Mountain site identification method and device, equipment, storage medium and program product

Abstract

The invention discloses a mountain site identification method and device, equipment, a storage medium and a program product, and relates to the technical field of position detection. The method comprises the following steps: acquiring position coordinates and altitude of a target station, and position coordinates and altitude of a plurality of preset positions in an area where the target station is located; subtracting a preset reference altitude from the altitude of the target station and the altitudes of the plurality of preset positions to obtain a regional height difference of the target station and a regional height difference of the plurality of preset positions; generating a grayscale image according to the region height difference of the target station, the region height difference of the plurality of preset positions, the position coordinates of the target station and the position coordinates of the plurality of preset positions; and inputting the grayscale image into the mountain site identification model, determining a classification value of the target site according to the model weight parameter of the mountain site identification model and the grayscale image, and determining the site type of the target site according to the classification value of the target site. According to the embodiment of the invention, the accuracy of mountain site identification is improved.

Description

Technical Field

[0001] This application belongs to the field of location detection technology, and in particular relates to a method, apparatus, equipment, storage medium and program product for identifying mountain sites. Background Technology

[0002] Mountain site identification determines whether a site is situated within a mountain, thereby assessing the rationality of the site selection and the appropriateness of the antenna mounting height. It also helps identify issues such as excessively close, excessively distant, excessively high, and excessively overlapping coverage, thus assisting in site selection.

[0003] Existing methods for identifying mountain sites often rely on manual verification during the site selection process, assessing the presence of mountains at the site location and the rationality of antenna height. However, manual verification suffers from poor accuracy and reliability. Furthermore, with the rapid development and expansion of operators' businesses, the task of site selection is increasing, and existing manual verification methods are prone to errors when faced with a large volume of site selection tasks. Therefore, the accuracy of existing methods for identifying mountain sites is relatively low. Summary of the Invention

[0004] This application provides a method, apparatus, device, storage medium, and program product for identifying mountain sites, in order to solve the problem of low accuracy in existing mountain site identification methods.

[0005] In a first aspect, embodiments of this application provide a method for identifying mountain sites, the method comprising: Acquire regional image information of the target site, including the location coordinates and altitude of the target site, as well as the location coordinates and altitude of multiple preset locations within the area where the target site is located; Subtract the preset baseline altitude from the altitude of the target station and the altitude of multiple preset locations to obtain the regional altitude difference of the target station and the regional altitude difference of multiple preset locations. A grayscale image is generated based on the regional height difference of the target site and the regional height differences of multiple preset locations, as well as the location coordinates of the target site and the location coordinates of multiple preset locations. The grayscale image is input into the mountain site recognition model. The classification value of the target site is determined based on the model weight parameters of the mountain site recognition model and the grayscale image. The site type of the target site is then determined based on the classification value of the target site.

[0006] Secondly, embodiments of this application provide a device for identifying mountain monitoring stations, the device comprising: The acquisition module is used to acquire regional image information of the target site. The regional image information includes the location coordinates and altitude of the target site, as well as the location coordinates and altitude of multiple preset locations within the area where the target site is located. The calculation module is used to subtract the preset reference altitude from the altitude of the target station and the altitude of multiple preset locations respectively to obtain the regional altitude difference of the target station and the regional altitude difference of multiple preset locations. The generation module is used to generate a grayscale image based on the regional height difference of the target site and the regional height difference of multiple preset locations, the location coordinates of the target site and the location coordinates of multiple preset locations. The determination module is used to input grayscale images into the mountain site recognition model, determine the classification value of the target site based on the model weight parameters of the mountain site recognition model and the grayscale image, and determine the site type of the target site based on the classification value of the target site.

[0007] Thirdly, embodiments of this application provide a terminal device, which includes: a processor and a memory storing computer program instructions; the processor executes the computer program instructions to implement the mountain site identification method as described in the first aspect.

[0008] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the mountain site identification method as described in the first aspect.

[0009] Fifthly, embodiments of this application provide a computer program product in which instructions, when executed by a processor of an electronic device, cause the electronic device to perform the mountain site identification method as described in the first aspect.

