A GNSS data-based method, device and medium for identifying obstructions above a survey station

Abstract

This invention belongs to the field of GNSS observation environment occlusion recognition technology, specifically involving a method, device, and medium for identifying occlusion over a GNSS station based on GNSS data. The method includes the following steps: collecting GNSS observation data under various occlusion environments; calculating the elevation angle, carrier-to-noise ratio (CNR), CNR fitting residual, and pseudorange consistency for each satellite; acquiring occlusion information for each station using a fisheye camera; dividing the airspace above the station into grids based on azimuth and elevation angles; constructing training samples using the data quality and corresponding occlusion information within the grids; training an occlusion recognition model based on the samples; predicting the occlusion situation of new stations; optimizing the prediction results; and obtaining the final occlusion recognition result. This invention requires no additional sensor support, achieving occlusion recognition over a station solely based on GNSS data. It is suitable for quality assessment and station site selection in complex environments such as urban areas and forest areas, and also provides a reference for optimizing stochastic models in subsequent positioning calculations.

Description

Technical Field

[0001] This invention belongs to the field of GNSS observation environment obstruction identification technology, specifically involving a method, device and medium for identifying obstruction over the station based on GNSS data. Background Technology

[0002] With the continuous development of Global Navigation Satellite System (GNSS) technology, it plays an increasingly important role in many fields such as traffic navigation, geographic information collection, disaster monitoring, and autonomous driving. Especially in complex environments such as cities, mountains, and forests, the demand for high-precision GNSS positioning is growing. However, obstacles such as tall buildings and trees in the surrounding environment can obstruct or interfere with GNSS signals, causing problems such as signal attenuation, reflection, and multipath effects, which seriously affect the quality of GNSS observation data.

[0003] In complex environments, GNSS data errors caused by obstruction factors (such as observation interruptions, decreased carrier-to-noise ratio, and enhanced multipath effects) often lead to gross errors and abnormal fluctuations in pseudorange or carrier measurements, posing a challenge to the accuracy and stability of GNSS high-precision positioning technology. Traditional gross error detection and multipath suppression methods mostly rely on statistical analysis or model constraints, which lack sufficient understanding of the spatial distribution information of obstructed areas and cannot accurately identify signal anomalies caused by obstruction.

[0004] Some existing research methods attempt to extract occlusion information with the help of auxiliary sensors (such as lidar, panoramic or fisheye cameras), but these methods usually face problems such as strong equipment dependence, complex deployment and high cost.

[0005] In summary, there is an urgent need for a method that can effectively identify obstructed areas above a GNSS station without relying on external sensors, using only GNSS observation data. This would provide a reference for GNSS station site selection optimization, data quality assessment, and data processing strategy optimization. Summary of the Invention

[0006] The purpose of this invention is to provide a method, device, and medium for identifying overhead obstruction based on GNSS data. The specific technical solution is as follows:

[0007] A method for identifying overhead obstruction based on GNSS data includes the following steps:

[0008] S1: Collect GNSS observation data under various obstruction environments, and calculate the elevation angle, carrier-to-noise ratio, carrier-to-noise ratio fitting residual, and pseudorange consistency for each satellite;

[0009] S2: Use a fisheye camera to collect sky images above the station and extract the occlusion area information corresponding to each station through image recognition methods;

[0010] S3: Divide the sky above the station into grids according to azimuth and elevation angles, construct a training sample set with GNSS observation quality indicators in each grid cell as input features and corresponding occlusion conditions as classification labels, train the model based on the training sample set, and use the trained model to predict the occlusion conditions of the station to obtain the prediction results.

[0011] S4: Optimize the prediction results obtained in S3 to obtain the final occlusion recognition results.

[0012] Optionally, in S1, the GNSS observation data acquisition duration is at least 20 hours; and the carrier-to-noise ratio fitting residual is the difference between the actual carrier-to-noise ratio of each satellite and its fitted carrier-to-noise ratio at the same elevation angle in an open environment, where the satellite is in an open environment. At elevation angle Fitting noise ratio The calculation method is as follows:

[0013] ;

[0014] in, Indicates the elevation angle. Indicates satellite The fitting coefficient;

[0015] The specific calculation method for the carrier-to-noise ratio fitting residual is as follows:

[0016] ;

[0017] in, To block satellites in the environment At elevation angle Carry-to-noise ratio fitting residual at that time To block satellites in the environment At elevation angle The carrier-to-noise ratio at that time.

