Satellite image-based photovoltaic panel position and orientation determination method and device

Abstract

This invention provides a method and apparatus for determining the location and orientation of photovoltaic (PV) panels based on satellite imagery. The method includes: processing satellite imagery using a selected recognition algorithm to obtain images of all suspected PV panels; determining the pixel coordinates of the two endpoints of the central axis of each PV panel and the pixel coordinates of the center point of the central axis; using the latitude and longitude coordinates of the center point of the central axis of each PV panel as its location; and determining the arrangement direction of each PV panel based on the latitude and longitude coordinates of the two endpoints of the central axis, using the direction perpendicular to its arrangement direction as its orientation. This invention provides a method and apparatus for determining the location and orientation of PV panels based on satellite imagery, addressing the problem in existing technologies where it is difficult to obtain accurate positioning and orientation of large quantities of PV panels, thus hindering new energy power generation companies from achieving refined production, construction, and operation management, and preventing them from gaining a competitive edge in the electricity market.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method and apparatus for determining the position and orientation of photovoltaic panels based on satellite imagery. Background Technology

[0002] For new energy photovoltaic power generation companies, photovoltaic power plants are numerous, generally located in remote areas, and occupy large areas (reaching thousands or even tens of thousands of acres), with a large number of devices (thousands to tens of thousands of photovoltaic panels). Currently, the difficulty in accurately locating large quantities of photovoltaic panels poses significant challenges to new energy power generation companies in project construction supervision, production operation and maintenance, and asset management, hindering the achievement of refined production, construction, and operation management. First, during the construction period, it is impossible to accurately grasp the actual number and progress of photovoltaic panels installed, making it impossible to determine whether there is any problem of "shortage" in the installation of the project; Secondly, after the photovoltaic system is put into operation, when problems need to be addressed, it is difficult for staff to find the corresponding photovoltaic panels in a timely manner, resulting in low work efficiency.

[0003] Furthermore, while a south-facing orientation of photovoltaic (PV) panels theoretically ensures maximum power generation during the midday sun's strongest hours, the orientation can vary depending on the terrain. Moreover, with the full implementation of market-based trading mechanisms for renewable energy, electricity prices are higher during morning / afternoon peak hours. Therefore, renewable energy power generation companies should prioritize generating more power in the morning / afternoon, rather than focusing on midday output. The difficulty in obtaining accurate orientation data for large quantities of PV panels makes it impossible to precisely analyze the peak power generation periods of PV farms, hindering renewable energy power generation companies from gaining a competitive edge in the electricity market. Summary of the Invention

[0004] This invention provides a method and apparatus for determining the location and orientation of photovoltaic panels based on satellite imagery, in order to solve the problem that it is difficult to obtain accurate positioning and orientation of a large number of photovoltaic panels in the prior art, which leads to the inability of new energy power generation enterprises to achieve refined production, construction and operation management and to take the initiative in the competition of the power market.

[0005] In a first aspect, the present invention provides a method for determining the position and orientation of a photovoltaic panel based on satellite imagery, comprising: Acquire satellite imagery of the target area, wherein the satellite imagery is application-grade orthorectified visible optical imagery; Based on the regional differences and image imaging differences of the satellite imagery, a recognition algorithm is selected. The satellite imagery is then processed based on the selected recognition algorithm to obtain all suspected photovoltaic panel images. The recognition algorithm includes a scanning contrast recognition algorithm, an edge detection recognition algorithm, and a color analysis recognition algorithm. Calculate the pixel coordinates of the two endpoints of the central axis of each suspected photovoltaic panel graphic; The central axis of each suspected photovoltaic panel pattern is plotted on the satellite image based on the pixel coordinates of the two endpoints of the central axis. Based on the central axis of all photovoltaic panels and all suspected photovoltaic panel graphics on the satellite image, the pixel coordinates of the two ends of the central axis of all photovoltaic panels are determined. The pixel coordinates of the center point of the central axis are determined based on the pixel coordinates of the two ends of the central axis of each photovoltaic panel; Convert the pixel coordinates of the two ends of the central axis of all photovoltaic panels and the pixel coordinates of the center point of the central axis into latitude and longitude coordinates; The latitude and longitude coordinates of the center point of the central axis of each photovoltaic panel are used as its position; The arrangement direction of each photovoltaic panel is determined based on the latitude and longitude coordinates of the two ends of its central axis, and the direction perpendicular to its arrangement direction is taken as its orientation.

