Photovoltaic station unmanned aerial vehicle intelligent operation and maintenance method and system based on image recognition

Abstract

The invention relates to the technical field of photovoltaic industry, in particular to a photovoltaic station unmanned aerial vehicle intelligent operation and maintenance method based on image recognition, and the method comprises the steps: carrying out the partitioning based on a photovoltaic station, generating an inspection task queue, and carrying out the dynamic adjustment of the inspection task queue based on real-time illumination; acquiring a real-time acquisition image when the inspection task is performed according to the dynamically adjusted inspection task queue, generating a defect comprehensive score of the photovoltaic module according to the real-time acquisition image, and generating a fault judgment result according to the defect comprehensive score; generating an operation and maintenance task level according to the fault judgment result, and generating a task work order according to the operation and maintenance task level; and obtaining task execution data after the task work order is executed, and performing task feedback optimization according to the task execution data. According to the invention, the inspection efficiency and the identification accuracy can be improved, and the dynamic control of the defect development trend and the accurate distribution of task resources are realized.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of photovoltaic industry, in particular to a photovoltaic station unmanned aerial vehicle intelligent operation and maintenance method and system based on image recognition. BACKGROUND

[0002] With the continuous increase of photovoltaic installed capacity, centralized photovoltaic stations are increasingly becoming an important part of clean energy, and their operation stability and power generation efficiency directly affect the economy and safety of the overall system. Therefore, using unmanned aerial vehicles to carry image acquisition equipment for component defect inspection has become a mainstream technical path.

[0003] In the prior art, an invention patent with publication number CN119906363A discloses a photovoltaic equipment state discrimination and fault diagnosis method based on AI, which comprises the following steps: step 1. photovoltaic equipment data acquisition, step 2. photovoltaic equipment state analysis, step 3. string fault inspection and step 4. string fault processing. By preliminarily monitoring each string belonging to the power plant, it is determined whether each string belonging to the power plant has any fault abnormality, so that subsequent classification inspection is facilitated. By controlling the unmanned aerial vehicle to perform inspection, the range of the string with the open circuit problem is determined, thereby avoiding the need for the maintenance person in charge to conduct extensive investigation.

[0004] Although there are methods for photovoltaic operation and maintenance in the prior art, there are still problems, such as the lack of pertinence in the existing route planning and inspection task scheduling mechanism. Generally, fixed routes or equidistant scanning strategies are used. This method cannot fully consider factors such as component layout density, historical fault hot zones, local shading and environmental disturbances, resulting in low inspection path efficiency, delayed coverage of important areas, and inability to achieve dynamic optimal allocation of inspection resources.

[0005] Therefore, it is necessary to design a photovoltaic station unmanned aerial vehicle intelligent operation and maintenance method based on image recognition. SUMMARY

[0006] Therefore, it is necessary to design a photovoltaic station unmanned aerial vehicle intelligent operation and maintenance method based on image recognition.

[0007] The technical scheme of the application is as follows: A photovoltaic station unmanned aerial vehicle intelligent operation and maintenance method based on image recognition, the method comprising: The photovoltaic station is divided into zones and an inspection task queue is generated, and the inspection task queue is dynamically adjusted based on real-time illumination; acquire real-time acquisition images when performing the inspection task according to the dynamically adjusted inspection task queue, generate a defect comprehensive score of the photovoltaic module according to the real-time acquisition images, and generate a fault judgment result according to the defect comprehensive score; generate an operation and maintenance task level according to the fault judgment result, and generate a task work order according to the operation and maintenance task level; acquire task execution data after executing the task work order, and perform task feedback optimization according to the task execution data.

[0008] Specifically, the photovoltaic station is partitioned and an inspection task queue is generated, the inspection task queue is dynamically adjusted based on real-time illumination, including: The component array of the photovoltaic station is partitioned and a plurality of regions are generated, and an inspection task queue is generated according to each region; The inspection task queue is dynamically adjusted according to real-time illumination dynamic sorting.

[0009] Specifically, the component array of the photovoltaic station is partitioned and a plurality of regions are generated, and an inspection task queue is generated according to each region, including: The component array of the photovoltaic station is partitioned and a plurality of regions are generated, and a priority score model is constructed according to each region; An inspection priority score is output according to the priority score model, and an inspection task queue is generated according to the inspection priority score.

[0010] Specifically, the inspection task queue is dynamically adjusted according to real-time illumination dynamic sorting, including: Real-time illumination of the region is acquired, and a current illumination interference factor is generated according to the real-time illumination; A scheduling priority is generated according to the current illumination interference factor and the inspection priority score corresponding to the region; The inspection task queue is dynamically adjusted according to the scheduling priority.

[0011] Specifically, real-time acquisition images when performing the inspection task according to the dynamically adjusted inspection task queue are acquired, a defect comprehensive score of the photovoltaic module is generated according to the real-time acquisition images, and a fault judgment result is generated according to the defect comprehensive score, including: Real-time acquisition images when performing the inspection task according to the dynamically adjusted inspection task queue are acquired, image fusion is performed according to the real-time acquisition images, and a defect comprehensive score of the photovoltaic module is generated; The defect comprehensive score is compared with historical recognition data recorded in the photovoltaic station database to generate a fault judgment result.

[0012] Specifically, the fault judgment result generated based on the comprehensive defect score is compared with the historical identification data recorded in the photovoltaic station database, and the fault judgment result is generated, including: Generating a fault judgment result based on the comprehensive defect score and substituting the historical identification data recorded in the photovoltaic station database into a trend score update model; The model is updated according to the trend score to generate a fault judgment result.