[0010] This application provides a method, apparatus, device, storage medium, and program product for identifying mountain sites. By acquiring regional image information of the target site and calculating the regional height difference of the target site, the interference of different reference elevations caused by different sea levels in different regions is eliminated. Furthermore, the regional height difference reflects the relative terrain changes within the area where the target site is located, thereby improving the accuracy of terrain information acquisition. A grayscale image is generated based on the regional height difference, the location coordinates of the target site, and the location coordinates of multiple preset locations. The geographical information represented by the regional height difference is converted into image features that the mountain site identification model can learn, thus converting multiple regional height differences into regional terrain features. This grayscale image is input into the mountain site identification model, and the classification value of the target site is determined through model weight parameters and the grayscale image. That is, the regional terrain features in the grayscale image are extracted using preset model parameters to obtain the classification value corresponding to the regional terrain features, improving the accuracy of terrain information extraction. Finally, the site type of the target site is classified according to the classification value, thereby improving the accuracy of mountain site identification. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the method for identifying mountain sites provided in an embodiment of this application; Figure 2 This is a topographical diagram of the area where the target site is located, provided in an embodiment of this application. Figure 3 This is a grayscale image diagram of the area where the target site is located, provided in an embodiment of this application. Figure 4 This application provides a method for determining a mountain site identification model. Figure 5 This is another method for determining a mountain site identification model provided in the embodiments of this application; Figure 6 This is a schematic diagram illustrating the verification results of the mountain site identification model provided in the embodiments of this application; Figure 7 This is a schematic diagram of the structure of the mountain site identification device provided in the embodiments of this application; Figure 8 This is a schematic diagram of the structure of the terminal device provided in the embodiments of this application. Detailed Implementation

[0013] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are intended only to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0014] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0015] Mountain site identification involves determining the rationality of a site selection and the appropriateness of antenna height by assessing whether the terrain is mountainous. For example, the antenna height for a macro site on a non-mountainous terrain should not be less than 15 meters, while for a macro site on a mountainous terrain, it should not be less than 6 meters. This process then assists in reviewing the planning scheme, helping reviewers assess the site's coverage for four violations (excessive proximity, excessive distance, excessive height, and excessive overlap), thus identifying sites that exceed these limits. Current mountain site identification methods rely on manual review of antenna height and other factors during site construction approval. However, with the rapid development and expansion of operators' business, the demand for site construction has increased, leading to a surge in site selection tasks. Manual review is costly and inefficient, and its reliance on the experience of reviewers results in poor accuracy and reliability. Therefore, traditional manual site selection methods can no longer meet the current needs.

[0016] To address the problems of existing technologies, this application discloses a method, apparatus, device, storage medium, and program product for identifying mountain sites. By acquiring regional image information of the target site and calculating the regional height difference of the target site, the interference of absolute landmark references in different regions is eliminated, reflecting the relative terrain changes within the area where the target site is located, thus improving the accuracy of terrain information acquisition. A grayscale image is generated based on the regional height difference and location coordinates, converting geographical information into visual features that the mountain site identification model can learn. The grayscale image includes the surrounding terrain features of the target site's area. This grayscale image is input into the mountain site identification model, which can calculate the classification value of the target site based on the surrounding terrain features, improving the comprehensiveness of terrain information extraction. Finally, the target site is classified according to the classification value, thereby improving the accuracy of mountain site identification.

[0017] The method for identifying mountain sites provided in the embodiments of this application is described below.

[0018] Figure 1 A flowchart illustrating a method for identifying mountain sites according to an embodiment of this application is shown. Figure 1 As shown, the method may include the following steps: S101 to S104.

[0019] S101, acquire regional image information of the target station. The regional image information includes the location coordinates and altitude of the target station, as well as the location coordinates and altitude of multiple preset locations within the area where the target station is located.

[0020] In some embodiments, obtaining regional image information of a target site may include: Obtain regional image information of a region of preset size and preset shape centered on the target site.

[0021] In some embodiments, the regional image information of the target site can be obtained by satellite remote sensing technology or directly extracted from the terrain database of an existing geographic information system. For scenarios with high precision, the location coordinates can also be obtained by GPS positioning, and the altitude can be measured on-site using equipment such as altimeters.