[0006] Optionally, the algorithm for selecting and recognizing regional differences and image imaging differences based on the satellite imagery includes: When the satellite imagery has a bright field of view with strong illumination, a simple background, and the shadow of a photovoltaic panel, the contrast scanning recognition algorithm is selected. When there is no shadow of the photovoltaic panel in the satellite image and there is a contrast between the brightness of the photovoltaic panel's own reflection and the environmental background, the edge detection and recognition algorithm is selected. When there is no shadow of the photovoltaic panel in the satellite image and the color of the photovoltaic panel contrasts with the environmental background, the color analysis and recognition algorithm is selected.

[0007] Optionally, the scanning contrast recognition algorithm includes: The satellite image is processed into a grayscale image; The grayscale image is scanned by setting the scanning window and scanning step size, and the average brightness of each scanning window is calculated. For each column, determine whether the difference in average brightness between two adjacent scanning windows is greater than the first threshold. If so, take the center point of the two adjacent scanning windows as the key point. Feature point clustering is performed on all key points to obtain all clustering graphs; Geometric morphology screening is performed on all the clustered graphics to obtain all the suspected photovoltaic panel graphics.

[0008] Optionally, the step of performing geometric morphology screening on all clustered graphics to obtain all suspected photovoltaic panel graphics includes: Determine whether the aspect ratio and area of ​​the clustered graph meet the preset conditions; If so, then it will be used as the suspected photovoltaic panel graphic.

[0009] Optionally, the edge detection and recognition algorithm includes: The satellite image is processed into a grayscale image; The grayscale image is then subjected to Gaussian smoothing. Perform edge detection and edge closure on the processed grayscale image; Extract the edge contours to obtain all the suspected photovoltaic panel graphics.

[0010] Optionally, the color analysis and recognition algorithm includes: Determine the number of channels for each pixel in the satellite image and obtain the channel values ​​for each pixel; For each pixel, determine whether the values ​​of each channel are within their preset range and whether the ratio of the values ​​of any two channels is within their preset range. If so, then the pixel is designated as a key pixel. Cluster all key pixels by feature point clustering to obtain all clustering graphics; Geometric morphology screening is performed on all the clustered graphics to obtain all the suspected photovoltaic panel graphics.

[0011] Optionally, the step of verifying the central axis of all photovoltaic panels and all suspected photovoltaic panel graphics on the satellite imagery to determine the pixel coordinates of the two endpoints of the central axis of all photovoltaic panels includes: When a central axis coincides with a photovoltaic panel, the pixel coordinates of the two endpoints of the central axis are taken as the pixel coordinates of the two endpoints of the central axis of the photovoltaic panel. When a central axis coincides with multiple closely arranged photovoltaic panels, the central axis is divided equally according to the actual number of photovoltaic panels, and the pixel coordinates of the two ends of the multiple central axes after equal division are respectively used as the pixel coordinates of the two ends of the central axis of the multiple closely arranged photovoltaic panels. If a central axis falls outside all photovoltaic panels, that central axis will be deleted. When a photovoltaic panel does not have a coincident central axis, set the pixel coordinates of the two endpoints of the central axis of the photovoltaic panel.

[0012] Secondly, the present invention provides a device for determining the position and orientation of a photovoltaic panel based on satellite imagery, comprising an acquisition module, an identification module, a calculation module, a drawing module, a verification module, a positioning module, a conversion module, a positioning module, and an orientation module, wherein: The acquisition module is used to acquire satellite images of the target area, and the satellite images are application-grade orthorectified visible optical images. The identification module is used to select an identification algorithm based on the regional differences and image imaging differences of the satellite image, and to process the satellite image based on the selected identification algorithm to obtain all suspected photovoltaic panel images. The identification algorithm includes a scanning contrast identification algorithm, an edge detection identification algorithm, and a color analysis identification algorithm. The calculation module is used to calculate the pixel coordinates of the two ends of the central axis of each suspected photovoltaic panel graphic; The drawing module is used to draw the central axis of each suspected photovoltaic panel graphic on the satellite image based on the pixel coordinates of the two ends of the central axis. The verification module is used to verify the central axis of all photovoltaic panels and all suspected photovoltaic panel graphics on the satellite image, and to determine the pixel coordinates of the two ends of the central axis of all photovoltaic panels. The positioning module is used to determine the pixel coordinates of the center point of the central axis of each photovoltaic panel based on the pixel coordinates of the two ends of the central axis. The conversion module is used to convert the pixel coordinates of the two ends of the central axis of all photovoltaic panels and the pixel coordinates of the center point of the central axis into latitude and longitude coordinates. The positioning module is used to determine the latitude and longitude coordinates of the center point of the central axis of each photovoltaic panel as its position; The orientation module is used to determine the arrangement direction of each photovoltaic panel based on the latitude and longitude coordinates of the two ends of the central axis of each photovoltaic panel, and the direction perpendicular to its arrangement direction is taken as its orientation.

