Infrared-based unmanned aerial vehicle photovoltaic panel subfissure detection system

Abstract

The invention belongs to the technical field of photovoltaic panel detection, and discloses an infrared-based unmanned aerial vehicle photovoltaic panel subfissure detection system. The range determining module is used for determining a monitoring range and constructing a monitoring surface, the monitoring surface is formed by connecting the central points of the surfaces of photovoltaic panels, and each monitoring point is associated with the serial number, specification and shape of the photovoltaic panel and the association relationship with other panels; the path setting module is used for generating a flight surface based on the monitoring surface, planning a flight path of the unmanned aerial vehicle through a weighted directed graph minimization objective function, and optimizing a flight path steering mode in combination with a multi-modal steering decision function; according to the invention, the photovoltaic panel with hidden cracks can be detected, and the accuracy of unmanned aerial vehicle positioning and infrared image comparison can be improved when the unmanned aerial vehicle detects the photovoltaic panel; moreover, through the setting of the flight path, the unmanned aerial vehicle can efficiently complete the infrared image collection of the photovoltaic panel in the monitoring range, and the detection efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic panel detection, and more specifically, to an infrared-based unmanned aerial vehicle photovoltaic panel hidden crack detection system. Background Art

[0002] As global demand for clean energy continues to rise, the photovoltaic industry, a core sector of the renewable energy sector, is experiencing rapid growth. Large-scale photovoltaic power plants are springing up like mushrooms after a rain, and the number of photovoltaic panel installations has also increased dramatically. These power plants act like giant "sunlight collectors," continuously converting sunlight into electricity. However, during this energy conversion process, subtle damage, known as cracks, can develop within the cells (panels) of photovoltaic modules. When current flows through these cracks, resistance increases, generating heat and triggering the hot spot effect. This highly harmful hot spot effect can not only completely destroy the cells, but can also burn the entire module, or even cause serious safety hazards such as fires.

[0003] Furthermore, hidden cracks in photovoltaic panels will not only reduce the power generation efficiency of photovoltaic panels, but in serious cases, it will also cause damage to the solar cells, thereby affecting the stable operation of the entire photovoltaic power station. Therefore, how to efficiently and accurately detect hidden cracks in photovoltaic panels has become a key problem that needs to be overcome in the process of achieving sustainable development of the photovoltaic industry.

[0004] In view of this, the present invention proposes an infrared-based UAV photovoltaic panel hidden crack detection system to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned objectives, the present invention provides the following technical solutions: an infrared-based UAV photovoltaic panel hidden crack detection system, comprising: Range determination module: determines the monitoring range and constructs a monitoring surface. The monitoring surface is composed of lines connecting the center points of the photovoltaic panel surface. Each monitoring point is associated with the photovoltaic panel number, specifications, shape and its relationship with other panels. Path setting module: Generates a flight surface based on the monitoring surface, plans the UAV flight path by minimizing the objective function through a weighted directed graph, and optimizes the flight path steering method by combining a multi-modal steering decision function; Transmission comparison module: Controls the drone to collect infrared images along the flight path, establishes a transmission channel from server to receiving point to collection azimuth point, and determines whether there are hidden cracks in the photovoltaic panel through temperature difference analysis and morphological characteristics.

[0006] Furthermore, the range determination module constructs a monitoring surface by the following steps: The geographical range corresponding to all photovoltaic panels monitored by the system is used as the monitoring range; Taking the center point of the photovoltaic panel surface as the monitoring point, all monitoring points within the monitoring range are connected to form a plane or non-plane monitoring surface; The characteristic information of each monitoring point is associated, and the characteristic information includes number, specification, shape and spatial relationship of adjacent panels.

[0007] Furthermore, the path setting module generates a flight path through the following steps: Taking the monitoring surface as the reference surface, copy and move it upward to generate the flight surface; Construct a weighted directed graph, where the nodes correspond to the starting point, end point, and collection location of the drone, and the edge weight of the weighted directed graph is the sum of the distance and the number of turns; The objective function is minimized by the Dijkstra algorithm to generate a flight path with the shortest total flight distance and the least number of turns.