[0022] In some embodiments, obtaining the location coordinates and altitude of multiple preset locations within the area where the target site is located may include: The target site area is evenly divided into a preset number of sub-regions, and the first preset position of each sub-region is selected as the preset position. The location coordinates and altitude of the preset position are obtained. The first preset position can be set according to actual needs, and can be the center point of the sub-region, the highest point of the sub-region, or one or more points randomly selected from the sub-region.

[0023] The target site area is evenly divided into a preset number of sub-regions, and the first preset position in each sub-region is selected as the preset position. This ensures that the selected first preset position can evenly cover the target site area, comprehensively capture the terrain features of different sub-regions, and avoid the omission of some terrain features due to the concentration of samples in a certain local area. This method is suitable for areas with many terrain variations.

[0024] In some embodiments, obtaining the location coordinates and altitude of multiple preset locations within the area where the target site is located may further include: A preset number of sampling points are randomly selected in the area where the target site is located as preset locations, and the location coordinates and altitude of the preset locations are obtained.

[0025] This application embodiment randomly selects sampling points as preset locations, eliminating the need for strict area division, reducing operational complexity, and enabling rapid acquisition of preset location information. It is suitable for areas with no significant terrain changes.

[0026] In one example, when obtaining the preset location of the target site, the location can be selected according to the actual terrain. In areas with relatively simple and regular terrain distribution, the area can be evenly divided into a small number of sub-regions, and the first preset location of each sub-region can be selected as the preset location, or a small number of sampling points can be randomly selected as the preset location.

[0027] In areas with simple and regular terrain, the changes in terrain features are relatively gentle and regular. A small number of evenly selected or randomly selected preset locations can cover the key terrain features in the area, reducing the cost of data acquisition and the computational load of subsequent processing operations, and improving the efficiency of the overall process.

[0028] In one example, when the terrain distribution is complex, the area is evenly divided into a large number of sub-regions, and the first preset position of each sub-region is selected as the preset position, or a large number of sampling points are randomly selected as the preset positions.

[0029] Regions with complex terrain have many undulating and abrupt terrain features. Dividing the region evenly into a large number of sub-regions or selecting a large number of sampling points can provide denser coverage of the area, avoid missing key terrain information due to insufficient sampling points, and ensure a comprehensive reflection of terrain changes within the region.

[0030] S102, subtract the preset reference altitude from the altitude of the target station and the altitude of multiple preset locations respectively to obtain the regional altitude difference of the target station and the regional altitude difference of multiple preset locations.

[0031] Among them, the preset reference altitude is the reference reference value for calculating the height difference in the region, that is, the reference value used to measure the relative height between the target station and the surrounding preset locations. It can be the altitude of the target station, the average altitude of the target station and multiple preset locations, or the lowest altitude of the target station and multiple preset locations; the regional height difference is the relative height of the target station or preset location relative to the preset reference altitude.

[0032] Because the altitude obtained in this application embodiment is a relative altitude relative to sea level, the absolute altitudes relative to sea level in different regions are not comparable. For example, a hill on a plateau may have an altitude of 4000 meters, which is 50 meters higher than the surrounding plain, while a hill on a plain may have an altitude of 200 meters, which is also 50 meters higher than the surrounding plain. The absolute altitude difference between the two is huge, but the degree of undulation in the local terrain is the same. Furthermore, absolute altitudes are affected by global sea level changes or measurement benchmarks, and the data may be inconsistent. If absolute altitudes are used directly, the terrain features of the target site area may be misjudged, masking the actual differences in local terrain elevation.

[0033] Converting the elevation of the target site area into regional elevation differences using a preset benchmark elevation can eliminate the interference of different relative elevations of different regions relative to sea level, focusing on the topographic relief characteristics within the region and more intuitively reflecting the relative elevation relationships of the terrain within the region.

[0034] In some embodiments, the preset reference altitude is the lowest altitude among the altitude of the target station and the altitudes of multiple preset locations. Calculating the regional altitude difference of the target station and the regional altitude difference of the multiple preset locations based on the altitude of the target station, the altitudes of the multiple preset locations, and the preset reference altitude may include: Subtract the preset baseline altitude from the altitude of the target station to obtain the regional altitude difference of the target station; Subtract the preset baseline altitude from the altitude of each preset location to obtain the regional altitude difference between the preset locations.