[0013] Thirdly, the present invention provides an electronic device, comprising: A processor; and a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform the steps of any of the methods described above.

[0014] Fourthly, the present invention provides a storage medium comprising: The storage medium stores a processing program for determining the position and orientation of photovoltaic panels based on satellite imagery. When the processing program for determining the position and orientation of photovoltaic panels based on satellite imagery is executed by a processor, it implements the steps of any of the methods described above.

[0015] The aforementioned solution identifies massive amounts of photovoltaic (PV) panel images based on satellite imagery of the target area and obtains the accurate location and orientation of individual PV panels. This lowers the barrier to image acquisition and analysis, helping new energy power generation companies establish a complete and accurate geographic information ledger for PV equipment. During the construction of PV projects, statistical analysis of the number and location distribution of PV panels over time helps monitor project progress. After PV power generation is put into operation, it enables navigation for individual PV devices, facilitating timely troubleshooting by staff and strongly supporting new energy power generation companies in achieving refined production, construction, and operation management. Furthermore, the orientation of PV panels can determine peak power generation periods, aiding in the accurate calculation of the expected peak output curve of the PV power field and enabling precise output assessment. This allows new energy power generation companies to better tailor their competitive bidding strategies, gaining a competitive edge in the electricity market. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a method for determining the position and orientation of a photovoltaic panel based on satellite imagery, as provided in Embodiment 1 of the present invention. Figure 2 This is a schematic diagram of a local grayscale image of a photovoltaic electric field provided in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the central axis and orientation of the photovoltaic panel provided in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the visualization result of local identification and rendering of a photovoltaic electric field provided in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of a device for determining the position and orientation of a photovoltaic panel based on satellite imagery, provided in Embodiment 2 of the present invention. Detailed Implementation

[0017] The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. The advantages and features of the present application will become clearer from the following description and claims. It should be noted that the drawings are all in a very simplified form and are not to scale, and are only used to facilitate and clarify the illustration of the embodiments of the present application.

[0018] It should be noted that, in order to clearly illustrate the content of this application, several embodiments are provided to further explain the different implementations of this application. These embodiments are enumerated rather than exhaustive. Furthermore, for the sake of brevity, content mentioned in the preceding embodiments is often omitted in the following embodiments. Therefore, content not mentioned in the following embodiments can be referred to in the preceding embodiments.

[0019] Example 1 Please refer to Figure 1 , Figure 1 The image shown illustrates a method for determining the location and orientation of a photovoltaic panel based on satellite imagery, as provided in this embodiment. The method includes: S1. Acquire satellite imagery of the target area. The satellite imagery is application-grade orthorectified visible optical imagery. It should be noted that the satellite imagery in this embodiment does not need to be hyperspectral satellite remote sensing imagery, but rather visible optical imagery provided by ordinary commercial satellite platforms, such as full-color imagery (PMS) or multispectral near-infrared imagery (MSI). This not only lowers the barrier to acquisition and analysis, but also provides the necessary imagery resources from multiple domestic and international commercial satellite platforms, offering a wide range of choices, low acquisition costs, and timely resource updates. Furthermore, commercial satellite platforms can now provide application-grade orthorectified imagery that has undergone target area fusion and radiometric / geometric / atmospheric correction, thus eliminating the need for additional preprocessing of satellite imagery.

[0020] In this embodiment, in order to meet the needs of target identification and monitoring, the satellite image is a satellite image of the target area when the cloud cover is 0 or the cloud cover does not obscure the target area.