[0013] Specifically, generating an operation and maintenance task level according to the fault judgment result, and generating a task work order according to the operation and maintenance task level, including: Constructing a task level mapping model according to the fault judgment result; An operation and maintenance task level is generated according to the task level mapping model, and a task work order is generated according to the operation and maintenance task level.

[0014] Specifically, a photovoltaic station drone intelligent operation and maintenance system based on image recognition is also provided, and the system includes: An inspection task generation module is used to partition the photovoltaic field and generate an inspection task queue, and dynamically adjust the inspection task queue based on real-time light intensity; a fault judgment generation module, configured to obtain real-time collected images when performing inspection tasks according to the dynamically adjusted inspection task queue, generate a comprehensive defect score of the photovoltaic module based on the real-time collected images, and generate a fault judgment result based on the comprehensive defect score; A task work order generation module is used to generate an operation and maintenance task level according to the fault judgment result, and generate a task work order according to the operation and maintenance task level; The task execution feedback module is used to obtain task execution data after executing the task work order, and perform task feedback tuning based on the task execution data.

[0015] Specifically, the inspection task generation module is further used to: partition and generate multiple areas based on the component array of the photovoltaic station, and generate an inspection task queue according to each of the areas; and dynamically adjust the inspection task queue according to real-time illumination dynamic sorting.

[0016] Specifically, the inspection task generation module is also used to: partition and generate multiple areas based on the component array of the photovoltaic station, and build a priority scoring model according to each of the areas; output the inspection priority score according to the priority scoring model, and generate an inspection task queue according to the inspection priority score.

[0017] Specifically, the inspection task generation module is further configured to: acquire real-time light of the region, and generate a current light interference factor according to the real-time light; generate a scheduling priority according to the current light interference factor and a region corresponding inspection priority score; and dynamically adjust the inspection task queue according to the scheduling priority.

[0018] Specifically, the fault judgment generation module is further configured to: acquire real-time acquisition images when performing the inspection task according to the dynamically adjusted inspection task queue, perform image fusion according to the real-time acquisition images, and generate a defect comprehensive score of the photovoltaic module; and generate a fault judgment result, compare the fault judgment result with historical identification data recorded in a photovoltaic power station database, and generate the fault judgment result.

[0019] Specifically, the fault judgment generation module is further configured to: generate a fault judgment result according to the defect comprehensive score, and input the fault judgment result into a trend score updating model recorded in the photovoltaic power station database; and generate the fault judgment result according to the trend score updating model.

[0020] Specifically, the task work order generation module is further configured to: construct a task level mapping model according to the fault judgment result; generate an operation and maintenance task level according to the task level mapping model, and generate a task work order according to the operation and maintenance task level.

[0021] Optionally, a computer device is also provided, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above-mentioned image recognition based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method when executing the computer program.

[0022] Optionally, a computer readable storage medium is also provided, which stores a computer program, and the computer program implements the steps of the above-mentioned image recognition based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method when executed by a processor.

[0023] The present application relates to machine learning and deep learning technology, which realizes the following technical effects: (1) The above-mentioned image recognition based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method solves the problem that the fixed template type setting inspection route method in the prior art easily leads to low inspection flexibility, and solves the problem that there is no differentiated inspection priority, and the flexible adjustment cannot be performed according to real-time fault density or dynamic environment of the power station.

[0024] (2) By acquiring the real-time acquisition image when performing the inspection task according to the dynamically adjusted inspection task queue, generating a defect comprehensive score of the photovoltaic module according to the real-time acquisition image, and generating a fault judgment result according to the defect comprehensive score, the problem of low inspection data acquisition efficiency caused by strong reflection is avoided; through the coupling processing mode of visible light and thermal imaging, the defect recognition of dual-mode data is realized, the visible light frame and the infrared frame of the same component are one-to-one corresponding through synchronous acquisition and timestamp alignment, and pseudo defects caused by frame difference are avoided.

[0025] (3) According to the fault judgment result, the operation and maintenance task level is generated, and the task work order is generated according to the operation and maintenance task level; and the task execution data after executing the task work order is acquired, and the task feedback optimization is performed according to the task execution data, which has the technical advantages of high intelligent degree, fast response speed and strong self-adaptation ability, and significantly enhances the operation and maintenance automation and safety protection ability of the photovoltaic power station. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 It is a flowchart of the image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method in one embodiment; Figure 2 It is a structural block diagram of the image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance system in one embodiment; Figure 3 It is a structural block diagram of the computer device in one embodiment. DETAILED DESCRIPTION

[0027] In the following description, specific details are set forth, such as particular system configurations, techniques, etc., in order to provide a thorough understanding of the present embodiments. However, persons skilled in the art will understand that the present application can be practiced without these specific details. In other instances, well-known structures, devices, circuits, and methods have not been described in detail in order to avoid obscuring the present application.

[0028] It should be understood that, when used in the specification and the appended claims of the present application, the term "comprising" indicates the presence of the described features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0029] It should also be understood that the term "and / or" used in the specification and the appended claims of the present application means any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0030] As used in the specification and the appended claims, the term "if' can be interpreted as meaning "when" or "upon" or "in response to determining" or "in response to detecting" depending on the context. Similarly, the phrase "if it is determined" or "if [the described condition or event] is detected" can be interpreted to mean "upon determining" or "in response to determining" or "upon detecting [the described condition or event]" or "in response to detecting [the described condition or event]" depending on the context.

[0031] In addition, in the description of the present application and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0032] Reference in the specification to "one embodiment" or "some embodiments" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. The appearances of the phrase "in one embodiment" or "in some embodiments" in various places in the specification are not necessarily all referring to the same embodiment, although it can. The terms "comprising", "including", "having" and their variants are meant to be construed as "including but not limited to", unless otherwise noted.

