9. The photovoltaic power plant monitoring method based on video perception and operational data fusion according to claim 8, characterized in that, The output target string's obstruction electrical response monitoring results include: The extracted actual voltage step events are sorted according to the event time to form an actual voltage step sequence; The actual voltage step sequence is compared with the voltage step reference sequence to determine whether the order of occurrence of each step event in the actual voltage step sequence is consistent with the voltage step reference sequence. When the actual voltage step sequence matches the voltage step reference sequence, the output shows a monitoring result indicating that the electrical response of the target string block is normal. When the actual voltage step sequence is inconsistent with the voltage step reference sequence, the monitoring results of the abnormal electrical response of the target string block are output.

10. A system using the photovoltaic power plant monitoring method based on video perception and operational data fusion as described in any one of claims 1-9, characterized in that, Includes the following modules: Occlusion projection determination module: used to identify occlusion areas on the component based on video images and determine the projection position of the occlusion area on the target component; Cross-boundary order generation module: used to determine the cross-boundary order of substrings corresponding to the occlusion boundary based on the positional relationship between the projection position and the substring boundary line inside the component; Bypass Response Reference Construction Module: Used to construct a bypass response reference sequence for a target string based on the substring crossover order and the substring and bypass diode. The bypass response reference sequence includes a bypass event timing sequence and its corresponding voltage step reference sequence. Response determination module: used to extract actual voltage step events from the voltage change process of the target string, and output the obstruction electrical response monitoring results of the target string based on the correspondence between the actual voltage step events and the voltage step reference sequence.

Images