[0021] In this embodiment, the resolution of the satellite image is less than or equal to 2 meters.

[0022] In one specific implementation, the satellite imagery is a 3-4 channel RGB / RGBA / BGRN image, in TIFF format, with a bit depth of 8 or 16 bits.

[0023] S2. Based on the regional differences and image imaging differences of satellite imagery, a selection and recognition algorithm is used to process satellite imagery and obtain all suspected photovoltaic panel images. The recognition algorithm includes scanning contrast recognition algorithm, edge detection recognition algorithm and color analysis recognition algorithm. Considering the regional differences in satellite imagery (such as different ground backgrounds, vegetation, and lighting conditions) and the differences in image imaging from different commercial satellite optical sensors that may be used for analysis, multiple recognition algorithms are used in combination for identification.

[0024] In this embodiment, the algorithm for selecting and recognizing regional differences and image imaging differences based on satellite imagery includes: When the satellite imagery has a bright field of view with strong illumination, a simple background, and the shadow of a photovoltaic panel, the contrast scanning recognition algorithm is selected. When there is no shadow of the photovoltaic panel in the satellite image and there is a contrast between the brightness of the photovoltaic panel's own reflection and the environmental background, the edge detection and recognition algorithm is selected. When there is no shadow of the photovoltaic panel in the satellite image and the color of the photovoltaic panel contrasts with the background, the color analysis and recognition algorithm is selected.

[0025] In one specific implementation, multiple recognition algorithms can be used simultaneously to recognize satellite images, and the best recognition result can be selected.

[0026] In this embodiment, the scanning contrast recognition algorithm includes: Process satellite images into grayscale images; Set the scanning window and scanning step size to scan the grayscale image, and calculate the average brightness of each scanning window; For each column, determine whether the difference in average brightness between two adjacent scanning windows is greater than the first threshold. If so, take the center point of the two adjacent scanning windows as the key point. Feature point clustering is performed on all key points to obtain all clustering graphs; Geometric morphology screening was performed on all clustered patterns to obtain all suspected photovoltaic panel patterns.

[0027] It's important to note that while a smaller scanning window generally results in higher accuracy, for large-area, high-resolution images, an excessively small window can lead to slow resolution. Therefore, the size of the scanning window must be considered in relation to the resolution. Generally, when the resolution is below 0.5 meters, a scanning window of 7×7 pixels or 9×9 pixels can be used; when the resolution is above 0.75 meters, a scanning window of 5×5 pixels or 3×3 pixels can be used.

[0028] In one specific implementation, the scanning window is set to 5×5 pixels, the scanning step size is set to 5 pixels, and the first threshold is set to 20.

[0029] In one specific implementation, the feature points are clustered using DBSCAN clustering, with a neighborhood radius of 10 and a minimum number of samples of 3.

[0030] It should be noted that since photovoltaic panels are rectangular, clusters that meet the geometric shape and size criteria are identified as suspected photovoltaic panel patterns. Other clusters that do not meet the criteria in terms of geometric shape and size, such as roads, vegetation and land boundaries, water edges, and general buildings, will be filtered out and not identified.

[0031] In this embodiment, geometric morphology screening is performed on all clustered patterns to obtain all suspected photovoltaic panel patterns, including: Determine whether the aspect ratio and area of ​​the clustered graph meet the preset conditions; If so, then it will be considered a suspected photovoltaic panel graphic.

[0032] In one specific implementation, geometric morphology filtering of all clustered graphs is performed using the OpenCV algorithm library.

[0033] In one specific implementation, the preset conditions are that the aspect ratio is greater than 3 and the area is greater than 100.

[0034] The scanning contrast recognition algorithm actually detects the brightness contrast between the photovoltaic panel's shadow and the ambient background, and identifies the photovoltaic panel's shadow as an equivalent substitute for the photovoltaic panel. Figure 2 The image shown is a partial grayscale image of a photovoltaic field. The dark shadows of the photovoltaic panels can be clearly seen, and the red dot is the center point of the central axis of the photovoltaic panels.

[0035] In this embodiment, the edge detection and recognition algorithm includes: Process satellite images into grayscale images; Apply Gaussian smoothing to the grayscale image; Perform edge detection and edge closure on the processed grayscale image; Extract the edge contours to obtain all possible photovoltaic panel graphics.