[0033] In one embodiment, a terminal is provided, which is configured to: divide a photovoltaic station into zones and generate a patrol task queue based on the photovoltaic station, dynamically adjust the patrol task queue based on real-time illumination; acquire real-time acquisition images when performing a patrol task according to the dynamically adjusted patrol task queue, generate a defect comprehensive score of a photovoltaic module according to the real-time acquisition images, and generate a fault judgment result according to the defect comprehensive score; generate an operation and maintenance task level according to the fault judgment result, and generate a task work order according to the operation and maintenance task level; acquire task execution data after executing the task work order, and perform task feedback optimization according to the task execution data.

[0034] The terminal can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers and portable wearable devices.

[0035] In one embodiment, as shown in Figure 1 An image recognition-based photovoltaic station unmanned aerial vehicle intelligent operation and maintenance method is provided, which includes: Step S100: dividing a photovoltaic station into zones and generating a patrol task queue based on the photovoltaic station, and dynamically adjusting the patrol task queue based on real-time illumination; Step S200: acquiring real-time acquisition images when performing the inspection task according to the dynamically adjusted inspection task queue, generating a defect comprehensive score of the photovoltaic module according to the real-time acquisition images, and generating a fault judgment result according to the defect comprehensive score; Step S300: generating an operation and maintenance task level according to the fault judgment result, and generating a task work order according to the operation and maintenance task level; Step S400: acquiring task execution data after executing the task work order, and performing task feedback optimization according to the task execution data.

[0036] In the present application, by partitioning based on a photovoltaic station and generating an inspection task queue, dynamically adjusting the inspection task queue based on real-time illumination, the problem that the method of using a fixed template to set an inspection route in the prior art easily leads to low inspection flexibility is solved, and the problem that there is no differentiated inspection priority and the problem that it is not possible to flexibly adjust according to real-time fault density or dynamic environment of the station is solved. Then, by acquiring real-time acquisition images when performing the inspection task according to the dynamically adjusted inspection task queue, generating a defect comprehensive score of the photovoltaic module according to the real-time acquisition images, and generating a fault judgment result according to the defect comprehensive score, the problem of low inspection data acquisition efficiency caused by strong reflection is avoided. Through the coupling processing mode of visible light and thermal imaging, defect recognition reinforcement of dual-mode data is realized, synchronous acquisition and timestamp alignment ensure that the visible light frame and the infrared frame of the same component correspond one by one, and pseudo-defects caused by frame differences are avoided. Then, an operation and maintenance task level is generated according to the fault judgment result, and a task work order is generated according to the operation and maintenance task level. And task execution data after executing the task work order is acquired, and task feedback optimization is performed according to the task execution data, which has the technical advantages of high intelligent degree, fast response speed, and strong self-adaptation capability, and significantly enhances the operation and maintenance automation and safety protection capability of the photovoltaic power station.

[0037] In one embodiment, step S100: partitioning based on a photovoltaic station and generating an inspection task queue, dynamically adjusting the inspection task queue based on real-time illumination, comprises: Step S110: partitioning based on a component array of a photovoltaic station and generating a plurality of regions, and generating an inspection task queue according to each region; Step S120: dynamically adjusting the inspection task queue according to real-time illumination dynamic sorting.

[0038] In this embodiment, in order to solve the problem that the method of setting the inspection route by using the fixed template in the prior art easily leads to low inspection flexibility, the component array of the photovoltaic station is partitioned and a plurality of regions are generated, and an inspection task queue is generated according to each region, so as to set the inspection queue based on different regions, and then in order to avoid the problem that the strong reflection situation causes low inspection data collection efficiency, the inspection task queue is dynamically adjusted by real-time light dynamic sorting.

[0039] In one embodiment, step S110: partitioning the component array of the photovoltaic station and generating a plurality of regions, and generating an inspection task queue according to each region, comprises: Step S111: partitioning the component array of the photovoltaic station and generating a plurality of regions, and constructing a priority score model according to each region. Step S112: outputting an inspection priority score according to the priority score model, and generating an inspection task queue according to the inspection priority score.

[0040] In this embodiment, considering that the existing route planning methods mainly include two types, and both types have disadvantages. The first type is a fixed template type inspection route, specifically, simple partitioning is performed according to an initial component arrangement diagram, a flight path is preset, there is no differentiated inspection priority, and the flight path cannot be flexibly adjusted according to real-time fault density or dynamic environment of the station. The second type is a uniform coverage path based on a grid method or an equal interval strategy, which has certain automation capability, but usually does not introduce historical state data of components, terrain information or local occlusion analysis, resulting in low efficiency, repeated invalid flight and important area detection lag.

[0041] Therefore, the component array of the photovoltaic station is first partitioned and a plurality of regions are generated, a priority score model is constructed according to each region, then an inspection priority score is outputted according to the priority score model, and an inspection task queue is generated according to the inspection priority score.

[0042] Specifically, the priority score model is specifically as follows:

[0043] wherein, is the inspection priority score of the i-th region, which is a dimensionless technical feature. is the maximum inspection priority score calculated in all regions, and the unit is pieces·s / m 3 .

[0044] is the historical average of the component failure rate in the ith region, which is a dimensionless technical feature. Based on the preset defect database, the average of "effective defect record number ÷ inspection frequency × 100%" of all components in the region i in the last 12 months is obtained.

[0045] is the component density per unit area. When importing the BIM or GIS layout of the photovoltaic power station, the component arrangement coordinates are read and the "total number of components in the region ÷ the horizontal projection area (m 2 ) of the region" is calculated. This index is recalculated only when the array layout is adjusted or new strings are added, and generally remains stable for more than half a year, with the unit being pieces / m 2 .