[0008] Furthermore, the multimodal steering decision function is: , Where, Indicates steering judgment, Indicates the time the drone makes an arc turn between the turning start point and the turning end point and flies along the turning edge. Indicates the time the drone makes a pivot turn at any point between the turning start point and the turning end point and flies along the turning edge.

[0009] Furthermore, the establishment of a transmission channel from server to receiving point to location for collecting azimuths includes: Receiving points corresponding to monitoring points are arranged in the server, and monitoring points corresponding to the receiving points and monitoring points are arranged in the server. A transmission channel is established between the acquisition azimuth points corresponding to the points.

[0010] Furthermore, the determination of whether there are hidden cracks in the photovoltaic panel through temperature difference analysis and morphological characteristics includes: A multi-level mapping relationship is established on the server side. The infrared image corresponding to the acquisition point is defined as the upper point, the associated monitoring point is defined as the intermediate mapping point, and the individual panels of the photovoltaic panel are defined as the lower point and encoded. This achieves hierarchical correspondence between the upper point and multiple lower points through the intermediate mapping points. Perform geometric correction on the infrared image, align the image with the monitoring point as the reference, extract the temperature difference between each lower point and its adjacent lower point in the corrected image, and form a temperature difference set ΔT={ΔT1,ΔT2,...,ΔTn}; Normal panel determination: If all temperature differences in the temperature difference set ΔT are within the preset temperature difference threshold range [ΔTmin, ΔTmax], the corresponding photovoltaic panel is determined to be a normal panel; Abnormal panel determination: If there is at least one temperature difference in the temperature difference set ΔT that exceeds a preset temperature difference threshold range, the corresponding photovoltaic panel is determined to be an abnormal panel.

[0011] Furthermore, the performing geometric correction on the infrared image and aligning the image based on the monitoring point as a reference includes: Taking the monitoring point as the center point, the edge of the photovoltaic panel and the intersection point are used as auxiliary correction lines; The infrared image and the monitoring surface coordinate system are aligned through affine transformation to ensure accurate mapping between the infrared image and the physical location.

[0012] Furthermore, the abnormal panel determination further includes: The temperature difference exceeding the preset temperature difference threshold range is regarded as an abnormal temperature difference, and the lower point corresponding to the abnormal temperature difference is extracted as an abnormal feature point; Perform morphological analysis on abnormal feature points. If their distribution characteristics conform to the typical characteristics of hidden cracks and the abnormal temperature difference ΔTabnormal is greater than 0°C, it is determined that the photovoltaic panel has hidden cracks. On the contrary, if the abnormal temperature difference ΔTabnormal ≤ 0°C or the distribution characteristics do not conform to the typical characteristics of micro-cracks, it is judged as other abnormalities.

[0013] Furthermore, the flight path of the UAV is also associated with an update rule, and the update rule specifically includes: when the monitoring point changes, the flight path of the UAV is updated according to the preset rule.

[0014] Furthermore, the preset temperature difference threshold interval is calibrated by the following steps: collecting infrared images of normal panels, calculating the mean μ and standard deviation σ of the temperature difference between adjacent panels; and setting the temperature difference threshold interval to [μ-3σ, μ+3σ], that is, ΔTmin is μ-3σ, and ΔTmax is μ+3σ.