[0035] When selecting a preset benchmark altitude, the regional average altitude is affected by a few extreme high or low values ​​within the region. However, when using the target station itself as the benchmark, all altitude differences revolve around the target station, only reflecting the local relationship between the target and its surroundings, failing to reflect the overall terrain characteristics of the region. Therefore, selecting the lowest altitude within the region as the preset benchmark altitude, compared to using the average altitude, avoids interference from a few extreme high or low values, ensuring that the benchmark value stably reflects the basic elevation level within the region, making the calculated regional altitude differences more closely match the actual terrain undulations.

[0036] In this application embodiment, the lowest altitude of the target site and multiple preset locations is selected as the preset reference altitude. Compared with the average altitude of the area, the interference of extreme values ​​on the reference value can be eliminated. Compared with the altitude of the target site, it can more comprehensively reflect the overall terrain framework of the area and reflect the characteristics of the target site in the terrain of the area.

[0037] S103, Generate a grayscale image based on the regional height difference of the target site and the regional height difference of multiple preset locations, the location coordinates of the target site and the location coordinates of multiple preset locations.

[0038] In some embodiments, generating a grayscale image based on the regional height difference of the target site and the regional height difference of multiple preset locations, the location coordinates of the target site and the location coordinates of the multiple preset locations, may include: The height difference of the target site and the height difference of the multiple preset locations are converted to grayscale to obtain the grayscale value of the target site and the grayscale value of the multiple preset locations. A grayscale image is generated based on the grayscale value of the target site and the grayscale values ​​of multiple preset locations, as well as the location coordinates of the target site and the location coordinates of multiple preset locations.

[0039] This application's embodiments generate grayscale images, directly reflecting the spatial distribution of elevation differences through brightness variations. Compared to selecting only specific preset locations, which results in discontinuous elevation differences, this method preserves the spatial proximity of points by using continuous grayscale pixels where each pixel has a corresponding grayscale value. This provides crucial features for analyzing the regional terrain of the target site. Furthermore, the grayscale image integrates information from both location coordinates and regional elevation differences into a single-channel image containing only grayscale values, facilitating subsequent model standardization.

[0040] S104. Input the grayscale image into the mountain site recognition model, determine the classification value of the target site based on the model weight parameters of the mountain site recognition model and the grayscale image, and determine the site type of the target site based on the classification value of the target site.

[0041] Among them, the mountain site identification model is a pre-trained classification model that extracts terrain features from grayscale images through model weight parameters.

[0042] In some embodiments, the category value and the site type have a corresponding preset mapping relationship.

[0043] In some embodiments, when the classification value is greater than a preset threshold, the target site can be determined to be a mountain site; when the classification value is not greater than the preset threshold, the target site can be determined to be a non-mountain site.

[0044] By mapping the classification value to the site type, the type of the target site can be directly determined based on the specific data of the classification value. The automated analysis of terrain features by the model replaces the traditional manual review, improving the efficiency and accuracy of identification.

[0045] This embodiment of the application eliminates the interference of absolute poster references in different regions by acquiring regional image information of the target site and calculating the regional elevation difference of the target site, reflecting the relative terrain changes within the region where the target site is located, and improving the accuracy of terrain information acquisition. A grayscale image is generated based on the regional elevation difference and location coordinates, converting geographical information into visual features that the mountain site recognition model can learn. The grayscale image includes the surrounding terrain features of the target site's region. This grayscale image is input into the mountain site recognition model, which can calculate the classification value of the target site based on the surrounding terrain features, improving the comprehensiveness of terrain information extraction. Finally, the target site is classified according to the classification value, thereby improving the accuracy of mountain site recognition.

[0046] In some embodiments, the area height difference of the target site and the area height differences of multiple preset locations are respectively converted to grayscale to obtain the grayscale value of the target site and the grayscale values ​​of the multiple preset locations, which may include: Obtain the maximum and minimum values ​​of the regional height difference of the target site and the regional height differences of multiple preset locations; The height difference between the target site and the height difference between multiple preset locations are normalized based on the maximum and minimum values. The height difference between the target site and the height difference between multiple preset locations are then converted into grayscale values ​​to obtain the grayscale values ​​of the target site and the multiple preset locations.