[0046] is the average of the actual wind speed in the region, with the unit being m / s. When accessing the real-time data stream of the weather station and the unmanned aerial vehicle onboard anemometer of the power station, the arithmetic mean of the instantaneous wind speed during the flight of the sub-region is calculated. If the area of the region is larger than the coverage radius of the onboard measuring point, inverse distance weighted interpolation is used, and the result is refreshed in real time before each inspection flight path planning.

[0047] is the average shading interference factor of the region. In the last three visible light aerial images, the area proportion of fixed shading (mountains, buildings, supports), semi-fixed shading (transformer boxes, combiner boxes) and temporary shading (vehicles, tree shadows) is detected by the light projection algorithm, and the score is accumulated according to the duration of the shading. The latest image is called before the first flight each day to reevaluate, which is a dimensionless technical feature.

[0048] is the equivalent wind speed conversion coefficient of the shading interference factor, with the unit being m / s.

[0049] The priority scoring model filters out the region that is currently most worthy of priority inspection by increasing the weight of the component failure rate and the component density, and penalizing the wind speed and shading interference.

[0050] Further, the inspection priority score is generated in order from high to low to generate an inspection task queue.

[0051] After generating the inspection task queue, the inspection task queue is dynamically adjusted according to the real-time light dynamic sorting.

[0052] In one embodiment, step S120: dynamically adjusting the inspection task queue according to the real-time light dynamic sorting, comprises: Step S121: acquiring the real-time light of the region, and generating a current light interference factor according to the real-time light; Step S122: generating a scheduling priority according to the current light interference factor and the inspection priority score corresponding to the region; Step S123: dynamically adjusting the inspection task queue according to the scheduling priority.

[0053] In the present embodiment, in the existing practice of unmanned aerial vehicle inspection of most photovoltaic stations, task scheduling usually adopts two common strategies. The first strategy is static sequential scheduling, which executes the shooting task according to the fixed flight route or simple grid partition designed initially, without considering real-time light, fault urgency or regional environmental differences. The second strategy is time slice polling scheduling, which divides the station into several blocks on average, and rotates according to the “first-in-first-shooting” or fixed time slice cycle, and at most avoids the noon strong reflection period.

[0054] The common defects of the two common strategies are three. The first is the lack of dynamic priority, which cannot weigh the regions with high real fault risk but poor current light conditions against the regions with low risk and good shooting conditions. The second is the unbalanced utilization of resources, and the high-value components and low-value components obtain the same flight attention, resulting in waste of flight time and battery life. The third is the inability to reflect the emergency warning. When a potential safety hazard such as hot spot outbreak occurs in a region, the existing scheduling mechanism often cannot immediately prioritize the task.

[0055] After the present embodiment generates the inspection task queue, the high-priority region task is prioritized and the low-priority region is delayed to improve the inspection efficiency and image value density. In addition, in order to improve the recognition accuracy, the optimal acquisition time window needs to be specified for each sub-task in the task queue, where the sub-task is the task of inspecting a specific region.

[0056] Further, the real-time light of the region is obtained first, and a current light interference factor is generated according to the real-time light. Specifically, the irradiance sensor, cloud camera and NOAA prediction data are read before the unmanned aerial vehicle takes off, the “light intensity normalized value x glare weight x cloud attenuation coefficient” of the region in the estimated shooting period is estimated, and the current light interference factor is generated. Then, it is updated every 3-5 minutes to ensure that it reflects the rapid change of the weather.

[0057] Then, a scheduling priority is generated according to the current light interference factor and the inspection priority score corresponding to the region, specifically, the scheduling priority is generated based on a priority scheduling model.

[0058] The priority scheduling model is as follows:

[0059] wherein, is the scheduling priority of the subtask corresponding to the jth inspection area. Wherein j represents the index of the inspection area, and the subtask refers to the task generated according to the inspection area in the inspection task queue. is the inspection priority score corresponding to the jth inspection area. is the current light interference factor corresponding to the area.

[0060] is a pre-set task urgency adjustment coefficient, set by the operation and maintenance scheduling background manually or by a rule engine, taking 1.0 in regular inspection, and automatically rising to 1.5-2.0 when receiving high temperature alarm or fire warning, immediately affecting the subsequent .

[0061] The priority scheduling model outputs will be used to dynamically sort the inspection task queue and drive the UAV to complete image acquisition in turn, that is, to dynamically adjust the inspection task queue according to the scheduling priority. Specifically, combined with the real-time light interference factor , the "shootability" is corrected to avoid the loss of efficiency caused by "high risk but cannot shoot" or "good light but low risk". At the same time, the task urgency adjustment coefficient allows the operation and maintenance center to improve the collection priority of a specific area when receiving manual or automatic warning, forming a soft real-time response.

[0062] In one embodiment, step S200: acquiring real-time acquisition images when performing the inspection task according to the dynamically adjusted inspection task queue, generating a defect comprehensive score of the photovoltaic module according to the real-time acquisition images, and generating a fault judgment result according to the defect comprehensive score, comprising: Step S210: acquiring real-time acquisition images when performing the inspection task according to the dynamically adjusted inspection task queue, performing image fusion according to the real-time acquisition images, and generating a defect comprehensive score of the photovoltaic module; Step S220: generating a fault judgment result according to the defect comprehensive score and comparing it with the historical identification data recorded in the photovoltaic power station database, and generating a fault judgment result.

[0063] In this embodiment, most of the current photovoltaic power station UAV inspection schemes still take a single imaging modality as the core in the fault detection link, only use visible light images for surface defect (cracking, shielding, dust accumulation) detection, and rely on convolutional neural networks or traditional threshold-edge methods, but are not sensitive to early internal cracks or thermal mismatch defects of poor contact of welding strips.