[0015] The technical effects and advantages of the infrared-based UAV photovoltaic panel hidden crack detection system of the present invention are as follows: 1. By setting the monitoring surface, collection points, receiving points, and corresponding transmission channels, the UAV can be accurately positioned and incorrect comparisons caused by inaccurate positioning and transmission errors of infrared images can be avoided, thus improving the accuracy of UAV positioning and infrared image comparison; 2. By mapping and encoding the upper, intermediate, and lower points within the server, the correspondence between the infrared image and multiple panels on the photovoltaic panel can be determined, enabling the system to quickly locate the panel with hidden cracks, improving the accuracy and efficiency of locating hidden cracks on the photovoltaic panel. By comparing the temperature difference between the lower point and adjacent lower points in the infrared image with a preset temperature difference threshold range, it can be determined whether the photovoltaic panel is normal. By obtaining abnormal feature points and performing morphological analysis, it can be determined whether the photovoltaic panel has hidden cracks or other abnormalities. 3. The set flight path enables the UAV to efficiently complete the infrared image acquisition of photovoltaic panels within the monitoring range, improving the accuracy and reliability of flight path planning; at the same time, the turning method corresponding to each turning number is decided through the multimodal conditional decision function to select the turning method that consumes less time, which can save the time spent by the UAV in turning, improve the efficiency of the UAV in collecting infrared images within the monitoring range, and reduce the flight time of the UAV on the flight path. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a structural schematic diagram of an infrared-based UAV photovoltaic panel hidden crack detection system of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0018] Application scenarios: During the operation of a photovoltaic power station, hidden cracks in photovoltaic panels not only reduce the panels' power generation efficiency but, in severe cases, can also damage the cells, further impacting the stable operation of the entire photovoltaic power station. Efficiently and accurately detecting hidden cracks in photovoltaic panels has become a key challenge that the photovoltaic industry urgently needs to overcome. Furthermore, when inspecting photovoltaic panels for hidden cracks, drones equipped with infrared cameras are typically used to capture images of the panels and determine if any hidden cracks are present. However, the monitoring area of a photovoltaic power station is generally large, necessitating more efficient flight paths and accurate image comparison when drones capture infrared images. Based on this, the present invention proposes an infrared-based UAV photovoltaic panel hidden crack detection system to solve the above problems.

[0019] See also Figure 1As shown, the infrared-based drone photovoltaic panel hidden crack detection system described in this embodiment includes a range determination module, a path setting module, and a transmission comparison module: The range determination module is used to determine the monitoring range and construct a monitoring surface. The monitoring surface is composed of lines connecting the center points of the photovoltaic panel surface. Each monitoring point is associated with the number, specification, shape of the photovoltaic panel and its relationship with other panels. The above-mentioned path setting module generates a flight surface based on the monitoring surface, plans the UAV flight path by minimizing the objective function through a weighted directed graph, and optimizes the flight path steering method in combination with a multimodal steering decision function. It should be noted that by setting the objective function of the flight path and the multimodal conditional decision function, the flight risk caused by frequent steering can be reduced and the time spent by the UAV on steering can be saved, allowing the UAV to efficiently complete the infrared image acquisition of photovoltaic panels within the monitoring range, thereby improving the accuracy and reliability of flight path planning. The above-mentioned transmission comparison module is used to control the UAV to collect infrared images according to the flight path, establish a transmission channel from server to receiving point to collection azimuth point, and determine whether there are hidden cracks in the photovoltaic panel through temperature difference analysis and morphological characteristics; it should be noted that by arranging receiving points corresponding to monitoring points in the server, establishing a transmission channel between the monitoring points corresponding to the receiving points and the collection azimuth points corresponding to the monitoring points, the UAV can be accurately positioned and incorrect comparisons caused by inaccurate positioning of infrared images and transmission errors can be avoided. It can also be determined whether there are hidden cracks or other abnormalities in the photovoltaic panel.

[0020] In the above embodiment, by setting the monitoring surface, collection azimuth points, receiving points and corresponding transmission channels, the accuracy of drone positioning and infrared image comparison can be improved; by setting the flight path, the drone can efficiently complete the infrared image collection of photovoltaic panels within the monitoring range, thereby improving the accuracy and reliability of flight path planning. At the same time, by using a multimodal conditional decision function to decide the turning method corresponding to each turning number, the efficiency of the drone in collecting infrared images within the monitoring range can be improved, and the flight time of the drone on the flight path can be reduced; by establishing the mapping and encoding of the upper point, the middle mapping point and the lower point in the server, the system can quickly locate the specific solar panel with hidden cracks, thereby improving the accuracy and efficiency of the positioning of hidden cracks on the photovoltaic panel. By comparing the temperature difference between the lower point and the adjacent lower point of the infrared image with the preset temperature difference threshold range, it can be determined whether the photovoltaic panel is a normal photovoltaic panel. By obtaining abnormal feature points and morphological analysis, it can be determined whether the photovoltaic panel has hidden cracks or other abnormalities.

