A power transmission and distribution line remote sensing image enhancement system based on digital twinning

Abstract

This invention belongs to the field of image processing technology and provides a remote sensing image enhancement system for power transmission and transformation lines based on digital twins. The system includes: a remote sensing image acquisition module for acquiring original remote sensing images of the power transmission and transformation lines; a digital twin model construction module for constructing a virtual digital twin model of the power transmission and transformation lines based on the original remote sensing images using digital twin technology; and an image enhancement processing module for enhancing the original remote sensing images based on the virtual digital twin model, and performing verification and optimization to obtain enhanced remote sensing images. This invention improves the image processing effect and image quality of remote sensing images by utilizing digital twin technology to enhance the original remote sensing images of power transmission and transformation lines, providing a foundation for the efficient utilization of remote sensing images.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a remote sensing image enhancement system for power transmission and transformation lines based on digital twins. Background Technology

[0002] Power transmission and transformation projects involve numerous locations, long distances, and complex terrain. The main construction processes include building demolition, tree felling, road construction, land leveling, foundation excavation, material transportation, foundation pouring, tower erection, conductor and ground wire deployment, and trial operation. Due to the long distances and time spans of these projects, they cause numerous surface disturbances, are prone to environmental and water conservation issues, and attract significant public attention.

[0003] Traditional environmental protection supervision relies on manual methods such as inspections and field investigations, making it difficult to conduct timely and comprehensive environmental and water conservation monitoring. However, the development of high-resolution remote sensing technology in recent years has made it possible to conduct high-frequency, rapid, and efficient comprehensive environmental and water conservation monitoring of the power grid, providing strong data support for the development of intelligent identification technology for environmental and water conservation targets.

[0004] In the process of processing remote sensing images, there are problems such as relatively simple processing techniques, low processing efficiency, high processing difficulty for some remote sensing images, and unsatisfactory processing results, which result in low utilization of remote sensing images.

[0005] Therefore, it is necessary to provide a remote sensing image enhancement system for power transmission and transformation lines based on digital twins. Summary of the Invention

[0006] This invention provides a digital twin-based remote sensing image enhancement system for power transmission and transformation lines. By utilizing digital twin technology to enhance the original remote sensing images of power transmission and transformation lines, the system can improve the image processing effect and image quality of remote sensing images, thus providing a foundation for the efficient utilization of remote sensing images.

[0007] This invention provides a digital twin-based remote sensing image enhancement system for power transmission and transformation lines, comprising:

[0008] The remote sensing image acquisition module is used to acquire raw remote sensing images of power transmission and transformation lines.

[0009] The digital twin model building module is used to construct a virtual digital twin model of a power transmission and transformation line based on the original remote sensing imagery using digital twin technology.

[0010] The image enhancement processing module is used to enhance the original remote sensing images based on the virtual digital twin model, and to verify and optimize them to obtain enhanced remote sensing images.

[0011] Furthermore, original remote sensing images of transmission and transformation lines are acquired, including:

[0012] Configure the image acquisition equipment for accessing high-resolution satellites and set the operating parameters of the image acquisition equipment;

[0013] Based on the operating parameters, image acquisition equipment is used to receive and acquire raw remote sensing images of power transmission and transformation lines via high-resolution satellites.

[0014] Furthermore, the original remote sensing images include engineering project images of power transmission and transformation lines and images of the surrounding environment; engineering project images include, but are not limited to, images of the specific route trajectory, deployment scale, and construction site location of the engineering project; and images of the surrounding environment include, but are not limited to, images of buildings, vegetation cover, and road topography.

[0015] Furthermore, the digital twin model building module includes an image processing unit and a model building unit;

[0016] The image processing unit is used to perform grayscale preprocessing on the original remote sensing images and to extract image features to obtain target image features for building the model.

[0017] The model building unit is used to construct a digital twin model based on the features of the target image, relying on the Unity3D platform, and based on the set model algorithm, model structure, and model parameters.

[0018] Furthermore, the image enhancement processing module includes a processing unit and a verification and optimization unit;

[0019] The processing unit is used to enhance the original remote sensing images based on a digital twin model using virtual reality technology to obtain enhanced remote sensing images;

[0020] The verification and optimization unit is used to simulate and analyze the image enhancement process, and to verify and optimize the image enhancement effect.

[0021] Furthermore, using virtual reality technology and based on a digital twin model, the original remote sensing images are enhanced to obtain enhanced remote sensing images, including:

[0022] Based on the digital twin model, a three-dimensional convolutional neural network algorithm is used to extract three-dimensional image features from the original remote sensing images to obtain three-dimensional image feature data.

[0023] Based on the feature data of three-dimensional images, three-dimensional modeling is performed using virtual reality technology to obtain a three-dimensional auxiliary model;

[0024] Based on a 3D auxiliary model, GAN image enhancement technology is used to enhance the original remote sensing images to obtain enhanced remote sensing images.