[0047] In some embodiments, the normalization method may include methods such as linear mapping, nonlinear mapping, logarithmic mapping, and piecewise mapping.

[0048] This application embodiment normalizes all regional height differences based on the maximum and minimum values ​​among the target site and multiple preset locations. It maps regional height differences that may originally be in different numerical ranges to a grayscale value standard range of 0-255, eliminating the difference in physical quantity units and numerical ranges. This avoids the imbalance in grayscale value distribution caused by large differences in the original height difference numerical ranges, and completely preserves the relative size relationship between regional height differences at each location, thereby improving the accuracy of subsequent mountain site identification.

[0049] In some embodiments, generating a grayscale image based on the grayscale value of the target site and the grayscale values ​​of multiple preset locations, the location coordinates of the target site and the location coordinates of the multiple preset locations, may include: The grayscale values ​​of the target site and multiple preset locations are filled into the initial image according to their corresponding location coordinates. The size of the initial image is determined according to the preset size and preset shape of the area where the target site is located. The grayscale image is obtained by interpolating the grayscale values ​​of the unfilled pixels in the initial image and filling them into the corresponding positions in the initial image.

[0050] Since the grayscale values ​​corresponding to the regional height difference between the target site and the preset location are calculated based on a finite number of sampling points, these sampling points are isolated and distributed in the initial image, belonging to discrete data. However, the terrain distribution features in natural scenes usually have spatial continuity, that is, the terrain features of adjacent locations are gradually changing regions. If only the grayscale values ​​of discrete sampling points are retained, a large number of pixels in the initial image will lack grayscale information and cannot fully reflect the continuous change features of the terrain.

[0051] This application embodiment uses an interpolation method to reasonably infer unknown points from known points, transforming discrete data into a continuous image, preventing feature loss due to pixel loss, and providing complete data input for subsequent mountain site recognition models.

[0052] In one example, the target site is located in a square area with sides of 200 meters. This square area is divided into 400 square sub-regions with sides of 10 meters each. The center point of each sub-region is selected as the preset position. The initial image can then be a blank image of 20x20 pixels, and the grayscale value corresponding to the preset position is filled into the blank image. When there are unfilled pixels due to different preset position selection methods, an interpolation method is used to calculate the grayscale value and fill it.

[0053] In some embodiments, the interpolation method may include inverse distance weighted interpolation and Kriging interpolation, etc.

[0054] In one example, such as Figure 2 and Figure 3 As shown, Figure 2 The black marker in the center indicates the target site. The upper left area of ​​the image is mountainous, and the lower right area is flat land. Figure 2 The central area is the area where the target site is located. Figure 2 The grayscale image of the target site area obtained after the above processing is as follows: Figure 3 As shown, mountainous areas appear white in the grayscale image, while flat areas appear black. The target site is located at the gray boundary between the mountainous and flat areas, indicating that its elevation difference lies between the mountainous and flat areas, perfectly matching the actual terrain features of the transition zone. This verifies the effectiveness of the regional elevation difference and grayscale value conversion logic, namely, that the level of grayscale value accurately reflects the relative undulation of the terrain.

[0055] In some embodiments, determining the classification value of a target site based on the model weight parameters of the mountain site identification model and the grayscale image, and determining the site type of the target site based on the classification value of the target site, may include: The grayscale image is converted into a one-dimensional vector, thereby transforming the spatial features of the image into a linear numerical form that the model can directly process, adapting to the input requirements of the mountain site recognition model, and facilitating subsequent mathematical calculations.

[0056] The distance from the target station to the preset hyperplane is calculated based on the model weight parameters and a one-dimensional vector; the preset hyperplane is the classification boundary learned by the mountain station identification model during training. The model weight parameters are the mathematical coefficients of the hyperplane, determining its position and orientation.

[0057] The station type of the target station is determined based on the distance of the target station from the preset hyperplane. In this example, the absolute value reflects the distance between the target station and the classification boundary. The larger the absolute value, the higher the classification confidence. The positive or negative sign of the distance corresponds to the position of the sample relative to the hyperplane, that is, the station type of the target station.