[0018] Optionally, in S1, pseudorange consistency is the difference between the inter-epoch differences in pseudorange and the average variation calculated from Doppler observations, and is calculated as follows:

[0019] ;

[0020] in, For pseudorange consistency, and These are the pseudorange observations for the current epoch and the previous epoch, respectively. The wavelength at the current frequency. and These are the Doppler observations for the current epoch and the previous epoch, respectively. The interval is the epoch.

[0021] Optionally, in S2, after acquiring images of the area above the station, the images are semantically segmented using image recognition methods to divide the obstructed areas into tree obstruction and building obstruction. Tree obstruction is an area that satellite signals can penetrate, while building obstruction is an area that satellite signals cannot penetrate.

[0022] Optionally, in S3, the area above the station can be divided into grids based on azimuth and elevation angles, as follows:

[0023] The hemispherical space above the station is divided into 2.5°×2.5° grid units based on azimuth and elevation angles.

[0024] Optionally, in S3, within each grid cell, GNSS observation quality indicators are statistically extracted as input features of the model, and the occlusion type of the corresponding grid cell obtained in S2 is used as a classification label to construct a training sample set. The GNSS observation quality indicators include at least:

[0025] Angle of elevation;

[0026] The mean, standard deviation, maximum, minimum, and number of carrier-to-noise ratio values;

[0027] The mean, standard deviation, maximum, minimum, and number of the carrier-to-noise ratio fitting residuals;

[0028] Mean and standard deviation of pseudorange consistency.

[0029] Optionally, in S3, the model uses a lightweight gradient booster model.

[0030] Optionally, in S4, the prediction results are optimized, as follows:

[0031] The predicted occlusion mask is analyzed using an 8-adjacent connected component analysis, and its area is measured by the number of grid cells.

[0032] When the area of ​​a tree-shaded or building-shaded area is less than a preset threshold of 10, it is considered an invalid area and removed, and the area is then filled with the category of the grid closest to the removed area.

[0033] Based on the category distribution within the 8×8 sliding neighborhood window, a majority voting strategy is used to reassign the category of each grid cell.

[0034] In addition, the present invention also provides a computer device, including a memory and a processor;

[0035] The memory is used to store computer programs that can run on the processor;

[0036] When the processor executes the computer program, it implements the steps of the above-described method for identifying obstructions above a measuring station.

[0037] In addition, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for identifying obstructions above a measuring station.

[0038] The application of the technical solution of the present invention has the following beneficial effects:

[0039] This invention proposes a method for identifying overhead occlusion based on GNSS data. The method divides the hemispherical space above the station into a two-dimensional grid according to azimuth and elevation angles. GNSS data quality indicators are statistically analyzed within each grid cell, and an occlusion category is generated for each grid cell using a fisheye camera. A training sample set containing data quality features and corresponding occlusion categories is constructed to train an occlusion classification model based on a lightweight gradient booster. After training, the model predicts the overhead occlusion distribution of a new station without relying on external sensors, using only GNSS observation data. To improve the spatial continuity and classification stability of the model output, this invention further performs 8-adjacent connected component analysis on the predicted occlusion mask map, removing isolated occlusion regions with areas smaller than a preset threshold. For the removed regions, interpolation is performed using the categories of neighboring pixels. Subsequently, based on a majority class statistical strategy within a local sliding window, the occlusion results are smoothed to finally obtain the occlusion distribution map of the station.

[0040] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description

[0041] To more clearly illustrate the technical solutions of the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a flowchart of the steps of the method for identifying overhead obstruction in a preferred embodiment of the present invention;

[0043] Figure 2 This is an occlusion recognition image of a fisheye camera in a preferred embodiment of the present invention;

[0044] Figure 3 This is an occlusion recognition diagram of the occlusion recognition method above the measuring station in a preferred embodiment of the present invention;

[0045] Figure 4 This is a satellite trajectory and carrier-to-noise ratio distribution map of the overhead obstruction identification method in a preferred embodiment of the present invention. Detailed Implementation

[0046] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] like Figure 1 As shown in the figure, this embodiment provides a method for identifying overhead obstruction based on GNSS data, including the following steps:

[0048] S1: Collect GNSS observation data under various obstruction environments, and calculate the elevation angle, carrier-to-noise ratio, carrier-to-noise ratio fitting residual, and pseudorange consistency for each satellite;

[0049] S2: Use a fisheye camera to collect sky images above the station and extract the occlusion area information corresponding to each station through image recognition methods;

[0050] S3: Divide the sky above the station into grids according to azimuth and elevation angles, construct a training sample set with GNSS observation quality indicators in each grid cell as input features and corresponding occlusion conditions as classification labels, train the model based on the training sample set, and use the trained model to predict the occlusion conditions of the station to obtain the prediction results.