[0025] Furthermore, the image enhancement process is simulated and analyzed to verify and optimize the image enhancement effect, including:

[0026] Construct a standard image database that meets image enhancement standards; the standard image database stores several standard images;

[0027] Using a set simulation model, the process of enhancing the original remote sensing image is simulated and analyzed to obtain the simulation analysis results and the simulated enhanced image.

[0028] The resolution of the simulated enhanced image is compared and analyzed with that of the standard image. Based on the results of the comparison analysis, the simulated enhanced image is verified. If the resolution comparison value is less than the set resolution comparison threshold, the verification result of the simulated enhanced image is confirmed. If the resolution comparison value is greater than the set resolution comparison threshold, the resolution of the original remote sensing image corresponding to the simulated enhanced image is further optimized.

[0029] Furthermore, the resolution of the original remote sensing image corresponding to the simulated enhanced image is further optimized, including:

[0030] Acquire the original remote sensing image, and cluster the distribution of pixels in the original remote sensing image to obtain several clustering regions;

[0031] For clustered regions where the number of pixels exceeds a set threshold, pixel interpolation is used to enhance the resolution of the original remote sensing image and obtain the first enhanced remote sensing image.

[0032] For clustered regions where the number of pixels is less than a set threshold, the missing pixels are predicted based on a neural network prediction model to obtain the prediction results. Based on the prediction results, the missing pixels are averaged, removed, or supplemented to obtain the second enhanced remote sensing image.

[0033] By combining the pixel density of clustered regions and the local texture complexity, the weights of pixel interpolation and neural network prediction are set and dynamically adjusted. Based on the comprehensive weight composed of the pixel interpolation weights and the neural network prediction weights, the first and second enhanced remote sensing images are optimized and enhanced to obtain the enhanced first and second enhanced remote sensing images. The formula for calculating the comprehensive weight is:

[0034]

[0035] W i Represents the overall weight, α represents the adjustment factor, and N represents the overall weight. i N represents the number of pixels in the i-th cluster region; t C represents the total number of pixels. i Let C be the texture complexity of the i-th cluster region, and let C be the local window grayscale variance. max This represents the maximum texture complexity value; when W iWhen the weight exceeds the set weight threshold, interpolation pixel enhancement is prioritized to optimize and enhance the first enhanced remote sensing image; when W i When the weight threshold is less than or equal to the set weight threshold, a neural network is used to predict and supplement pixels to optimize and enhance the second enhanced remote sensing image.

[0036] By combining the enhanced first and second enhanced remote sensing images, a new remote sensing image with further optimized resolution is obtained.

[0037] Furthermore, it also includes the management and updating of the standard image database, specifically including:

[0038] Based on the cloud platform, the standard image database is set as a cloud database;

[0039] The standard image database is managed using management programs on the cloud platform.

[0040] By utilizing big data from geographic information systems, the first updated data is periodically retrieved;

[0041] Image data acquired by taking pictures with a camera mounted on a drone is used as the second update data;

[0042] Based on the work records at the construction site of the power transmission and transformation line, close-range ground image data is obtained as the third update data;

[0043] The first, second, and third update data are used as update data for the standard image database, and the standard image database is updated accordingly.

[0044] Furthermore, it also includes a result output and verification module for verifying the enhanced remote sensing imagery in practical applications, specifically including:

[0045] The verification of enhanced remote sensing image texture features is carried out as follows: using a set extraction template, the texture features of the enhanced remote sensing image are extracted and texture feature description information is obtained; the texture feature description information is compared and analyzed with the data in the set texture description information database to obtain the comparison results; and the first verification result is obtained based on the comparison results.

[0046] The identification of specific targets in enhanced remote sensing imagery is carried out as follows: using a target identification model, targets are identified in the enhanced remote sensing imagery to obtain target identification results; based on the accuracy of the target identification results, a second verification result is obtained.

[0047] Based on the first and second verification results, the verification results for the enhanced remote sensing image are obtained.

[0048] Compared with existing technologies, this invention has the following advantages and beneficial effects: it can achieve efficient and accurate enhancement processing of remote sensing images of power transmission and transformation lines; firstly, through the remote sensing image acquisition module, high-quality, high-resolution original remote sensing images can be acquired, providing a reliable data foundation for subsequent processing; the digital twin model construction module utilizes advanced digital twin technology to construct a virtual digital twin model that is highly consistent with the actual situation, providing strong support for image enhancement processing; the image enhancement processing module effectively improves the clarity and detail of the images; the function of managing and updating the standard image database can dynamically update and optimize the standard image database according to actual conditions and needs, thereby improving the adaptability and practicality of the system; the result output and verification module can perform texture feature verification and target recognition verification on the enhanced remote sensing images, further ensuring the quality and reliability of the images.