[0058] This embodiment converts grayscale images into one-dimensional vectors and calculates the distance to a preset hyperplane using model weight parameters. This method has low computational complexity and can quickly process large amounts of target site data in batches. Furthermore, the preset hyperplane serves as a clear classification boundary, with the sign of the distance directly corresponding to the site type. The absolute value of the distance intuitively reflects the confidence level of the classification, making the results highly interpretable and facilitating site type determination based on distance. This improves the efficiency of site type identification while ensuring accuracy.

[0059] In some embodiments, converting a grayscale image into a one-dimensional vector may include: Flatten the two-dimensional pixel matrix of a grayscale image into a continuous numerical sequence in row or column order.

[0060] In one example, the grayscale image is a two-dimensional pixel matrix of size 20x20 pixels, which is flattened into a one-dimensional vector of 400 elements in row order.

[0061] In some embodiments, the mountain site identification model may include a support vector machine model, a linear classification model, and a logistic regression model, etc.

[0062] In some embodiments, calculating the distance from the target site to the preset hyperplane based on model weight parameters and a one-dimensional vector may include: Where d is the point ( , The distance from the decision boundary, where A, B, and C are the model weight parameters of the mountain site identification model. and These are the x and y coordinates of the target site, respectively.

[0063] In one example, a positive distance from the target station to the preset hyperplane corresponds to a mountain station, while a negative distance corresponds to a non-mountain station.

[0064] In some embodiments, such as Figure 4 As shown, before inputting the grayscale image into the mountain site identification model and determining the site type of the target site based on the pre-trained model weight parameters of the mountain site identification model and the grayscale image, the method may further include: S401 to S406.

[0065] S401, obtain multiple parameter values, training data and validation data corresponding to multiple model parameters of the initial mountain site identification model.

[0066] The initial mountain site identification model is an untrained basic model framework, and its parameters need to be optimized through training.

[0067] In some embodiments, the model parameters of the mountain site identification model may include a kernel function, a regularization parameter, and a kernel function parameter.

[0068] In some embodiments, the parameter values ​​of the kernel function include linear kernel functions, polynomial kernel functions, radial basis function kernel functions, and sigmoid kernel functions. The parameter values ​​of the regularization parameter include 0.1, 1, 10, 100, and 1000. The parameter values ​​of the kernel function parameter include 0.001, 0.01, 0.1, 1, and 10.

[0069] S402, based on the multiple parameter values ​​corresponding to multiple model parameters, determine all parameter value combinations of the model parameters of the initial mountain site identification model.

[0070] Among them, the parameter value combination is to arrange and combine the candidate values ​​of different model parameters to form a complete parameter configuration.

[0071] In one example, a combination of parameter values ​​could include: a kernel function with a parameter value of a linear kernel function, a regularization parameter with a parameter value of 0.1, and a kernel function parameter with a parameter value of 0.001.

[0072] S403, the training data is input into the initial mountain station identification model corresponding to different parameter value combinations for training, and the first mountain station identification model corresponding to multiple different parameter value combinations is obtained.

[0073] For each combination of parameter values ​​generated in S402, training data is input into the corresponding initial mountain site identification model for training. During training, the model adjusts its internal weights by learning features from the training data, enabling it to initially distinguish site types. Ultimately, each parameter combination corresponds to a trained first mountain site identification model, forming multiple sets of models to be evaluated.

[0074] S404. Input the verification data into the first mountain site identification model corresponding to different parameter value combinations to obtain the identification results of the first mountain site identification model corresponding to different parameter value combinations.

[0075] Using validation data, the accuracy of the actual prediction results of each first mountain site identification model was tested, thereby examining the generalization ability of models with different parameter combinations on unseen data.

[0076] S405 compares the identification results with the labels in the verification data to determine the accuracy of the identification results.

[0077] The labels are the real site types pre-labeled in the verification data, serving as a benchmark for evaluating the accuracy of the model's predictions; the accuracy of the identification results is the proportion of data whose model predictions match the labels to the total data sample.

[0078] In some embodiments, the accuracy of the identification result may include the accuracy of evaluation metrics, F1 score, etc.

[0079] In one example, if the validation data contains 100 samples, and the identification model for a certain first mountain site correctly predicts 95 of them, then its accuracy is 95%.

[0080] The accuracy obtained from the identification results quantifies the performance of different parameter combination models, providing a basis for selecting the optimal identification model for the first mountain site.