[0051] S4: Optimize the prediction results obtained in S3 to obtain the final occlusion recognition results.

[0052] Optionally, in S1, the GNSS observation data acquisition duration is at least 20 hours; and the carrier-to-noise ratio fitting residual is the difference between the actual carrier-to-noise ratio of each satellite and its fitted carrier-to-noise ratio at the same elevation angle in an open environment, where the satellite is in an open environment. At elevation angle Fitting noise ratio The calculation method is as follows:

[0053] ;

[0054] in, Indicates the elevation angle. Indicates satellite The fitting coefficient;

[0055] The specific calculation method for the carrier-to-noise ratio fitting residual is as follows:

[0056] ;

[0057] in, To block satellites in the environment At elevation angle Carry-to-noise ratio fitting residual at that time To block satellites in the environment At elevation angle The carrier-to-noise ratio at that time.

[0058] Optionally, in S1, pseudorange consistency is the difference between the inter-epoch differences in pseudorange and the average variation calculated from Doppler observations, and is calculated as follows:

[0059] ;

[0060] in, For pseudorange consistency, and These are the pseudorange observations for the current epoch and the previous epoch, respectively. The wavelength at the current frequency. and These are the Doppler observations for the current epoch and the previous epoch, respectively. The interval is the epoch.

[0061] Optionally, in S2, after acquiring images of the area above the station, the images are semantically segmented using image recognition methods to divide the obstructed areas into tree obstruction and building obstruction. Tree obstruction is an area that satellite signals can penetrate, while building obstruction is an area that satellite signals cannot penetrate.

[0062] Optionally, in S3, the area above the station can be divided into grids based on azimuth and elevation angles, as follows:

[0063] The hemispherical space above the station is divided into 2.5°×2.5° grid units based on azimuth and elevation angles.

[0064] Optionally, in S3, within each grid cell, GNSS observation quality indicators are statistically extracted as input features of the model, and the occlusion type of the corresponding grid cell obtained in S2 is used as a classification label to construct a training sample set. The GNSS observation quality indicators include at least:

[0065] Angle of elevation;

[0066] The mean, standard deviation, maximum, minimum, and number of carrier-to-noise ratio values;

[0067] The mean, standard deviation, maximum, minimum, and number of the carrier-to-noise ratio fitting residuals;

[0068] Mean and standard deviation of pseudorange consistency.

[0069] Optionally, in S3, the model uses a lightweight gradient booster model.

[0070] Furthermore, the model training in this embodiment adopted the following metrics:

[0071] Precision: Represents the proportion of samples predicted to belong to a certain class that actually do not; it is used to measure the accuracy of the prediction results.

[0072] ;

[0073] Recall: Represents the proportion of samples that actually belong to a certain class that are correctly predicted; it measures the model's ability to cover samples of that class.

[0074] ;

[0075] F1 score: The harmonic mean of precision and recall, which comprehensively reflects the accuracy and completeness of the model.

[0076] ;

[0077] Accuracy: The proportion of correct predictions out of all predictions, reflecting overall prediction performance.

[0078] ;

[0079] in, TP The number of samples that are true to that category is also the number of samples that are predicted to be of that category. FP This represents the number of samples that do not actually belong to this category, but the model predicts they belong to this category. TN This refers to the number of samples that do not actually belong to this category, and whose predictions also do not fall into this category. FN This represents the number of samples that actually belong to this category, but the model did not predict as belonging to this category.

[0080] Optionally, in S4, the prediction results are optimized, as follows:

[0081] The predicted occlusion mask is analyzed using an 8-adjacent connected component analysis, and its area is measured by the number of grid cells.

[0082] When the area of ​​a tree-shaded or building-shaded area is less than a preset threshold of 10, it is considered an invalid area and removed, and the area is then filled with the category of the grid closest to the removed area.

[0083] Based on the category distribution within the 8×8 sliding neighborhood window, a majority voting strategy is used to reassign the category of each grid cell.