[0049] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0050] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0051] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0052] Figure 1 A schematic diagram of the structure of a digital twin-based remote sensing image enhancement system for power transmission and transformation lines;

[0053] Figure 2 A schematic diagram of the modular structure for building a digital twin model;

[0054] Figure 3 This is a schematic diagram of the image enhancement processing module. Detailed Implementation

[0055] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0056] This invention provides a remote sensing image enhancement system for power transmission and transformation lines based on digital twins, such as... Figure 1 As shown, it includes:

[0057] The remote sensing image acquisition module is used to acquire raw remote sensing images of power transmission and transformation lines.

[0058] The digital twin model building module is used to construct a virtual digital twin model of a power transmission and transformation line based on the original remote sensing imagery using digital twin technology.

[0059] The image enhancement processing module is used to enhance the original remote sensing images based on the virtual digital twin model, and to verify and optimize them to obtain enhanced remote sensing images.

[0060] The working principle of the above technical solution is as follows: To realize the remote sensing image enhancement system for power transmission and transformation lines based on digital twins, firstly, the remote sensing image acquisition module uses high-precision remote sensing equipment, such as high-definition cameras mounted on satellites or drones, to capture images of the power transmission and transformation lines, obtaining raw remote sensing images. These images contain detailed information about the power transmission and transformation lines and their surrounding environment, but the image quality may be low due to factors such as weather and lighting, making them difficult to use directly for analysis and judgment. Next, the digital twin model construction module uses digital twin technology to construct a virtual digital twin model of the power transmission and transformation lines based on these raw remote sensing images. This model not only includes the physical structure information of the power transmission and transformation lines but also simulates their operating status and changes in the surrounding environment. This virtual digital twin model provides the foundation for subsequent image enhancement processing. Then, the image enhancement module, based on this virtual digital twin model, enhances the original remote sensing image. By comparing the differences between the virtual model and the actual image, the module can identify problems such as blurriness and noise in the image and use advanced image processing algorithms to correct and optimize them. The processed remote sensing image not only has improved clarity but also better reflects the actual condition of the power transmission and transformation lines. Finally, the enhanced remote sensing image is verified and optimized to ensure image quality and accuracy. This step typically includes manual inspection of the image and comparison with field survey data to ensure that the final enhanced remote sensing image meets the needs of practical applications.

[0061] The beneficial effects of the above technical solution are as follows: The solution provided in this embodiment can significantly improve the quality and usability of remote sensing images of power transmission and transformation lines; through the introduction of digital twin technology, the system can construct a virtual model that is highly consistent with the actual situation, providing a precise foundation for image enhancement; at the same time, the application of advanced image processing algorithms effectively solves problems such as blurring and noise in the images, improving image clarity and accuracy; furthermore, by verifying and optimizing the enhanced remote sensing images, the quality and accuracy of the images are further ensured, enabling them to better serve the monitoring, maintenance, and management of power transmission and transformation lines; this not only improves work efficiency but also reduces the cost and risk of manual inspection, which is of great significance for improving the safety and stability of the power system.

[0062] In one embodiment, acquiring raw remote sensing images of transmission and transformation lines includes:

[0063] Configure the image acquisition equipment for accessing high-resolution satellites and set the operating parameters of the image acquisition equipment;

[0064] Based on the operating parameters, image acquisition equipment is used to receive and acquire raw remote sensing images of power transmission and transformation lines via high-resolution satellites.

[0065] The working principle of the above technical solution is as follows: by configuring the image acquisition equipment and setting its working parameters, such as satellite orbit, acquisition time, and resolution, it can be ensured that the acquired raw remote sensing images have the required quality and accuracy; subsequently, the image acquisition equipment uses the observation capabilities of high-resolution satellites to receive and acquire raw remote sensing images of power transmission and transformation lines. These raw image data contain detailed information about the power transmission and transformation lines and their surrounding environment, providing a foundation for subsequent image enhancement processing.

[0066] The beneficial effects of the above technical solution are as follows: by using the solution provided in this embodiment, and by utilizing image acquisition equipment connected to high-resolution satellites, it is possible to ensure the acquisition of accurate original remote sensing images of power transmission and transformation lines.

[0067] In one embodiment, the original remote sensing imagery includes engineering project images of the power transmission and transformation line and images of the surrounding environment; the engineering project images include, but are not limited to, images of the specific route trajectory, deployment scale, and construction site location of the engineering project; the images of the surrounding environment include, but are not limited to, images of buildings, vegetation cover, and road topography.

[0068] The working principle of the above technical solution is as follows: the original remote sensing images serve as the basic data for the construction of the digital twin model, and their completeness is crucial for subsequent digital twin modeling and image enhancement processing; the engineering project images record in detail the actual layout of the transmission and transformation lines, including the line trajectory, deployment scale, and specific location of the construction site. This information is crucial for understanding the physical characteristics and spatial distribution of the transmission and transformation lines, while the environmental images along the line capture key environmental factors around the line, such as the distribution of buildings, vegetation cover, and road terrain features. This environmental information helps to analyze the natural and human factors that may affect the operation of the line.

[0069] The beneficial effects of the above technical solution are as follows: by using the solution provided in this embodiment, the target for image acquisition is provided through the definition of the original remote sensing image.