[0064] In this embodiment, the real-time acquisition image obtained according to the dynamically adjusted inspection task queue for performing the inspection task includes a visible light image and an infrared thermal image, the image fusion according to the real-time acquisition image is to fuse the visible light image and the infrared thermal image, and feature enhancement and fault extraction are further performed, typical fault types including fragment shielding, crack, thermal spot anomaly and connection rupture are recognized, and the image fusion recognition output result is represented as follows:

[0065] wherein, is the defect comprehensive score of the kth photovoltaic component. In this application, the defect comprehensive score is calculated as follows: Three threshold values are set, which are slight anomaly, severe anomaly and dangerous shutdown, and are marked in different colors on the GIS base map.

[0066] is an anomaly score recognized based on the visible light image, which is a score value obtained by performing target detection and semantic segmentation processing on the visible light image collected by the unmanned aerial vehicle. Specifically, first, brightness normalization and edge enhancement preprocessing are performed on the image, and then a pre-trained YOLO or Mask R-CNN image recognition model is used to recognize the defect types in the image, including component damage, bird droppings shielding and foreign matter accumulation. According to the recognized defect types, distribution range and morphological features, including crack length, foreign matter area and shielding position, a weighted scoring method is used to calculate the abnormality degree, and finally normalized to a score between 0 and 1 , and the higher the score, the more serious the abnormality.

[0067] is a thermal anomaly score obtained based on infrared thermal image analysis, which is obtained by analyzing the temperature distribution data of each photovoltaic component in the infrared thermal image. Specifically, first, the temperature mean and standard deviation of the region corresponding to each photovoltaic component are extracted, and the temperature difference mutation area and typical thermal spot features are recognized, then the parameters of the abnormal area temperature rise amplitude, thermal spot shape and thermal zone continuity are quantitatively evaluated, and compared with the set normal thermal threshold value, and the thermal anomaly grade of each component is calculated. The abnormal temperature rise amplitude and spatial distribution range are taken as the core weight factor, and the normalized is obtained after comprehensive calculation, which reflects the possible electrical hidden danger or connection fault degree of the photovoltaic component.

[0068] is a fusion consistency confidence coefficient, which is used to represent whether the recognition results between the two modalities are consistent. First, the IoU and defect type consistency rate of the visible light and infrared two anomaly masks are calculated, and then the weighted average is taken as the confidence, one value for each component, which is derived in real time. is a score adjustment coefficient, which is used to avoid that one modality dominates the score too much. For empirical hyperparameters, determine at model deployment time via cross-validation (typically 0.8-1.2), Modified only at version upgrade, not dynamically changed with running.

[0069] Therefore, the embodiment realizes defect identification reinforcement of dual-mode data by the coupling processing mode of visible light and thermal imaging, synchronous acquisition and timestamp alignment ensure that the visible light frame and infrared frame of the same component correspond one by one, and pseudo defects caused by frame differences are avoided. Pixel-level fusion enhancement superimposes the texture boundary of visible light and the temperature gradient of infrared, fills the blind area of both, and unifies the quantitative score Sort various defects on the same risk scale, thereby laying a foundation for subsequent trend analysis and work order classification.

[0070] In one embodiment, step S220: generate a fault judgment result according to the defect comprehensive score and compare it with historical identification data recorded in the photovoltaic power station database, and generate a fault judgment result, including: Step S221: According to the defect comprehensive score, generate a fault judgment result and substitute it into a trend score updating model with historical identification data recorded in the photovoltaic power station database; Step S222: According to the trend score updating model, generate a fault judgment result.

[0071] In the traditional photovoltaic operation and maintenance platform, historical trend analysis is mostly evaluated in units of whole strings or whole areas, ignoring single component differences, and cannot flexibly adjust sensitivity according to the historical fluctuation amplitude of different components; the analysis result only stays at the report level, and it is difficult to directly drive subsequent task classification.

[0072] In the embodiment, trend score updating model is constructed for comparative analysis to determine whether the current defect is a new fault, a more serious fault or no obvious change, specifically, according to the defect comprehensive score, a fault judgment result is generated and substituted into a trend score updating model with historical identification data recorded in the photovoltaic power station database.

[0073] The trend score updating model is as follows:

[0074] Wherein, is the fault change trend score of the kth photovoltaic component. is the defect comprehensive score. is the historical defect score average of the component, specifically, a sliding average of scores of the last N inspections (default N=8) is maintained for each component, and every time a new one is added The earliest one is automatically kicked out to ensure timeliness.

[0075] is a trend response coefficient, used to adjust the weight of historical data on the current judgment. The initial value is set according to the component manufacturing batch or the electrical load grading of the inverter string, and in operation, it is adaptively converged according to the standard deviation of the component score. The greater the fluctuation , the smaller the value, and the batch reevaluation is performed once a month. For example, if the standard deviation of the defect score of a component in the last 8 inspections is 0.12, the value of is set to 0.85; and for another component with greater fluctuation, the standard deviation is 0.25, the value of is set to 0.60. This can enhance the sensitivity of the trend to components with stable score fluctuations and suppress abnormal fluctuations that interfere with the judgment.

[0076] According to the value range of , the classification judgment of new faults, aggravated faults or no obvious changes is as follows: First, new fault: when the historical average defect score of the component is close to 0, it means that the component has not been identified as having obvious defects in previous inspections, i.e., "no historical fault"; at this time, if the current identification is significantly greater than 0, i.e., , then is also greater than a certain positive value, i.e., , indicating a new fault. Second, aggravated fault: when the component has a prior defect record, i.e., , and the current is significantly greater than the historical average, then is a larger positive value, i.e., , indicating that the defect is further expanding or deteriorating, and is determined as an aggravated fault. Third, no obvious change: if is basically the same as , i.e., the difference between the two is less than ±0.05, then is close to 0, i.e., , indicating that the fault state remains stable and there is no obvious change.