[0036] Gaussian smoothing is used to suppress image noise; in one specific implementation, the Gaussian kernel is set to 7.

[0037] In one specific implementation, edge detection is achieved using the OpenCV-Canny method, with upper and lower limits set to 50 and 150, respectively.

[0038] In one specific implementation, edge closure is achieved through morphological operations, with a kernel size set to 3×3.

[0039] In this embodiment, the color analysis and recognition algorithm includes: Determine the number of channels for each pixel in the satellite image and obtain the channel values ​​for each pixel; For each pixel, determine whether the values ​​of each channel are within their preset range and whether the ratio of the values ​​of any two channels is within their preset range. If so, then the pixel is designated as a key pixel. Cluster all key pixels by feature point clustering to obtain all clustering graphics; Geometric morphology screening was performed on all clustered patterns to obtain all suspected photovoltaic panel patterns.

[0040] By setting preset ranges for the values ​​of each channel and the ratios between pairs of channel values, the color difference between the photovoltaic panel and the environmental background can be determined. It should be noted that due to the combined effects of differences in imaging from different satellite sensors, atmospheric conditions at the time of imaging, and geographical environment, remote sensing images can vary greatly. Therefore, the preset range is not fixed and must be set according to the actual situation of the image.

[0041] In this embodiment, the preset range of each channel value and the preset range of the ratio of each pair of channel values ​​are determined by interval statistics of each channel value of the satellite image.

[0042] It should be noted that for 8-bit satellite imagery, the preset range of values ​​for each channel is determined between 0 and 255, and for 16-bit satellite imagery, the preset range of values ​​for each channel is determined between 0 and 65535. Alternatively, the 16-bit satellite imagery can be converted to 8-bit satellite imagery before determining the preset range.

[0043] In one specific implementation, taking an RGB image as an example, each pixel has 3 channels. The preset range of channel 1 values ​​is [80, 255], the preset range of channel 2 values ​​is [80, 255], the preset range of channel 3 values ​​is [110, 255], the preset range of the ratio of channel 1 to channel 2 values ​​is [0.89, 1.2], the preset range of the ratio of channel 1 to channel 3 values ​​is less than 0.8, and the preset range of the ratio of channel 2 to channel 3 values ​​is less than 0.8.

[0044] In one specific implementation, the feature points are clustered using DBSCAN clustering, with a neighborhood radius of 5 and a minimum number of samples of 4.

[0045] It should be noted that the techniques used for geometric shape selection in the color analysis and recognition algorithm are the same as those used in the scanning contrast recognition algorithm.

[0046] S3. Calculate the pixel coordinates of the two endpoints of the central axis of each suspected photovoltaic panel graphic; Figure 3 The diagram shows the central axis and orientation of the photovoltaic panel, where the central axis is the centerline along the major axis. Figure 3 In the diagram, line AB, the central axis actually represents the arrangement direction of the photovoltaic panels themselves. The orientation of the photovoltaic panels is perpendicular to their arrangement direction, i.e. Figure 3 The direction of the dashed line.

[0047] In one specific implementation, the pixel coordinates of the two ends of the central axis of all suspected photovoltaic panel graphics are saved as a CSV file.

[0048] S4. Draw the central axis of each suspected photovoltaic panel graphic on the satellite image based on the pixel coordinates of the two ends of the central axis. S5. Verify the central axis of all photovoltaic panels and all suspected photovoltaic panel graphics on the satellite imagery to determine the pixel coordinates of the two ends of the central axis of all photovoltaic panels; In this embodiment, the pixel coordinates of the two endpoints of the central axis of all photovoltaic panels and all suspected photovoltaic panel graphics on the satellite image are checked to determine the pixel coordinates of the two endpoints of the central axis of all photovoltaic panels, including: When a central axis coincides with a photovoltaic panel, the pixel coordinates of the two endpoints of the central axis are taken as the pixel coordinates of the two endpoints of the central axis of the photovoltaic panel. When a central axis coincides with multiple closely arranged photovoltaic panels, the central axis is divided equally according to the actual number of photovoltaic panels, and the pixel coordinates of the two ends of the multiple central axes after equal division are respectively used as the pixel coordinates of the two ends of the central axes of the multiple closely arranged photovoltaic panels. If a central axis falls outside all photovoltaic panels, that central axis will be deleted. When a photovoltaic panel does not have a coincident central axis, set the pixel coordinates of the two endpoints of the central axis of the photovoltaic panel.