[0021] As an optional embodiment, the range determination module constructs a monitoring surface through the following steps: The geographical range corresponding to all photovoltaic panels monitored by the system is used as the monitoring range; Taking the center point of the photovoltaic panel surface as the monitoring point, all monitoring points within the monitoring range are connected to form a plane or non-plane monitoring surface; Associating characteristic information of each monitoring point, the characteristic information including number, specification, shape and spatial relationship of adjacent panels; It should be noted that a single photovoltaic panel or multiple adjacent photovoltaic panels can be considered a monitoring point (the size of the photovoltaic panel here should be based on the maximum area that can be captured by the drone at the optimal shooting altitude, and the resolution of the infrared image must meet the requirements). By using the surface center point of each photovoltaic panel (single or multiple) as the monitoring point and connecting the monitoring points corresponding to all photovoltaic panels within the monitoring range, a monitoring surface is obtained. If the monitoring points of all photovoltaic panels are on the same horizontal plane, the monitoring surface is planar; if the monitoring points of all photovoltaic panels are not on the same horizontal plane, the monitoring surface is non-planar. Therefore, the monitoring surface can reflect the geographical distribution and location distribution of the photovoltaic panels within the monitoring range. The characteristic information includes the photovoltaic panel number, specifications, shape, and spatial relationship with adjacent photovoltaic panels. The spatial relationship includes orientation information and distance information. This spatial relationship can further help determine the location information of the photovoltaic panel. Through the characteristic information of the photovoltaic panel and the setting of the monitoring points, the monitoring points can reflect all the information of the corresponding photovoltaic panel, facilitating the comparison and judgment of subsequent infrared images collected by the drone, so as to detect whether the photovoltaic panel corresponding to the infrared image has hidden cracks.

[0022] As an optional embodiment: the path setting module generates a flight path through the following steps: Taking the monitoring surface as the reference surface, copy and move it upward to generate the flight surface; Construct a weighted directed graph, where the nodes correspond to the starting point, end point, and collection location of the drone, and the edge weight of the weighted directed graph is the sum of the distance and the number of turns; Minimize the objective function through the Dijkstra algorithm to generate a flight path with the shortest total flight distance and the least number of turns; Specifically, the drone carries an infrared camera, and the height at which the drone can capture infrared images of the entire photovoltaic panel and meet the required image resolution is used as the optimal shooting height of the drone. By taking the monitoring surface as the reference surface, which is not necessarily a plane, the reference surface is copied and moved upward according to the optimal shooting height of the drone to generate a flight surface, so that the drone is at the optimal shooting height when collecting each infrared image; the optimal azimuth point is determined according to the orientation and layout of the photovoltaic panel and the camera parameters of the infrared camera, and this optimal azimuth point is used as the collection azimuth point of each photovoltaic panel collected by the drone, so that the infrared images collected by the drone can be more uniform, which is convenient for subsequent comparison of infrared images; It should be noted that in the flight path, the flight speed of the drone is set to be constant, and the lines between the nodes are connected as edges. The flight time of the drone on this edge is obtained by dividing the distance of each edge by the speed. If the monitoring area of the system is large, the monitoring area can be divided according to the average flight range of the drone (the range corresponding to the average flight time). After the division, the photovoltaic panels are inspected according to the system; the starting point, end point and each collection azimuth point of the drone are taken as a node, and the lines between the two nodes form edges and a weighted directed graph is constructed. The distance and number of turns corresponding to each edge are used as the weight factors of each edge in the weighted directed graph. Then, the laboratory calibration method (such as regression analysis of the historical data of the drone in the flight path) is used to determine the reasonable ratio of the weight coefficients corresponding to the distance and the number of turns (such as 6:4). The impact of historical data such as distance and number of turns on the total inspection time, based on the above-mentioned edges and edge weight factors, the objective function is set. The objective function is expressed as , where bi represents the weight of the i-th edge, K represents the number of all edges, and i represents the index of the edge. represents the sum of K bi, Represents the flight path corresponding to the minimized objective function; the expression of bi is: , where JLi represents the distance of the i-th edge, ωJL represents the weight coefficient corresponding to the distance, ZXi represents the number of turns of the i-th edge, and ωZX represents the weight coefficient corresponding to the number of turns; the Dijkstra algorithm or the A* algorithm is used to minimize the objective function (i.e., find the shortest path of the weighted directed graph). By minimizing the objective function, a flight path with the shortest total flight distance and the least number of turns can be set, thereby improving the flight efficiency of the UAV and reducing the flight risk caused by frequent turns. The set flight path can make the UAV more in line with the actual flight situation when collecting infrared images, so that the UAV can efficiently complete the infrared image collection of photovoltaic panels within the monitoring range, thereby improving the accuracy and reliability of flight path planning.