[0077] In one embodiment, step S300: generating an operation and maintenance task level according to the fault judgment result, and generating a task work order according to the operation and maintenance task level, includes: Step S310: constructing a task level mapping model according to the fault judgment result; Step S320: generating an operation and maintenance task level according to the task level mapping model, and generating a task work order according to the operation and maintenance task level.

[0078] In this embodiment, considering that in most current photovoltaic station operation and maintenance systems, the mapping relationship between fault identification results and actual operation and maintenance tasks is still extensive or highly dependent on manual intervention. In this application, the fault trend score is converted to the actual operation and maintenance task level by constructing a task level mapping model .

[0079] The task level mapping model is as follows:

[0080] wherein, is the task execution priority value of the kth component. is the fault trend score. is the risk exposure factor currently accumulated by the kth component. Specifically, by counting the number of hours in the "severe abnormality" state, the number of fault records and the number of times the temperature rise exceeds the threshold in the last 90 days, the scores are aggregated according to the weights, i.e. , which is calculated once a day in the early morning. is the risk adjustment coefficient set by the operation and maintenance system, which is a platform strategy parameter. The conventional value is 1.0, which automatically increases to 1.3 if it enters the high temperature season, and decreases to 0.8 if the station is in a low load maintenance period, to balance the maintenance resources.

[0081] The task level mapping model outputs , which will be used to automatically sort all component tasks. The higher the priority, the earlier the task order scheduling, that is, the operation and maintenance task level is generated according to the task execution priority value , and the subsequent task execution is deployed according to the operation and maintenance task level, that is, the task order is generated.

[0082] In actual deployment, users set the task level judgment strategy through the front-end platform, for example: set the components with high-temperature running time ≥ 30 days . If and , it is set as a red emergency task. In addition, a soft manual intervention mode is also supported, that is, the on-duty personnel can mark "delay" or "priority processing" for individual tasks, at which time the local weight of can be automatically adjusted based on the marking, thereby realizing manual intervention.

[0083] In this step, the "trend" and "cumulative" judgment indexes are fused, so that the situations of "dangerous trend intensifying but current impact is not big" and "current impact is big but has been handled" can be automatically identified, so that a more scene-adapted task response level is made.

[0084] which can directly drive the queue sorting and GIS layer positioning of the task order, and truly realize the closed-loop linkage from the identification result to the maintenance instruction. Adaptive and policy interface is provided to support automatic adjustment of operation and maintenance policy, such as emergency mode, night patrol policy switching, etc., to improve task flexibility and system decision-making ability.

[0085] In another embodiment, step S400: obtaining task execution data after executing the task order, and performing task feedback optimization according to the task execution data, the specific description includes: First, for the "order execution-model improvement" link of the existing technology, the off-line batch processing + manual annotation review method is generally used. However, the present application is different. After the task order is completed, a feedback optimization factor model is constructed to compare the with the re-output after processing, to calculate the task response effectiveness and model decision accuracy, and to perform feedback optimization on image fusion recognition accordingly.

[0086] The feedback optimization factor model is as follows:

[0087] Among them, is the error feedback factor of the kth component. is the defect comprehensive score. is the defect comprehensive score re-identified after operation and maintenance, that is, a new is obtained by running step S210 again. After the order is completed, the same image is taken on site and the model in step S210 is run again to output a new comprehensive score as a comparison benchmark. is a very small positive number, used to avoid division by zero error. is a constant, such as 1x10 -3 , to prevent the denominator from being zero. The model code is fixed and constant after deployment.

[0088] If , it is considered as accurate prediction, and the sample enters the "stable set"; 0.05 ≤ 0.20 enters the "mild error set" for light threshold fine-tuning; > 0.20 enters the "high error set", immediately increases the training weight of the sample, and locally amplifies the learning rate of the corresponding model layer by 10-30%, to ensure fast convergence.

[0089] From the user's perspective, the maintenance engineer only needs to upload "before and after maintenance comparison photos" and "processing notes" on the mobile terminal, and the platform will automatically package them as "feedback events". The whole calculation and model updating are completed in the background without feeling. The system's monitoring panel will display a line chart of error distribution in real time, which is convenient for managers to evaluate the model health. If a certain type of fault appears high , the platform will pop up a "data scarcity warning" to prompt the collection of special samples or adjustment of identification strategy. At the same time, the platform will also put Cross analysis with the task completion time, on-site retest photo clarity and other dimensions is used to evaluate the operation quality of the outsourcing maintenance team.

[0090] In this embodiment, after the completion of the task order, the method automatically converts the fault type, component number, and historical state information in the image recognition result into standardized semantic labels and links them to the pre-constructed photovoltaic operation and maintenance knowledge graph.

[0091] Specifically, in another embodiment, after completing the task scheduling and execution response closed loop, to further improve the knowledge accumulation ability and intelligent operation and maintenance level of the photovoltaic power station, after image recognition and order closed loop, the image semantic label extraction and knowledge graph linkage will be automatically started. Specifically, first, based on each recognized and confirmed image defect result, the following information fields are extracted, including but not limited to: component number, component coordinates, identification time, fault type (such as hot spot, crack, foreign matter obstruction, etc.), fault severity level, historical work order processing record, etc. The above fields will be formatted into semantic label items and automatically embedded in structured data nodes.