[0070] In one embodiment, such as Figure 2 As shown, the digital twin model building module includes an image processing unit and a model building unit;

[0071] The image processing unit is used to perform grayscale preprocessing on the original remote sensing images and to extract image features to obtain target image features for building the model.

[0072] The model building unit is used to construct a digital twin model based on the features of the target image, relying on the Unity3D platform, and based on the set model algorithm, model structure, and model parameters.

[0073] The working principle of the above technical solution is as follows: The image processing unit first performs grayscale processing on the acquired raw remote sensing images. This step helps to reduce color information interference in the images and highlight the structural and textural features of the images. Subsequently, the image processing unit uses advanced image feature extraction algorithms to accurately identify and extract key image features related to the transmission and transformation lines. These features constitute the basic data for building a digital twin model. The model building unit then takes these target image features and relies on the powerful Unity3D platform, which is known for its high flexibility and powerful 3D rendering capabilities. Based on the preset model algorithm, the refined model structure, and the optimized model parameters, the model building unit gradually builds a digital twin model that is highly consistent with the real transmission and transformation lines. In this process, the selection of algorithms, the setting of structures, and the adjustment of parameters have been repeatedly verified and optimized to ensure that the final digital twin model can not only realistically reflect the actual condition of the transmission and transformation lines, but also meet the needs of subsequent image enhancement and analysis.

[0074] The beneficial effects of the above technical solution are as follows: by using the solution provided in this embodiment, grayscale preprocessing of the original remote sensing image and construction of a digital twin model provide a foundation for subsequent enhancement processing of the original remote sensing image.

[0075] In one embodiment, such as Figure 3 As shown, the image enhancement processing module includes a processing unit and a verification and optimization unit;

[0076] The processing unit is used to enhance the original remote sensing images based on a digital twin model using virtual reality technology to obtain enhanced remote sensing images;

[0077] In enhancing raw remote sensing images, various noises often exist, affecting image quality. This invention measures noise intensity by calculating the standard deviation of noise in local image regions. For each pixel, a window of size n×n (e.g., n=3) is defined centered on it, and the standard deviation of pixel values ​​within the window is calculated:

[0078]

[0079] K local x represents the standard deviation of noise. ijRepresents the pixel value in the i-th row and j-th column within the window, reflecting the pixel information at each specific location within the window. average This represents the average pixel value within the window; n represents the number of windows; n 2 This represents the set window size centered on the current pixel;

[0080] Based on the calculated noise standard deviation, the parameters of the Gaussian filter are adaptively adjusted. When the noise standard deviation is large, the standard deviation of the Gaussian filter is increased to enhance the noise suppression effect; when the noise standard deviation is small, the standard deviation of the Gaussian filter is decreased to avoid over-smoothing of image details. The response formula of the Gaussian filter is:

[0081]

[0082] In the above formula, G(x, y) represents the weight matrix, and (x, y) represents the position coordinates of the Gaussian filter on the image plane, used to determine the effect of the filter on pixels at different positions in the image; σ represents the standard deviation of the Gaussian filter, which determines the smoothness and filtering effect of the Gaussian filter; the larger the value of σ, the flatter the curve of the Gaussian function, the wider the range of the filter, the stronger the smoothing effect on the image, and the more neighboring pixels will participate in the filtering calculation; the smaller the value of σ, the steeper the curve, the more concentrated the range of the filter is near the center pixel, the relatively weaker the smoothing effect on the image, and the better it can preserve the details of the image; the response formula of the Gaussian filter is used to generate the weight matrix G(x, y), which determines the weight of each pixel in the filtering process. When filtering remote sensing images, the weight matrix of the filter is weighted and summed with the pixel values ​​in the window to obtain the filtered pixel values. For pixels closer to the center pixel, their position coordinates in the Gaussian function (x, y) are relatively smaller. 2 +y 2 A smaller value corresponds to a larger weight matrix G(x,y), which in turn contributes more to the filtering result; while pixels farther from the center pixel have smaller weights and have a relatively smaller impact on the filtering result.

[0083] The verification and optimization unit is used to simulate and analyze the image enhancement process, and to verify and optimize the image enhancement effect.

[0084] The working principle of the above technical solution is as follows: The image enhancement processing module of this invention first receives raw remote sensing image data, which typically comes from images acquired by means of UAV aerial photography, satellite remote sensing, etc. After the processing unit starts working, it uses virtual reality technology, combined with a pre-constructed digital twin model, to perform detailed analysis and enhancement processing on the received raw remote sensing images. In this process, not only are the basic attributes of the image, such as overall brightness and contrast, considered, but special attention is also paid to the noise problem in the image. By calculating the noise standard deviation of local areas of the image, the intensity and distribution of noise in the image can be accurately identified. Subsequently, according to the magnitude of the noise standard deviation, the parameters of the Gaussian filter are adaptively adjusted to ensure that while effectively suppressing noise, as much detail information in the image as possible is preserved. This adaptive adjustment mechanism enables the system to provide the best enhancement effect when facing different types of remote sensing images. After the initial processing by the processing unit, the enhanced remote sensing image is generated and passed to the verification and optimization unit. In this stage, the system uses simulation analysis technology to conduct in-depth verification and optimization of the image enhancement process. By comparing the image quality before and after processing, the system can accurately evaluate the effect of image enhancement and fine-tune the processing parameters based on the evaluation results to ensure that the final output enhanced remote sensing image reaches the best quality level.