[0023] As an optional embodiment: the multimodal steering decision function is: A multimodal condition decision function is set, and the steering mode of the UAV is determined according to the multimodal condition decision function. The multimodal condition decision function is: , Where, Indicates steering judgment, Indicates the time the drone makes an arc turn between the turning start point and the turning end point and flies along the turning edge. Indicates the time the drone makes a pivot turn at any point between the turning start point and the turning end point and flies along the turning edge. It should be noted that arc turning and in-place turning represent the corresponding turning methods of the drone, respectively. The turning edge, turning starting point, and turning end point are the edge corresponding to each turning number and the two nodes corresponding to the edge. The time expression for in-place turning and flying on the turning edge is: , Thover represents the hovering stability time, which is generally 0.5~1 second and is obtained by experience or set value; Δθ represents the required steering angle (∘) of the UAV, which is obtained by the difference between the angle of the UAV at the starting point of the steering and the attitude angle of the UAV at the end of the steering (measured by the gyroscope, etc.); Krotate represents the in-situ steering rate, which is generally obtained from the UAV performance parameter manual or experimental test, and the unit is generally degrees per second (∘ / s); JLzxb represents the turning edge distance, which is generally measured in meters (m); v represents the flight speed of the UAV, which is measured in real time by the UAV's speed sensor (such as airspeed meter, GPS, etc.), and the unit is generally meters per second (m / s); the expression for the arc steering and flying time on the turning edge is: , where L represents the arc length of the arc turn, which is generally obtained from the trajectory length of the UAV flight, and the unit is generally meter (m); , , , is the turning radius, generally in meters (m), The arc length corresponds to the radian (rad). The steering angle can be converted into radians (rad). The arc length can be determined based on the determined radian (rad) and turning radius. c represents the chord length corresponding to the arc turn, and the unit is generally meters (m). ; amax represents the maximum centripetal acceleration of the drone, and the unit is generally meters per second squared (m / s2). amax can be dynamically estimated through real-time sensor data; when the drone needs to turn during flight, the turning method corresponding to each turning number is decided through the multimodal conditional decision function to select the turning method that consumes less time, which can save the time spent by the drone in turning, thereby improving the efficiency of the drone in collecting infrared images within the monitoring range and reducing the flight time of the drone on the flight path.

[0024] As an optional embodiment: the step of establishing a transmission channel includes: Receiving points corresponding to the monitoring points are arranged in the server, and a transmission channel is established between the monitoring points corresponding to the receiving points and the collection azimuth points corresponding to the monitoring points; It should be noted that by arranging receiving points in the server (receiving points are virtual points), when the drone transmits the collected infrared images, each infrared image corresponds to a unique collection orientation point, monitoring point and receiving point. The transmission channel formed by these three points can make the collected infrared image have a unique transmission channel during transmission, which can facilitate the mapping of this infrared image to the corresponding multiple lower points and comparison, so as to facilitate the correct positioning and comparison of this infrared image, avoid the incorrect comparison caused by inaccurate positioning of the infrared image and transmission errors, and thus improve the accuracy of drone positioning and infrared image comparison; furthermore, the drone collects infrared images according to the preset flight path, which can quickly and comprehensively cover the monitoring area. Photovoltaic panels in the domain improve the detection efficiency of photovoltaic panels. By establishing a transmission channel, the transmission channel constructed by the drone can be used to transmit after the drone collects infrared images. The drone can be positioned when transmitting infrared images, and the photovoltaic panels corresponding to the collected infrared images can be matched, which can solve the problems of inaccurate drone positioning and mismatched infrared image transmission. In addition, after the receiving point receives the infrared image, the monitoring point corresponding to the receiving point is covered. By covering or marking the corresponding monitoring point after the receiving point receives the infrared image, it is easy to determine whether the photovoltaic panel corresponding to each monitoring point has been imaged, thereby avoiding omission of infrared image collection corresponding to the photovoltaic panel within the monitoring range.