[0081] S406, determine the first mountain site identification model corresponding to the combination of different parameter values ​​with the highest accuracy as the mountain site identification model.

[0082] This application embodiment generates all parameter value combinations and trains models with different parameter value combinations using training data. This ensures a comprehensive exploration of the possible parameter configuration space of the model, avoids missing potential optimal solutions, maximizes the possibility of finding a high-performance model, and selects the model with the highest accuracy from all parameter value combinations as the mountain site identification model, thereby improving the identification accuracy of the obtained mountain site identification model.

[0083] The traditional practice of splitting the training and validation sets into single sets results in some data being used only for validation and not for training, leading to wasted samples. This is especially problematic when data volume is limited, as insufficient training data can result in inadequate model learning. Furthermore, training results can be affected by random partitioning factors, such as the partitioning of the training set including a large number of easily identifiable mountain site types or a large number of difficult-to-identify edge samples, leading to inaccurate model accuracy.

[0084] In some embodiments, such as Figure 5 As shown, the training data is input into the initial mountain station identification model corresponding to different parameter value combinations for training, and multiple first mountain station identification models corresponding to different parameter value combinations are obtained, which may include: S501 to S505.

[0085] S501 divides the training data and validation data into a preset number of data subsets.

[0086] In one example, if the total data volume is 782 samples, when using 5-fold cross-validation, it will be divided into 5 data subsets, each containing approximately 156-157 samples.

[0087] This partitioning method, by distributing the data into multiple parts, avoids evaluation bias caused by a single data partition, ensuring that subsequent training and validation processes can cover all data and improving the reliability of model evaluation.

[0088] S502, each data subset is used as the first verification data, and the data of the remaining subset of the data subset corresponding to the first verification data is used as the first training data, thus obtaining multiple first verification data and corresponding first training data.

[0089] The first validation data is a subset of data designated for validation purposes in a certain round of cross-validation; the first training data is the collection of all other data subsets except the first validation data, used for training the model in that round.

[0090] In one example, in 5-fold cross-validation, the first subset is used as the first validation data in the first round, and the remaining four subsets are used as the first training data; in the second round, the second subset is used as the first validation data, and the remaining four subsets are used as the first training data, and so on, generating a total of 5 sets of first training data and first validation data.

[0091] In this way, each subset of data is used once as validation data, ensuring that all samples participate in both training and validation, making full use of the data information.

[0092] S503, the first training data is input into the initial mountain station identification model corresponding to each parameter value combination for training, so as to obtain multiple first mountain station identification models corresponding to each parameter value combination.

[0093] By combining cross-validation and parameter optimization, multiple rounds of training are performed on each parameter combination. This ensures that the model performance of each parameter combination is not affected by a single training data partition, and can more comprehensively reflect its generalization ability.

[0094] S504, where accuracy includes accuracy rate; the accuracy of the identification result is determined by comparing the identification result with the labels in the verification data, including: The recognition results of multiple first mountain site recognition models corresponding to each parameter value combination are compared with the labels in the verification data, and the accuracy of the recognition results of multiple first mountain site recognition models corresponding to each parameter combination is determined based on the comparison results. S505, determine the first mountain site identification model corresponding to the different combinations of parameter values ​​with the highest accuracy, including: The first mountain site identification model corresponding to the parameter value combination with the highest accuracy is determined as the mountain site identification model.

[0095] This application's embodiments utilize cross-validation, dividing the data into multiple subsets that are used alternately as training and validation sets, thus avoiding insufficient model learning caused by idle data. Furthermore, each data point participates in the training process, reducing the interference of random data partitioning on the evaluation results and improving the reliability of model accuracy calculations. Simultaneously, training on multiple subsets strengthens the model's generalization ability, reduces the risk of overfitting, and ultimately improves the recognition accuracy of the final determined mountain site identification model.

[0096] In some embodiments, determining the first mountain site identification model corresponding to the parameter value combination with the highest accuracy as the mountain site identification model may include: Calculate the average recognition accuracy of multiple first mountain site recognition models corresponding to each parameter value combination; The parameter value combination with the highest average recognition accuracy is used as the parameter value of the pre-trained mountain site recognition model. The weight parameters of multiple first mountain site recognition models corresponding to the parameter value combination with the highest average recognition accuracy are used as the model weight parameters of the mountain site recognition model, thus obtaining the pre-trained mountain site recognition model.