[0049] Due to differences in satellite sensors, atmospheric conditions, geographical environment, etc., the quality of remote sensing images obtained by users varies greatly. At present, there is no way to guarantee that tens of thousands of photovoltaic panels can be identified in images of a large area without any errors or omissions. In order to ensure that the accurate location and orientation of photovoltaic panels are obtained, the identification results need to be manually checked and confirmed.

[0050] In one specific implementation, S4 and S5 can be implemented by the user using a visualization tool. Based on this tool, the user can load satellite imagery as a base map, read a CSV file containing the pixel coordinates of the two endpoints of the central axis of all suspected photovoltaic panel graphics, draw the central axis based on the pixel coordinates, and overlay it onto the base map for verification from an overall or locally zoomed perspective. The verification includes: When a central axis coincides with a photovoltaic panel, it is considered an accurate and effective identification, and no further action is taken. When a central axis coincides with multiple closely arranged photovoltaic panels, it is determined that the multiple closely arranged photovoltaic panels are identified as a whole strip. The central axis can be selected and divided equally according to the actual number of photovoltaic panels shown in the satellite image. When a central axis falls outside all photovoltaic panels, it is judged as an identification error, and can be deleted by selecting a large area or by selecting each panel individually. If a photovoltaic panel does not have a coincident central axis, it is considered an omission in the identification process. The pixel coordinates can be automatically calculated and set by clicking on the two ends of the central axis of the photovoltaic panel with the mouse.

[0051] Figure 4 The image shows the visualization result of a localized photovoltaic field identification rendering. It can be observed that in addition to the many photovoltaic panels in the photovoltaic area being identified (rendered in yellow), there are also a small number of rendering marks outside the photovoltaic area that are not photovoltaic panels. These can be deleted by selecting a large area with a visualization tool or by clicking on them one by one.

[0052] S6. Determine the pixel coordinates of the center point of the central axis of each photovoltaic panel based on the pixel coordinates of the two ends of the central axis; S7. Convert the pixel coordinates of the two ends of the central axis of all photovoltaic panels and the pixel coordinates of the center point of the central axis into latitude and longitude coordinates; S8. Use the latitude and longitude coordinates of the center point of the central axis of each photovoltaic panel as its position; S9. Determine the arrangement direction of each photovoltaic panel based on the latitude and longitude coordinates of the two ends of the central axis, and take the direction perpendicular to its arrangement direction as its orientation.

[0053] The aforementioned solution identifies massive amounts of photovoltaic (PV) panel images based on satellite imagery of the target area and obtains the accurate location and orientation of individual PV panels. This lowers the barrier to image acquisition and analysis, helping new energy power generation companies establish a complete and accurate geographic information ledger for PV equipment. During the construction of PV projects, statistical analysis of the number and location distribution of PV panels over time helps monitor project progress. After PV power generation is put into operation, it enables navigation for individual PV devices, facilitating timely troubleshooting by staff and strongly supporting new energy power generation companies in achieving refined production, construction, and operation management. Furthermore, the orientation of PV panels can determine peak power generation periods, aiding in the accurate calculation of the expected peak output curve of the PV power field and enabling precise output assessment. This allows new energy power generation companies to better tailor their competitive bidding strategies, gaining a competitive edge in the electricity market.