[0025] As an optional embodiment, the method of determining whether a photovoltaic panel has hidden cracks by temperature difference analysis and morphological characteristics includes: A multi-level mapping relationship is established on the server side. The infrared image corresponding to the acquisition point is defined as the upper point, the associated monitoring point is defined as the intermediate mapping point, and the individual panels of the photovoltaic panel are defined as the lower point and encoded. This achieves hierarchical correspondence between the upper point and multiple lower points through the intermediate mapping points. Perform geometric correction on the infrared image, align the image with the monitoring point as the reference, extract the temperature difference between each lower point and its adjacent lower point in the corrected image, and form a temperature difference set ΔT={ΔT1,ΔT2,...,ΔTn}; Normal panel determination: If all temperature differences in the temperature difference set ΔT are within the preset temperature difference threshold range [ΔTmin, ΔTmax], the corresponding photovoltaic panel is determined to be a normal panel; Abnormal panel determination: If at least one temperature difference in the temperature difference set ΔT exceeds the preset temperature difference threshold range, the corresponding photovoltaic panel is determined to be an abnormal panel; It should be noted that the lower points can also be in a grid shape, and their specifications can correspond to the specifications of a single solar panel. By establishing the mapping and encoding of the upper points, intermediate mapping points and lower points in the server, the correspondence between the infrared image and the multiple solar panels on the photovoltaic panel can be determined, so that the system can quickly locate the specific solar panel with hidden cracks, which can improve the accuracy and efficiency of locating hidden cracks on the photovoltaic panel; by mapping the infrared image to multiple lower points and obtaining the temperature difference set between each lower point and the adjacent lower points, the lower points that may have abnormalities can be quickly screened out, thereby improving the detection efficiency; by setting a preset temperature difference threshold interval, when all the temperature differences in the temperature difference set are within the temperature difference threshold interval, the photovoltaic panel is judged to be a normal panel (that is, the photovoltaic panel is in a normal state); conversely, when any temperature difference exceeds the preset temperature difference threshold interval, it is judged to be an abnormal panel (that is, the photovoltaic panel has an abnormality), which can quickly screen out abnormal panels that may have abnormalities, greatly shortening the detection time and improving the detection efficiency.

[0026] As an optional embodiment, the step of geometrically correcting the infrared image and aligning the image based on the monitoring point includes: Taking the monitoring point as the center point, the edge of the photovoltaic panel and the intersection point are used as auxiliary correction lines; Align the infrared image with the monitoring surface coordinate system through affine transformation to ensure accurate mapping between the infrared image and the physical location; It should be noted that by taking the monitoring point as the center point and using the edge and edge intersection of the photovoltaic panel as auxiliary correction lines and auxiliary correction points for correction, it is possible to facilitate the subsequent mapping of infrared images to the corresponding multiple lower points, and facilitate the comparison of each lower point (single solar panel) in the infrared image mapped on the multiple lower points with the adjacent lower points, so as to facilitate the subsequent detection of whether there are hidden cracks in the photovoltaic panel.