[0085] The beneficial effects of the above technical solution are as follows: By using the solution provided in this embodiment, the effect of enhancing the original remote sensing image can be improved through enhancement processing, verification, and optimization; by calculating the standard deviation of pixel values ​​within the window and adaptively adjusting the parameters of the Gaussian filter based on the calculated noise standard deviation, image smoothing can be achieved. While suppressing noise, the edge and detail information of the image can be preserved as much as possible. When processing remote sensing images of power transmission and transformation lines, by adjusting the σ value, the filtering effect can be flexibly controlled according to the noise situation and image detail requirements of different areas, thereby achieving a better image enhancement purpose.

[0086] In one embodiment, virtual reality technology is used to enhance original remote sensing images based on a digital twin model to obtain enhanced remote sensing images, including:

[0087] Based on the digital twin model, a three-dimensional convolutional neural network algorithm is used to extract three-dimensional image features from the original remote sensing images to obtain three-dimensional image feature data;

[0088] Based on the feature data of three-dimensional images, three-dimensional modeling is performed using virtual reality technology to obtain a three-dimensional auxiliary model;

[0089] Based on a 3D auxiliary model, GAN image enhancement technology is used to enhance the original remote sensing images to obtain enhanced remote sensing images.

[0090] The working principle of the above technical solution is as follows: To more effectively enhance the original remote sensing image, this invention first utilizes the high simulation capability of the digital twin model to map the original remote sensing image into the digital twin space. Through a three-dimensional convolutional neural network algorithm, it deeply mines the three-dimensional feature information in the image. This feature data contains important information such as the spatial structure and texture details of the image, providing a solid foundation for subsequent enhancement processing. Next, based on the extracted three-dimensional image feature data and combined with virtual reality technology, a three-dimensional auxiliary model corresponding to the original remote sensing image is constructed. This model not only has a high degree of realism but also intuitively displays key information such as terrain features and route orientation in the image, providing technicians with a more intuitive and convenient observation method. Finally, based on the constructed three-dimensional auxiliary model, GAN (Generative Adversarial Network) image enhancement technology is used to further enhance the original remote sensing image. GAN technology, through adversarial training between the generator and discriminator, can continuously optimize the quality of the generated image, making it clearer and more detailed. In this process, the three-dimensional auxiliary model provides rich prior knowledge for GAN technology, guiding it to perform image enhancement more accurately, ultimately obtaining high-quality enhanced remote sensing images.

[0091] The beneficial effects of the above technical solution are as follows: by using the solution provided in this embodiment, high-quality enhanced remote sensing images can be obtained by enhancing the original remote sensing images based on the digital twin model and the three-dimensional auxiliary model.

[0092] In one embodiment, the image enhancement process is simulated and analyzed to verify and optimize the image enhancement effect, including:

[0093] Construct a standard image database that meets image enhancement standards; the standard image database stores several standard images;

[0094] Using a set simulation model, the process of enhancing the original remote sensing image is simulated and analyzed to obtain the simulation analysis results and the simulated enhanced image.

[0095] The resolution of the simulated enhanced image is compared and analyzed with that of the standard image. Based on the results of the comparison analysis, the simulated enhanced image is verified. If the resolution comparison value is less than the set resolution comparison threshold, the verification result of the simulated enhanced image is confirmed. If the resolution comparison value is greater than the set resolution comparison threshold, the resolution of the original remote sensing image corresponding to the simulated enhanced image is further optimized.

[0096] The working principle of the above technical solution is as follows: In order to simulate and analyze the image enhancement process and verify and optimize the image enhancement effect, this invention first constructs a standard image database containing numerous high-quality standard images that have undergone professional calibration. These images serve as reference benchmarks for image enhancement effects, ensuring the objectivity and accuracy of the evaluation. Using an advanced simulation model, the system can simulate the entire process of enhancing the original remote sensing image, thereby generating a simulated enhanced image. This step not only helps to understand the actual effect of the enhancement process but also provides data support for subsequent optimization. Next, the simulated enhanced image and the standard image are compared in detail at different resolutions. The comparative analysis process, based on high-precision image processing algorithms, accurately calculates the resolution difference between the simulated enhanced image and the standard image. By setting a resolution comparison threshold, the system can automatically determine whether the simulated enhanced image meets the preset quality standards. If the resolution comparison value of the simulated enhanced image is less than the set resolution comparison threshold, it is considered to meet the quality standards, and the system will confirm the verification result of the simulated enhanced image and use it as a candidate for subsequent processing or application. Conversely, if the resolution comparison value is greater than the set resolution comparison threshold, the system will automatically trigger a further optimization process for the resolution of the original remote sensing image in order to obtain a higher quality enhanced image.