[0084] This invention provides a method for identifying overhead occlusion based on GNSS data. This method overcomes the problems of strong dependence on external sensors, high cost, and complex deployment in existing technologies. Specifically, this method divides the hemispherical space above the station into a 2.5°×2.5° two-dimensional grid according to azimuth and elevation angles. It statistically analyzes the GNSS data quality indicators within each grid cell and uses a fisheye camera to generate an occlusion category for each grid. A training sample set containing data quality features and corresponding occlusion categories is constructed to train an occlusion classification model based on a lightweight gradient booster. After training, the model does not rely on external sensors and predicts the overhead occlusion distribution of new stations solely based on GNSS observation data. To improve the spatial continuity and classification stability of the model output, this invention further performs 8-adjacent connected component analysis on the predicted occlusion mask map, removing isolated occlusion regions with areas smaller than a preset threshold. For the removed regions, interpolation is performed using the categories of neighboring pixels. Subsequently, based on the majority category statistical strategy within the local sliding window, the occlusion results were smoothed to finally obtain the occlusion distribution map of the station.

[0085] To verify the effectiveness of the algorithm, this embodiment set up 6 stations. GNSS data from stations 1, 2, 3, 4, and 5 were selected for training and divided into training and validation sets in an 8:2 ratio for model training and evaluation. The classification performance of the model on the validation set is shown in Table 1.

[0086] Table 1 Model training results

[0087]

[0088] In the sky category recognition, the model achieved a precision of 0.87, indicating that 87% of the samples predicted as sky actually belonged to the sky region; and a recall of 0.84, indicating that 84% of all real sky samples were successfully identified. The balance between these two metrics resulted in an F1 score of 0.86, demonstrating that the model performed well in this category with few false negatives and false positives.

[0089] For the building occlusion category, the model achieved the highest recall (0.93), indicating that the vast majority of building occlusion areas were accurately identified; its precision was 0.85, suggesting that some other categories were still misclassified as building occlusion. Overall, the identification of this category demonstrates good completeness and strong discriminative ability.

[0090] In terms of tree occlusion recognition, the model's recall rate was 0.82, slightly lower than other categories, indicating that a certain proportion of tree occlusion areas were not identified; however, the precision rate was 0.87, reflecting that the model's judgment of the identified tree occlusion areas was relatively accurate.

[0091] In summary, the model constructed by the method in this embodiment has balanced recognition performance across the three occlusion categories, with an overall recognition accuracy of 86%, which can meet the practical application requirements of GNSS station occlusion recognition in complex environments.

[0092] To verify the generalization ability of the constructed model, station 6, which was not involved in the training, was selected as the test sample. The trained model was used to predict the overhead occlusion situation at this station, and the identification statistics are shown in Table 2. The results show that the model's recognition performance in the tree occlusion category is relatively poor, with many other types of grids being misclassified as tree occlusion areas. The recognition results output by the optimized model are shown in Tables 3 and 4. Statistical analysis shows that the optimized model significantly reduces misclassifications in the tree occlusion category, with only a small number of tree areas being misclassified as sky areas, and the overall recognition accuracy improving to 93%.

[0093] Table 2 shows the model's prediction results for the new data.

[0094]

[0095] Table 3 Prediction results of the optimized model

[0096]

[0097] Table 4. Confusion matrix of the optimized model prediction results

[0098]

[0099] In addition, through comparison Figure 2 Occlusion map generated by fisheye camera and Figure 3 The occlusion maps identified by the method of this invention show that the two have good consistency in spatial distribution. Further combining... Figure 4 Analysis of satellite trajectory and carrier-to-noise ratio distribution information shows that: in areas where the model determines that the satellite signal is blocked by buildings, it is almost impossible to receive the satellite signal, indicating that the satellite signal is completely blocked; in areas blocked by trees, the satellite signal can be received, but its carrier-to-noise ratio is significantly reduced, which verifies the rationality and effectiveness of the method of the present invention in occlusion identification.

[0100] This invention also discloses a computer device, including a memory and a processor;

[0101] The memory is used to store computer programs that can run on the processor;

[0102] The processor is used to implement the steps of the land use classification method described above when executing the computer program.

[0103] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the computer device.

[0104] The computer device may be a mobile phone, desktop computer, laptop, handheld computer, or cloud server, etc. The computer device may include, but is not limited to, a processor and memory. For example, the computer device may also include input / output devices, network access devices, buses, etc.

[0105] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the computer device, connecting various parts of the computer device via various interfaces and lines.

[0106] The memory can be used to store the computer program and / or modules. The processor implements the computer program by running or executing the computer program and / or modules stored in the memory, and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital card (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0107] Wherein, if the modules / units integrated into the computer device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0108] In addition, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the land use classification method described above.