[0027] As an optional embodiment: the abnormal panel determination further includes: The temperature difference exceeding the preset temperature difference threshold range is regarded as an abnormal temperature difference, and the lower point corresponding to the abnormal temperature difference is extracted as an abnormal feature point; Perform morphological analysis on abnormal feature points. If their distribution characteristics conform to the typical characteristics of hidden cracks and the abnormal temperature difference ΔTabnormal is greater than 0°C, it is determined that the photovoltaic panel has hidden cracks. On the contrary, if the abnormal temperature difference ΔTabnormal ≤ 0°C or the distribution characteristics do not conform to the typical characteristics of micro-cracks, it is judged as other abnormalities; It should be noted that the abnormal temperature difference is the temperature difference corresponding to the abnormal feature point. By obtaining the distribution characteristics of the abnormal panel and whether the temperature difference between the abnormal feature point and the adjacent lower point is greater than zero, it is possible to accurately distinguish whether it is a hidden crack fault or other abnormalities (such as obstruction by bird droppings, leaves, and other debris). Among them, under normal circumstances, hidden cracks in photovoltaic panels will cause a hot spot effect (that is, the temperature of the abnormal feature point is greater than the temperature of the adjacent lower point, that is, the abnormal temperature difference ΔTabnormal>0°C), and the morphological distribution characteristics of the abnormal feature point will be spiderweb-like or tree-like, which are typical characteristics of hidden cracks. In contrast, for other abnormalities (such as obstruction by bird droppings, leaves, and other debris), the temperature of the abnormal feature point corresponding to the abnormal feature point is lower than the temperature of other adjacent lower points (abnormal temperature difference ΔTabnormal≤0°C), and the morphological distribution characteristics are generally not spiderweb-like or tree-like. Furthermore, by performing morphological analysis on the abnormal feature points, it is possible to promptly and accurately determine whether the corresponding photovoltaic panel has hidden cracks or other abnormalities, facilitating subsequent relevant personnel to take appropriate treatment measures.

[0028] As an optional embodiment: the flight path of the UAV is further associated with an update rule, and the update rule specifically includes: when a monitoring point changes (such as a monitoring point is added or deleted), the flight path of the UAV is updated according to the preset rule; It should be noted that the flight path is associated with an update rule, and the update rule is to update the flight path when the monitoring point changes. The update path can be updated after the photovoltaic panels within the monitoring range change, so that the flight path of the drone always fits the actual geographical distribution of the photovoltaic panels within the monitoring range, thereby improving the efficiency of detecting hidden cracks in the photovoltaic panels within the monitoring range.

[0029] As an optional embodiment: the preset temperature difference threshold interval is calibrated by the following steps: collecting infrared images of normal panels, calculating the mean μ and standard deviation σ of the temperature difference between adjacent solar panels; the temperature difference threshold interval is set to [μ-3σ, μ+3σ], that is, ΔTmin is μ-3σ, and ΔTmax is μ+3σ.

[0030] It should be noted that under normal operating conditions, the temperature distribution of adjacent panels of a photovoltaic panel usually conforms to a normal distribution (Gaussian distribution); when there are no hidden cracks or other anomalies, the temperature fluctuation amplitude is small, and the vast majority of data points are concentrated near the mean; 3σ principle: According to statistical theory, for normally distributed data, approximately 99.7% of the data points will fall within the range of ±3 times the standard deviation (σ) of the mean, and temperature differences outside this range are extremely likely to be abnormal (such as the hot spot effect caused by hidden cracks); threshold sensitivity balance: mean (μ): reflects the average temperature difference under normal operating conditions, eliminating the influence of ambient temperature fluctuations, standard deviation (σ): quantifies the fluctuation amplitude of normal temperature differences, and is used to define the strictness of anomalies; ±3σ range: Experimental data show that taking a ±3σ range can not only cover the vast majority of normal situations (false positive rate ≤ 0.3%), but also sensitively capture significant temperature difference abnormalities caused by hidden cracks.

[0031] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0032] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only one type. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0033] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

[0034] Finally: 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 in the scope of protection of the present invention.

Claims

1. An infrared-based UAV photovoltaic panel hidden crack detection system, characterized in that: include: Range determination module: determines the monitoring range and constructs a monitoring surface. The monitoring surface is composed of lines connecting the center points of the photovoltaic panel surface. Each monitoring point is associated with the photovoltaic panel number, specifications, shape and its relationship with other panels. Path setting module: Generates a flight surface based on the monitoring surface, plans the UAV flight path by minimizing the objective function through a weighted directed graph, and optimizes the flight path steering method by combining a multi-modal steering decision function; Transmission comparison module: Controls the drone to collect infrared images along the flight path, establishes a transmission channel from server to receiving point to collection azimuth point, and determines whether there are hidden cracks in the photovoltaic panel through temperature difference analysis and morphological characteristics.

2. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 1 is characterized in that: The scope determination module constructs the monitoring surface by the following steps: The geographical range corresponding to all photovoltaic panels monitored by the system is used as the monitoring range; Taking the center point of the photovoltaic panel surface as the monitoring point, all monitoring points within the monitoring range are connected to form a plane or non-plane monitoring surface; The characteristic information of each monitoring point is associated, and the characteristic information includes number, specification, shape and spatial relationship of adjacent panels.

3. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 1 is characterized in that: The path setting module generates a flight path by the following steps: Taking the monitoring surface as the reference surface, copy and move it upward to generate the flight surface; Construct a weighted directed graph, where the nodes correspond to the starting point, end point, and collection location of the drone, and the edge weight of the weighted directed graph is the sum of the distance and the number of turns; The objective function is minimized by the Dijkstra algorithm to generate a flight path with the shortest total flight distance and the least number of turns.

4. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 3 is characterized in that: The multimodal steering decision function is: ， Where, Indicates steering judgment, Indicates the time the drone makes an arc turn between the turning start point and the turning end point and flies along the turning edge. Indicates the time the drone makes a pivot turn at any point between the turning start point and the turning end point and flies along the turning edge.

5. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 4 is characterized in that: The establishment of the transmission channel from server to receiving point to collecting position includes: Receiving points corresponding to monitoring points are arranged in the server, and monitoring points corresponding to the receiving points and monitoring points are arranged in the server. A transmission channel is established between the acquisition azimuth points corresponding to the points.

6. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 1, It is characterized by: The determination of whether there are hidden cracks in the photovoltaic panel through temperature difference analysis and morphological characteristics includes: A multi-level mapping relationship is established on the server side. The infrared image corresponding to the acquisition point is defined as the upper point, the associated monitoring point is defined as the intermediate mapping point, and the individual panels of the photovoltaic panel are defined as the lower point and encoded. This achieves hierarchical correspondence between the upper point and multiple lower points through the intermediate mapping points. Perform geometric correction on the infrared image, align the image with the monitoring point as the reference, extract the temperature difference between each lower point and its adjacent lower point in the corrected image, and form a temperature difference set ΔT={ΔT1,ΔT2,...,ΔTn}; Normal panel determination: If all temperature differences in the temperature difference set ΔT are within the preset temperature difference threshold range [ΔTmin, ΔTmax], the corresponding photovoltaic panel is determined to be a normal panel; Abnormal panel determination: If there is at least one temperature difference in the temperature difference set ΔT that exceeds a preset temperature difference threshold range, the corresponding photovoltaic panel is determined to be an abnormal panel.

7. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 6, It is characterized by: The geometric correction of the infrared image and alignment of the image based on the monitoring point include: Taking the monitoring point as the center point, the edge of the photovoltaic panel and the intersection point are used as auxiliary correction lines; The infrared image and the monitoring surface coordinate system are aligned through affine transformation to ensure accurate mapping between the infrared image and the physical location.

8. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 7, characterized in that: The abnormal panel determination further includes: The temperature difference exceeding the preset temperature difference threshold range is regarded as an abnormal temperature difference, and the lower point corresponding to the abnormal temperature difference is extracted as an abnormal feature point; Perform morphological analysis on abnormal feature points. If their distribution characteristics conform to the typical characteristics of hidden cracks and the abnormal temperature difference ΔTabnormal is greater than 0°C, it is determined that the photovoltaic panel has hidden cracks. On the contrary, if the abnormal temperature difference ΔTabnormal ≤ 0°C or the distribution characteristics do not conform to the typical characteristics of micro-cracks, it is judged as other abnormalities.

9. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 5, characterized in that: The flight path of the UAV is also associated with an update rule, and the update rule specifically includes: when the monitoring point changes, the flight path of the UAV is updated according to the preset rule.

10. The infrared-based UAV photovoltaic panel hidden crack detection system according to claim 6, characterized in that: The preset temperature difference threshold interval is calibrated by the following steps: collecting infrared images of normal panels, calculating the mean μ and standard deviation σ of the temperature difference between adjacent solar panels; setting the temperature difference threshold interval to [μ-3σ, μ+3σ], that is, ΔTmin is μ-3σ, and ΔTmax is μ+3σ.

Images