[0097] The beneficial effects of the above technical solution are as follows: by using the solution provided in this embodiment, the image enhancement process is simulated and analyzed, and the image enhancement effect is verified and optimized, thus ensuring the quality of image enhancement.

[0098] In one embodiment, the resolution of the original remote sensing image corresponding to the simulated enhanced image is further optimized, including:

[0099] Acquire the original remote sensing image, and cluster the distribution of pixels in the original remote sensing image to obtain several clustering regions;

[0100] For clustered regions where the number of pixels exceeds a set threshold, pixel interpolation is used to enhance the resolution of the original remote sensing image and obtain the first enhanced remote sensing image.

[0101] For clustered regions where the number of pixels is less than a set threshold, the missing pixels are predicted based on a neural network prediction model to obtain the prediction results. Based on the prediction results, the missing pixels are averaged, removed, or supplemented to obtain the second enhanced remote sensing image.

[0102] By combining the pixel density of clustered regions and the local texture complexity, the weights of pixel interpolation and neural network prediction are set and dynamically adjusted. Based on the comprehensive weight composed of the pixel interpolation weights and the neural network prediction weights, the first and second enhanced remote sensing images are optimized and enhanced to obtain the enhanced first and second enhanced remote sensing images. The formula for calculating the comprehensive weight is:

[0103]

[0104] W i Represents the overall weight, α represents the adjustment factor, and N represents the overall weight. i N represents the number of pixels in the i-th cluster region; t C represents the total number of pixels. i Let C be the texture complexity of the i-th cluster region, and let C be the local window grayscale variance. max This represents the maximum texture complexity value; when W i When the weight exceeds the set weight threshold, interpolation pixel enhancement is prioritized to optimize and enhance the first enhanced remote sensing image; when W i When the weight threshold is less than or equal to the set weight threshold, a neural network is used to predict and supplement pixels to optimize and enhance the second enhanced remote sensing image.

[0105] By combining the enhanced first and second enhanced remote sensing images, a new remote sensing image with further optimized resolution is obtained.

[0106] The working principle of the above technical solution is as follows: Traditional resolution enhancement methods usually use a fixed threshold to divide the interpolation and prediction regions, but the pixel density and texture complexity of different regions in actual images vary significantly. For example, the conductor area of ​​a power transmission line (high density, low complexity) and the surrounding vegetation (low density, high complexity) require differentiated enhancement strategies. In order to further optimize the resolution of the original remote sensing image corresponding to the simulated enhanced image, this invention first acquires the original remote sensing image, clusters the distribution of pixels in the original remote sensing image, and obtains several clustering regions. The image is divided into multiple regions using a clustering algorithm (such as K-means), and the number of pixels in each region is counted. The density weight reflects the sufficiency of sampling in the region. High-density regions (such as conductors) are suitable for interpolation enhancement because their original information is sufficient. Then, clustering regions with a number of pixels greater than a set threshold are further... The resolution of the original remote sensing image is enhanced by pixel interpolation to obtain a first enhanced remote sensing image. Then, for clustered regions with fewer than a set threshold number of pixels, a neural network prediction model is used to predict missing pixels. High-complexity regions (such as vegetation) require neural network prediction to supplement detailed information. Based on the prediction results, missing pixels are averaged, removed, or added to obtain a second enhanced remote sensing image. Combining the pixel density and local texture complexity of the clustered regions, the weights of pixel interpolation and neural network prediction are set and dynamically adjusted. Based on the comprehensive weight composed of the pixel interpolation weight and the neural network prediction weight, the first and second enhanced remote sensing images are optimized and enhanced to obtain the enhanced first and second enhanced remote sensing images. The formula for calculating the comprehensive weight is:

[0107]

[0108] W i Represents the overall weight, α represents the adjustment factor, and N represents the overall weight. i N represents the number of pixels in the i-th cluster region; t C represents the total number of pixels. i Let C be the texture complexity of the i-th cluster region, and let C be the local window grayscale variance. max This represents the maximum texture complexity value; when W i When the weight exceeds the set weight threshold, interpolation pixel enhancement is prioritized to optimize and enhance the first enhanced remote sensing image; when W i When the weight threshold is less than or equal to the set weight threshold, a neural network is used to predict and supplement pixels to optimize and enhance the second enhanced remote sensing image, prioritizing the accuracy of high-density areas. Depending on actual needs (such as traverse clarity requirements), the weight threshold can be set to 0.5. iWhen the value is greater than 0.5, bicubic interpolation is used to enhance the resolution; otherwise, a convolutional neural network is used to predict missing pixels.

[0109] Finally, the first and second enhanced remote sensing images are combined to obtain the original remote sensing image with further optimized resolution.