[0092] These structured labels will then be dynamically associated with the pre-set photovoltaic fault knowledge graph. This knowledge graph not only includes the type classification system of common faults of photovoltaic components, but also includes the cause analysis path associated therewith (such as "hot spot → poor solder strip contact → encapsulation aging"), the equipment operation background (such as "high temperature + high humidity → adhesive film delamination"), the historical maintenance method recommendation (such as "available non-destructive infrared detection review → cleaning → replacement") and the risk extension path (such as "local failure → busbar abnormality → string voltage unevenness") and the like.

[0093] The linkage process uses a combination of keyword mapping and graph structure traversal. When comparing keywords between the recognition semantic label and the nodes in the knowledge graph, the domain word library and the upper and lower relationship reasoning mechanism are called to ensure the accuracy of fuzzy matching and semantic expansion. For example, for the recognition label "frame corrosion", the system can associate with the "poor grounding" and "encapsulation damage" nodes and automatically supplement the possible causes and disposal schemes.

[0094] In addition, the knowledge graph will automatically recommend similar processing strategies for the current defect label based on the number of historical similar cases and the success rate of maintenance, and generate an "explanation chain" to explain the possible source of the fault, the evolution trend and the recommended action. Finally, these graph results are packaged into "intelligent reference suggestion packages" and attached to the work order in the form of a prompt card for on-site personnel to refer to for decision-making, reducing experience dependence.

[0095] In one embodiment, as Figure 2Also shown, there is provided an image recognition-based unmanned aerial vehicle intelligent operation and maintenance system for a photovoltaic power station, the system comprising: a patrol task generation module configured to divide the photovoltaic power station into zones and generate a patrol task queue, and dynamically adjust the patrol task queue based on real-time illumination; a fault judgment generation module configured to acquire real-time acquisition images when performing a patrol task according to the dynamically adjusted patrol task queue, generate a defect comprehensive score of a photovoltaic module based on the real-time acquisition images, and generate a fault judgment result based on the defect comprehensive score; a task work order generation module configured to generate an operation and maintenance task level based on the fault judgment result, and generate a task work order based on the operation and maintenance task level; a task execution feedback module configured to acquire task execution data after executing the task work order, and perform task feedback optimization based on the task execution data.

[0096] In another embodiment, the patrol task generation module is further configured to divide a component array of the photovoltaic power station into multiple zones, and generate a patrol task queue for each zone; and dynamically adjust the patrol task queue based on real-time illumination dynamic sorting.

[0097] In another embodiment, the patrol task generation module is further configured to divide a component array of the photovoltaic power station into multiple zones, construct a priority score model based on each zone; output a patrol priority score based on the priority score model, and generate a patrol task queue based on the patrol priority score.

[0098] In another embodiment, the patrol task generation module is further configured to acquire real-time illumination of a zone, and generate a current illumination interference factor based on the real-time illumination; generate a scheduling priority based on the current illumination interference factor and a patrol priority score corresponding to the zone; and dynamically adjust the patrol task queue based on the scheduling priority.

[0099] In another embodiment, the fault judgment generation module is further configured to acquire real-time acquisition images when performing a patrol task according to the dynamically adjusted patrol task queue, perform image fusion based on the real-time acquisition images, and generate a defect comprehensive score of a photovoltaic module; compare the defect comprehensive score with historical recognition data recorded in a photovoltaic power station database, and generate a fault judgment result.

[0100] In another embodiment, the fault judgment generation module is further configured to generate a fault judgment result based on the defect comprehensive score, and input the fault judgment result into a trend score update model based on historical recognition data recorded in a photovoltaic power station database; and generate a fault judgment result based on the trend score update model.

[0101] In another embodiment, the task order generation module is further configured to: construct a task level mapping model according to the fault judgment result; generate an operation and maintenance task level according to the task level mapping model, and generate a task order according to the operation and maintenance task level.

[0102] In one embodiment, as shown in FIG. 13, a computer device is also provided, which includes a memory and a processor, the memory stores a computer program and an operating system, and the processor implements the steps of the above-mentioned machine vision-based separation detection method for a separation screen surface when executing the computer program. The computer device further includes a system bus, an internal memory, a network structure, a display screen, an input device, and the like. Figure 3

[0103] In one embodiment, a computer readable storage medium is also provided, which stores a computer program, and the computer program implements the steps of the above-mentioned image recognition-based intelligent operation and maintenance method for a photovoltaic field station by a processor when executing the computer program.

[0104] It should be noted that the information interaction, execution process and the like between the above modules are based on the same concept as the method embodiments of the present application, and the specific functions and technical effects brought by them can be referred to the method embodiments part, which will not be repeated here.

[0105] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is taken as an example for illustration, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or software. In addition, the specific names of each functional unit and module are only for easy distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0106] It should be noted that the information interaction, execution process and the like between the above modules are based on the same concept as the method embodiments of the present application, and the specific functions and technical effects brought by them can be referred to the method embodiments part, which will not be repeated here.

[0107] ​Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the above described functions. Each functional unit or module in the embodiment can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit. In addition, the specific name of each functional unit or module is only for convenient distinction, and does not limit the protection scope of the present application. The specific working process of the unit or module in the system can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0108] The embodiments of the present application further provide a network device, comprising at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps in any of the above method embodiments when executing the computer program.

[0109] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program, wherein the computer program is executable by a processor to implement the steps in any of the above method embodiments.

[0110] The embodiments of the present application provide a computer program product, which, when running on a mobile terminal, enables the mobile terminal to implement the steps in any of the above method embodiments.

[0111] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the present application can implement all or part of the processes in the above-mentioned embodiment methods through a computer program to instruct relevant hardware to complete, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of each method embodiment described above can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms. The computer readable medium can at least include any entity or device capable of carrying the computer program code to the photographing device / terminal equipment, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium. For example, U disk, mobile hard disk, magnetic disk or optical disk, etc. In some jurisdictions, according to legislation and patent practice, the computer readable medium can not be an electrical carrier signal and a telecommunication signal.