[0109] Compared to existing technologies, this invention divides the hemispherical space above the station into a two-dimensional grid based on azimuth and elevation angles. It statistically analyzes the GNSS data quality indicators within each grid cell and uses a fisheye camera to generate occlusion categories for each grid. This constructs a training sample set containing data quality features and corresponding occlusion categories, enabling the training of an occlusion classification model based on LightGBM. After training, the model of this invention does not rely on external sensors, but only on GNSS observation data to predict the occlusion distribution above a new station. To improve the spatial continuity and classification stability of the model output, this invention further performs 8-adjacent connected component analysis on the predicted occlusion mask map, removing isolated occlusion regions with areas smaller than a preset threshold. For the removed regions, interpolation is performed using the categories of neighboring pixels. Subsequently, based on a majority class statistical strategy within a local sliding window, the occlusion results are smoothed, ultimately obtaining the occlusion distribution map of the station.

[0110] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0111] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

Claims

1. A GNSS data-based identification method for over-station occlusion, characterized in that, The method comprises the following steps: S1: Collect GNSS observation data under various shielding environments, and calculate the elevation angle, carrier-to-noise ratio, carrier-to-noise ratio fitting residual and pseudorange consistency of each satellite; S2: Collect the sky image above the station by using a fisheye camera, and extract the shielding area information corresponding to each station by using an image recognition method; S3: Divide the space above the station into grids according to the azimuth angle and the elevation angle, construct a training sample set taking the GNSS observation quality indicators in each grid unit as input features and the corresponding shielding conditions as classification labels, train a model based on the training sample set, and predict the shielding conditions of the station by using the trained model to obtain a prediction result; S4: Optimize the prediction result obtained in S3 to obtain a final shielding recognition result; In S3, the GNSS observation quality indicators are extracted in each grid unit as input features of the model, and the shielding type of the corresponding grid unit obtained in S2 is taken as a classification label to construct a training sample set. The GNSS observation quality indicators at least include: Elevation angle; Mean, standard deviation, maximum value, minimum value and number of carrier-to-noise ratio; Mean, standard deviation, maximum value, minimum value and number of carrier-to-noise ratio fitting residual; Mean and standard deviation of pseudorange consistency; The model adopts a light gradient boosting machine model; In S4, the prediction result is optimized as follows: Perform 8-neighbor connected component analysis on the predicted shielding mask, and measure the area of the shielding mask by the number of grid units; When the area of a tree shielding area or a building shielding area is less than a preset threshold 10, the area is regarded as an invalid area and is removed, and the nearest grid category to the removed area is filled in the removed area; Based on the category distribution in the 8*8 sliding neighborhood window, the majority voting strategy is adopted to reassign the category of each grid unit.

2. The method of station sky obstruction identification according to claim 1, characterized in that, In S1, the collection time length of GNSS observation data is at least 20 hours; and the fitting residual of the carrier-to-noise ratio is the difference between the actual carrier-to-noise ratio of each satellite and the fitting carrier-to-noise ratio at the same height angle in an open environment, and the satellite in the open environment is at the same height angle The calculation method of the fitting carrier-to-noise ratio of the satellite at the height angle is as follows: ； wherein, denotes the elevation angle, denotes the satellite fitting coefficients; The specific calculation method of the carrier-to-noise ratio fitting residual is as follows: ； wherein, Satellite in occluded environment Carrier-to-noise ratio fitting residual at elevation angle Satellite in occluded environment Carrier-to-noise ratio fitting residual at elevation angle Satellite in occluded environment Carrier-to-noise ratio fitting residual at elevation angle 3. The method of claim 1, wherein In S1, the pseudorange consistency is the difference between the pseudorange epoch difference value and the average change value calculated from the Doppler observation, and the calculation method is as follows: ； wherein, is the pseudorange consistency, and are the pseudorange observations at the current epoch and the previous epoch, respectively, is the wavelength of the current frequency, and are the Doppler observations at the current epoch and the previous epoch, respectively, is the epoch interval.

4. The method of claim 1, wherein, In S2, after collecting the image above the station, the image is processed by semantic segmentation through an image recognition method, and the shielding area is divided into tree shielding and building shielding. The tree shielding is a satellite signal penetrable area, and the building shielding is a satellite signal non-penetrable area.

5. The method of claim 1, wherein, In S3, the space above the station is divided into grids according to the azimuth angle and the elevation angle, and the specific method is as follows: The hemispherical space above the station is divided into 2.5°*2.5° grid units according to the azimuth angle and the elevation angle.

6. A computer device, comprising: The computer program is stored in the memory and executed on the processor. The computer program is stored in the memory and executed on the processor. The computer program is stored in the memory and executed on the processor.

7. A computer readable storage medium characterized in that, The computer program is stored in the memory and executed on the processor.

Images