[0110] The beneficial effects of the above technical solution are as follows: By adopting the solution provided in this embodiment, the resolution of the original remote sensing image corresponding to the simulation enhancement image is further optimized, which can avoid the blurring problem caused by interpolation in low-density areas, and the details in high-complexity areas are supplemented by neural networks, which can improve the recognition accuracy of targets such as tower base and vegetation boundary, and further ensure the effect of image enhancement.

[0111] In one embodiment, the method further includes managing and updating a standard image database, specifically including:

[0112] Based on the cloud platform, the standard image database is set as a cloud database;

[0113] The standard image database is managed using management programs on the cloud platform.

[0114] By utilizing big data from geographic information systems, the first updated data is periodically retrieved;

[0115] Image data acquired by taking pictures with a camera mounted on a drone is used as the second update data;

[0116] Based on the work records at the construction site of the power transmission and transformation line, close-range ground image data is obtained as the third update data;

[0117] The first, second, and third update data are used as update data for the standard image database, and the standard image database is updated accordingly.

[0118] The working principle of the above technical solution is as follows: In order to achieve efficient utilization of the standard image database, the present invention also includes management and updating of the standard image database. First, based on the cloud platform, the standard image database is set as a cloud database; then, based on the management program in the cloud platform, the standard image database is managed; finally, using big data from the geographic information system, the first update data is periodically retrieved; image data obtained by taking pictures using a camera mounted on a drone is used as the second update data; based on the work records of the power transmission and transformation line construction site, ground close-range image data is obtained as the third update data; finally, the first update data, the second update data, and the third update data are used as the update data of the standard image database to update the standard image database.

[0119] The beneficial effects of the above technical solution are as follows: by adopting the solution provided in this embodiment, efficient management of the standard image database can be achieved through management and updating of the standard image database.

[0120] In one embodiment, a result output and verification module is also included, used to verify the enhanced remote sensing imagery in practical applications, specifically including:

[0121] The verification of enhanced remote sensing image texture features is carried out as follows: using a set extraction template, the texture features of the enhanced remote sensing image are extracted and texture feature description information is obtained; the texture feature description information is compared and analyzed with the data in the set texture description information database to obtain the comparison results; and the first verification result is obtained based on the comparison results.

[0122] The identification of specific targets in enhanced remote sensing imagery is carried out as follows: using a target identification model, targets are identified in the enhanced remote sensing imagery to obtain target identification results; based on the accuracy of the target identification results, a second verification result is obtained.

[0123] Based on the first and second verification results, the verification results for the enhanced remote sensing image are obtained.

[0124] The working principle of the above technical solution is as follows: To better verify the application of enhanced remote sensing images, this invention also includes a result output and verification module, used to verify the enhanced remote sensing images in practical applications. First, the texture features of the enhanced remote sensing images are verified. Specifically, using a set extraction template, the texture features of the enhanced remote sensing images are extracted, and texture feature description information is obtained. Based on the texture feature description information, a comparative analysis is performed with data in a set texture description information database to obtain a comparison result. Based on the comparison result, a first verification result is obtained. Next, the specific target identification of the enhanced remote sensing images is performed. Specifically, using a target identification model, target identification is performed on the enhanced remote sensing images to obtain target identification results. Based on the accuracy of the target identification results, a second verification result is obtained. Finally, based on the first and second verification results, the verification result of the enhanced remote sensing image is obtained.

[0125] The beneficial effects of the above technical solution are as follows: by adopting the solution provided in this embodiment, the application of enhanced remote sensing images is verified, which provides a foundation for the efficient utilization of enhanced remote sensing images.