[0054] Example 2 Please refer to Figure 5 , Figure 5 The image shown is a device for determining the location and orientation of a photovoltaic panel based on satellite imagery, provided in this embodiment. It includes an acquisition module, an identification module, a calculation module, a drawing module, a verification module, a positioning module, a conversion module, a positioning module, and an orientation module, wherein: The acquisition module is used to acquire satellite imagery of the target area. The satellite imagery is application-grade orthorectified visible optical imagery. The recognition module is used to select a recognition algorithm based on regional differences and image imaging differences in satellite imagery. Based on the selected recognition algorithm, the satellite imagery is processed to obtain all suspected photovoltaic panel images. The recognition algorithms include scanning contrast recognition algorithm, edge detection recognition algorithm, and color analysis recognition algorithm. The calculation module is used to calculate the pixel coordinates of the two ends of the central axis of each suspected photovoltaic panel graphic; The drawing module is used to draw the central axis of each suspected photovoltaic panel graphic on satellite imagery based on the pixel coordinates of the two ends of the central axis. The verification module is used to verify the central axis of all photovoltaic panels and all suspected photovoltaic panel graphics on satellite imagery, and to determine the pixel coordinates of the two ends of the central axis of all photovoltaic panels; The positioning module is used to determine the pixel coordinates of the center point of the central axis of each photovoltaic panel based on the pixel coordinates of the two ends of the central axis. The conversion module is used to convert the pixel coordinates of the two ends of the central axis of all photovoltaic panels and the pixel coordinates of the center point of the central axis into latitude and longitude coordinates; The positioning module is used to determine the latitude and longitude coordinates of the center point of the central axis of each photovoltaic panel as its location; The orientation module is used to determine the arrangement direction of each photovoltaic panel based on the latitude and longitude coordinates of the two ends of the central axis of each panel, and the direction perpendicular to its arrangement direction is taken as its orientation.

[0055] Example 3 This embodiment provides a device for determining the location and orientation of photovoltaic panels based on satellite imagery. Specifically, the device may include: A processor; and a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform the steps as described in the above embodiments.

[0056] Example 4 This embodiment provides a storage medium for determining the location and orientation of photovoltaic panels based on satellite imagery. Specifically, the storage medium may include: The storage medium stores a processing program for determining the position and orientation of photovoltaic panels based on satellite imagery. When the processing program for determining the position and orientation of photovoltaic panels based on satellite imagery is executed by the processor, it implements the steps as described in the above embodiments.

[0057] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for determining the position and orientation of photovoltaic panels based on satellite imagery, characterized in that, include: Acquire satellite imagery of the target area, wherein the satellite imagery is application-grade orthorectified visible optical imagery; Based on the regional differences and image imaging differences of the satellite imagery, a recognition algorithm is selected. The satellite imagery is then processed based on the selected recognition algorithm to obtain all suspected photovoltaic panel images. The recognition algorithm includes a scanning contrast recognition algorithm, an edge detection recognition algorithm, and a color analysis recognition algorithm. Calculate the pixel coordinates of the two endpoints of the central axis of each suspected photovoltaic panel graphic; The central axis of each suspected photovoltaic panel pattern is plotted on the satellite image based on the pixel coordinates of the two endpoints of the central axis. Based on the central axis of all photovoltaic panels and all suspected photovoltaic panel graphics on the satellite image, the pixel coordinates of the two ends of the central axis of all photovoltaic panels are determined. The pixel coordinates of the center point of the central axis are determined based on the pixel coordinates of the two ends of the central axis of each photovoltaic panel; Convert the pixel coordinates of the two ends of the central axis of all photovoltaic panels and the pixel coordinates of the center point of the central axis into latitude and longitude coordinates; The latitude and longitude coordinates of the center point of the central axis of each photovoltaic panel are used as its position; The arrangement direction of each photovoltaic panel is determined based on the latitude and longitude coordinates of the two ends of its central axis, and the direction perpendicular to its arrangement direction is taken as its orientation.

2. The method according to claim 1, characterized in that, The algorithm for selecting and recognizing regional differences and image imaging differences based on the satellite imagery includes: When the satellite imagery has a bright field of view with strong illumination, a simple background, and the shadow of a photovoltaic panel, the contrast scanning recognition algorithm is selected. When there is no shadow of the photovoltaic panel in the satellite image and there is a contrast between the brightness of the photovoltaic panel's own reflection and the environmental background, the edge detection and recognition algorithm is selected. When there is no shadow of the photovoltaic panel in the satellite image and the color of the photovoltaic panel contrasts with the environmental background, the color analysis and recognition algorithm is selected.

3. The method according to claim 1, characterized in that, The scanning contrast recognition algorithm includes: The satellite image is processed into a grayscale image; The grayscale image is scanned by setting the scanning window and scanning step size, and the average brightness of each scanning window is calculated. For each column, determine whether the difference in average brightness between two adjacent scanning windows is greater than the first threshold. If so, take the center point of the two adjacent scanning windows as the key point. Feature point clustering is performed on all key points to obtain all clustering graphs; Geometric morphology screening is performed on all the clustered graphics to obtain all the suspected photovoltaic panel graphics.