[0112] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0113] Those skilled in the art can appreciate that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0114] In the embodiments provided by the present application, it should be understood that the disclosed apparatus / network device and method can be implemented in other ways. For example, the above-described apparatus / network device embodiments are merely schematic, for example, the division of the modules or units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0115] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e. may be located in one place, or may be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0116] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

[0117] An embodiment of the present application further provides a computer device, the computer device of the embodiment comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps in any of the above methods when executing the computer program.

[0118] The computer device can include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above description is an example of the computer device, and does not constitute a limitation on the computer device, and can include more or fewer components than the above description, or combine certain components, or different components, for example, can also include input / output devices, network access devices, etc.

[0119] The processor can be a central processing unit (CPU), and the processor can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0120] The memory can be an internal storage unit of the computer device in some embodiments, such as a hard disk or a memory of the computer device. The memory can also be an external storage device of the computer device in other embodiments, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, and the like. Further, the memory can include both an internal storage unit and an external storage device of the computer device. The memory is used to store an operating system, an application program, a boot loader, data, and other programs, such as program codes of the computer program, and the like. The memory can also be used to temporarily store data that has been output or is to be output.

[0121] The technical features of the above embodiments can be combined in any manner. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described, but it should be understood that any combination of the technical features is within the scope of the present disclosure as long as the combination does not result in a contradiction.

[0122] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the present application. It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these are within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. An image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method, characterized in that, The method comprises: Based on the photovoltaic station, the area is divided and the inspection task queue is generated, and the inspection task queue is dynamically adjusted based on the real-time illumination; Obtain the real-time acquisition image when performing the inspection task according to the dynamically adjusted inspection task queue, generate the defect comprehensive score of the photovoltaic module according to the real-time acquisition image, and generate the fault judgment result according to the defect comprehensive score; Generate the operation and maintenance task level according to the fault judgment result, and generate the task work order according to the operation and maintenance task level; Obtain the task execution data after executing the task work order, and perform task feedback optimization according to the task execution data.

2. The image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method according to claim 1, characterized in that, Based on the photovoltaic station, the area is divided and the inspection task queue is generated, and the inspection task queue is dynamically adjusted based on the real-time illumination, comprising: Based on the component array of the photovoltaic station, multiple areas are generated, and an inspection task queue is generated according to each area; The inspection task queue is dynamically adjusted according to the real-time illumination dynamic sorting.

3. The photovoltaic station drone intelligent operation and maintenance method based on image recognition according to claim 2 is characterized in that: Based on the component array of the photovoltaic station, multiple areas are generated, and an inspection task queue is generated according to each area, comprising: Based on the component array of the photovoltaic station, multiple areas are generated, and an inspection task queue is generated according to each area, comprising: According to the priority score model, the inspection priority score is output, and the inspection task queue is generated according to the inspection priority score.

4. The image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method according to claim 2, characterized in that, According to the real-time illumination dynamic sorting, the inspection task queue is dynamically adjusted, comprising: Obtain the real-time illumination of the area, and generate the current illumination interference factor according to the real-time illumination; Generate the scheduling priority according to the current illumination interference factor and the inspection priority score corresponding to the area; The inspection task queue is dynamically adjusted according to the scheduling priority.

5. The image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method according to claim 1, characterized in that, Obtain the real-time acquisition image when performing the inspection task according to the dynamically adjusted inspection task queue, generate the defect comprehensive score of the photovoltaic module according to the real-time acquisition image, and generate the fault judgment result according to the defect comprehensive score, comprising: Obtain the real-time acquisition image when performing the inspection task according to the dynamically adjusted inspection task queue, generate the defect comprehensive score of the photovoltaic module according to the real-time acquisition image, and generate the fault judgment result according to the defect comprehensive score, comprising: According to the defect comprehensive score, the fault judgment result is compared with the historical identification data recorded in the photovoltaic station database, and the fault judgment result is generated.

6. The image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method according to claim 5, characterized in that, According to the defect comprehensive score, the fault judgment result is compared with the historical identification data recorded in the photovoltaic station database, and the fault judgment result is generated, comprising: According to the defect comprehensive score, the fault judgment result is compared with the historical identification data recorded in the photovoltaic station database, and the fault judgment result is generated. According to the defect comprehensive score, the fault judgment result is compared with the historical identification data recorded in the photovoltaic station database, and the fault judgment result is generated.

7. The image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance method according to claim 1, characterized in that, According to the fault judgment result, the operation and maintenance task level is generated, and the task work order is generated according to the operation and maintenance task level, comprising: According to the fault judgment result, the task level mapping model is constructed; According to the task level mapping model, the operation and maintenance task level is generated, and the task work order is generated according to the operation and maintenance task level.

8. An image recognition-based photovoltaic power station unmanned aerial vehicle intelligent operation and maintenance system, characterized in that, The system comprises: The inspection task generation module is configured to divide a photovoltaic station into zones and generate an inspection task queue based on the photovoltaic station, and dynamically adjust the inspection task queue based on real-time illumination; The fault judgment generation module is configured to acquire real-time collected images when performing the inspection task according to the dynamically adjusted inspection task queue, generate a defect comprehensive score of the photovoltaic module according to the real-time collected images, and generate a fault judgment result according to the defect comprehensive score; The task work order generation module is configured to generate an operation and maintenance task level according to the fault judgment result, and generate a task work order according to the operation and maintenance task level; The task execution feedback module is configured to acquire task execution data after executing the task work order, and perform task feedback optimization according to the task execution data. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 7.

Images