[0126] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A remote sensing image enhancement system for power transmission and transformation lines based on digital twins, characterized in that, include: The remote sensing image acquisition module is used to acquire raw remote sensing images of power transmission and transformation lines. The digital twin model building module is used to construct a virtual digital twin model of a power transmission and transformation line based on the original remote sensing imagery using digital twin technology. The image enhancement processing module is used to enhance the original remote sensing images based on the virtual digital twin model, and to verify and optimize them to obtain enhanced remote sensing images. Specifically: Using virtual reality technology and based on a digital twin model, the original remote sensing images are enhanced to obtain enhanced remote sensing images; The image enhancement process is simulated and analyzed to verify and optimize the image enhancement effect; specifically including: A standard image database meeting image enhancement standards is constructed; the standard image database stores several standard images; using a set simulation model, the process of enhancing the original remote sensing images is simulated and analyzed to obtain simulation analysis results and simulated enhanced images; the resolution of the simulated enhanced images is compared and analyzed with the standard images, and the simulated enhanced images are verified based on the comparison analysis results; if the resolution comparison value is less than the set resolution comparison threshold, the verification result of the simulated enhanced images is confirmed; if the resolution comparison value is greater than the set resolution comparison threshold, the resolution of the original remote sensing images corresponding to the simulated enhanced images is further optimized; Further optimization of the resolution of the original remote sensing image corresponding to the simulated enhanced image, including: The process involves acquiring original remote sensing images, clustering the pixel distribution of these images to obtain several clustered regions. For clustered regions with more than a set threshold number of pixels, pixel interpolation is used to enhance the resolution of the original remote sensing images, resulting in a first enhanced remote sensing image. For clustered regions with fewer than the set threshold number of pixels, a neural network prediction model is used to predict missing pixels, and the prediction results are then used to average, remove, or supplement the missing pixels, resulting in a second enhanced remote sensing image. By combining the pixel density of clustered regions and the local texture complexity, the weights of pixel interpolation and neural network prediction are set and dynamically adjusted. Based on the comprehensive weight composed of the pixel interpolation weights and the neural network prediction weights, the first and second enhanced remote sensing images are optimized and enhanced to obtain the enhanced first and second enhanced remote sensing images. The formula for calculating the comprehensive weight is: W i Represents the overall weight, α represents the adjustment factor, and N represents the overall weight. i N represents the number of pixels in the i-th cluster region; t C represents the total number of pixels. i Let C be the texture complexity of the i-th cluster region, and let C be the local window grayscale variance. max This represents the maximum texture complexity value; when W i When the weight exceeds the set weight threshold, interpolation pixel enhancement is prioritized to optimize and enhance the first enhanced remote sensing image; when W i When the weight threshold is less than or equal to the set weight threshold, a neural network is used to predict and supplement pixels to optimize and enhance the second enhanced remote sensing image. By combining the enhanced first and second enhanced remote sensing images, a new remote sensing image with further optimized resolution is obtained.

2. The remote sensing image enhancement system for power transmission and transformation lines based on digital twins according to claim 1, characterized in that, Acquire raw remote sensing images of power transmission and transformation lines, including: Configure the image acquisition equipment for accessing high-resolution satellites and set the operating parameters of the image acquisition equipment; Based on the operating parameters, image acquisition equipment is used to receive and acquire raw remote sensing images of power transmission and transformation lines via high-resolution satellites.

3. The remote sensing image enhancement system for power transmission and transformation lines based on digital twins according to claim 2, characterized in that, The original remote sensing images include engineering project images of power transmission and transformation lines and images of the surrounding environment; engineering project images include, but are not limited to, images of the specific route trajectory, deployment scale, and construction site location of the engineering project; images of the surrounding environment include, but are not limited to, images of buildings, vegetation cover, and road topography.

4. The remote sensing image enhancement system for transmission and transformation lines based on digital twins according to claim 1, characterized in that, The digital twin model building module includes an image processing unit and a model building unit; The image processing unit is used to perform grayscale preprocessing on the original remote sensing images and to extract image features to obtain target image features for building the model. The model building unit is used to construct a digital twin model based on the features of the target image, relying on the Unity3D platform, and based on the set model algorithm, model structure, and model parameters.

5. The remote sensing image enhancement system for transmission and transformation lines based on digital twins according to claim 1, characterized in that, Using virtual reality technology and based on a digital twin model, the original remote sensing images are enhanced to obtain enhanced remote sensing images, including: Based on the digital twin model, a three-dimensional convolutional neural network algorithm is used to extract three-dimensional image features from the original remote sensing images to obtain three-dimensional image feature data; Based on the feature data of three-dimensional images, three-dimensional modeling is performed using virtual reality technology to obtain a three-dimensional auxiliary model; Based on a 3D auxiliary model, GAN image enhancement technology is used to enhance the original remote sensing images to obtain enhanced remote sensing images.

6. The remote sensing image enhancement system for transmission and transformation lines based on digital twins according to claim 1, characterized in that, It also includes the management and updating of the standard image database, specifically including: Based on the cloud platform, the standard image database is set as a cloud database; The standard image database is managed using management programs on the cloud platform. By utilizing big data from geographic information systems, the first updated data is periodically retrieved; Image data acquired by taking pictures with a camera mounted on a drone is used as the second update data; Based on the work records at the construction site of the power transmission and transformation line, close-range ground image data is obtained as the third update data; The first, second, and third update data are used as update data for the standard image database, and the standard image database is updated accordingly.

7. The remote sensing image enhancement system for transmission and transformation lines based on digital twins according to claim 1, characterized in that, It also includes a results output and verification module, used to verify the enhanced remote sensing imagery in practical applications, specifically including: The verification of enhanced remote sensing image texture features is carried out as follows: using a set extraction template, the texture features of the enhanced remote sensing image are extracted and texture feature description information is obtained; the texture feature description information is compared and analyzed with the data in the set texture description information database to obtain the comparison results; and the first verification result is obtained based on the comparison results. The identification of specific targets in enhanced remote sensing imagery is carried out as follows: using a target identification model, targets are identified in the enhanced remote sensing imagery to obtain target identification results; based on the accuracy of the target identification results, a second verification result is obtained. Based on the first and second verification results, the verification results for the enhanced remote sensing image are obtained.

Images