4. The method according to claim 3, characterized in that, The geometric morphology screening of all clustered graphics to obtain all suspected photovoltaic panel graphics includes: Determine whether the aspect ratio and area of ​​the clustered graph meet the preset conditions; If so, then it will be used as the suspected photovoltaic panel graphic.

5. The method according to claim 1, characterized in that, The edge detection and recognition algorithm includes: The satellite image is processed into a grayscale image; The grayscale image is then subjected to Gaussian smoothing. Perform edge detection and edge closure on the processed grayscale image; Extract the edge contours to obtain all the suspected photovoltaic panel graphics.

6. The method according to claim 1, characterized in that, The color analysis and recognition algorithm includes: Determine the number of channels for each pixel in the satellite image and obtain the channel values ​​for each pixel; For each pixel, determine whether the values ​​of each channel are within their preset range and whether the ratio of the values ​​of any two channels is within their preset range. If so, then the pixel is designated as a key pixel. Cluster all key pixels by feature point clustering to obtain all clustering graphics; Geometric morphology screening is performed on all the clustered graphics to obtain all the suspected photovoltaic panel graphics.

7. The method according to claim 1, characterized in that, The process of verifying the central axis of all photovoltaic panels and all suspected photovoltaic panel images on the satellite imagery to determine the pixel coordinates of the two endpoints of the central axis of all photovoltaic panels includes: When a central axis coincides with a photovoltaic panel, the pixel coordinates of the two endpoints of the central axis are taken as the pixel coordinates of the two endpoints of the central axis of the photovoltaic panel. When a central axis coincides with multiple closely arranged photovoltaic panels, the central axis is divided equally according to the actual number of photovoltaic panels, and the pixel coordinates of the two ends of the multiple central axes after equal division are respectively used as the pixel coordinates of the two ends of the central axis of the multiple closely arranged photovoltaic panels. If a central axis falls outside all photovoltaic panels, that central axis will be deleted. When a photovoltaic panel does not have a coincident central axis, set the pixel coordinates of the two endpoints of the central axis of the photovoltaic panel.

8. A device for determining the position and orientation of a photovoltaic panel based on satellite imagery, characterized in that, It includes modules for acquisition, recognition, calculation, drawing, verification, positioning, conversion, location, and orientation, among which: The acquisition module is used to acquire satellite images of the target area, and the satellite images are application-grade orthorectified visible optical images. The identification module is used to select an identification algorithm based on the regional differences and image imaging differences of the satellite image, and to process the satellite image based on the selected identification algorithm to obtain all suspected photovoltaic panel images. The identification algorithm includes a scanning contrast identification algorithm, an edge detection identification algorithm, and a color analysis identification algorithm. The calculation module is used to calculate the pixel coordinates of the two ends of the central axis of each suspected photovoltaic panel graphic; The drawing module is used to draw the central axis of each suspected photovoltaic panel graphic on the satellite image based on the pixel coordinates of the two ends of the central axis. The verification module is used to verify the central axis of all photovoltaic panels and all suspected photovoltaic panel graphics on the satellite image, and to determine the pixel coordinates of the two ends of the central axis of all photovoltaic panels. The positioning module is used to determine the pixel coordinates of the center point of the central axis of each photovoltaic panel based on the pixel coordinates of the two ends of the central axis. The conversion module is used to convert the pixel coordinates of the two ends of the central axis of all photovoltaic panels and the pixel coordinates of the center point of the central axis into latitude and longitude coordinates. The positioning module is used to determine the latitude and longitude coordinates of the center point of the central axis of each photovoltaic panel as its position; The orientation module is used to determine the arrangement direction of each photovoltaic panel based on the latitude and longitude coordinates of the two ends of the central axis of each photovoltaic panel, and the direction perpendicular to its arrangement direction is taken as its orientation.

9. An electronic device, characterized in that, include: processor; And a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform the steps of the method as described in any one of claims 1-7.

10. A storage medium, characterized in that, include: The storage medium stores a processing program for determining the position and orientation of photovoltaic panels based on satellite imagery, which, when executed by a processor, implements the steps of the method as described in any one of claims 1-7.

Images