[0097] Each parameter value combination corresponds to multiple models trained on different data subsets. In this embodiment, the recognition accuracy of the model corresponding to each parameter value combination is determined by calculating the average recognition accuracy of these models. This can more realistically reflect the average performance of the model for each parameter value combination, reduce random fluctuations caused by a single data partitioning, and improve the reliability of the mountain site recognition model determination.

[0098] In one example, this application embodiment selects 942 sites from the intelligent planning system and manually labels them, including 446 mountain sites and 496 non-mountain sites. The rule for determining mountain sites is that the site belongs to a mountainous area and there are points in the surrounding area with an altitude difference of more than 10 meters.

[0099] Based on the latitude and longitude of the target station, a square area with sides of 200 meters is obtained, centered on the station. Latitude and longitude sampling is performed within this 200m x 200m square area, resulting in 1024 x 1024 sampling points. The altitude of each sampling point is obtained, and the altitude of the lowest point in the area is subtracted from the altitude of each point to obtain the regional altitude difference. Finally, the regional altitude differences are visualized using Python and Matplotlib libraries to generate a visualization image. The original image is input to generate a grayscale image of the altitude difference for each station, where white represents areas with larger altitude differences and black represents areas with smaller altitude differences.

[0100] From 942 manually selected sites, 160 sites were randomly chosen as the test set, including 80 sites on mountains and 80 sites not on mountains. The remaining 782 sites were used as the training set, and a support vector machine model was selected as the initial site recognition model.

[0101] When training the initial mountain site recognition model, the image of each site was first converted into a grayscale image, and the generated grayscale image was read using OpenCV's `imread` function. The 1024 * 1024 grayscale image was scaled down to 100 * 100, and the `flatten` method was used to flatten the two-dimensional image into a one-dimensional vector, which served as the input features for the initial mountain site recognition model. Finally, 5-fold cross-validation was used, dividing the site dataset into 5 site subsets. Each time, one site subset was used as the validation set, and the remaining 4 site subsets were used as the training set. This process was repeated 5 times, each time selecting a different site subset as the validation set. Finally, the average result of the 5 validations was calculated as the model's performance evaluation metric.

[0102] Finally, training samples are constructed by combining image data from both mountain and non-mountain sites. The model is then trained on these samples, and the weight vectors are adjusted to find the optimal decision boundary. After training, the optimized SVM model can determine whether a new site is a mountain site based on the elevation difference features around the site. The trained model is then validated on a test set, and the predicted results are compared with the labels to obtain the validation results. Figure 6 As shown in the confusion matrix, Figure 6 The verification labels were manually assigned, with 1 representing mountains and 0 representing non-mountains. The prediction labels were model predictions, with 1 representing mountains and 0 representing non-mountains. The verification results included 160 test sites, with 156 correctly identified and 4 incorrectly identified, for an accuracy of 97.50%. Among the 80 mountain sites, 79 were correctly identified and 1 was incorrectly identified, for an accuracy of 98.75%. Among the 80 non-mountain sites, 77 were correctly identified and 3 were incorrectly identified, for an accuracy of 96.25%.

[0103] Figure 7 This application illustrates a device 700 for identifying mountain sites, which may include: The acquisition module 701 is used to acquire regional image information of the target station. The regional image information includes the location coordinates and altitude of the target station, as well as the location coordinates and altitude of multiple preset locations within the area where the target station is located. The calculation module 702 is used to calculate the regional height difference of the target station and the regional height difference of multiple preset locations based on the altitude of the target station, the altitude of multiple preset locations and the preset reference altitude. The generation module 703 is used to generate a grayscale image based on the regional height difference of the target site and the regional height difference of multiple preset locations, the location coordinates of the target site and the location coordinates of multiple preset locations; The determination module 704 is used to input the grayscale image into the mountain site recognition model, determine the classification value of the target site based on the model weight parameters of the mountain site recognition model and the grayscale image, and determine the site type of the target site based on the classification value of the target site.

[0104] In some embodiments, the calculation module 702 is further configured to subtract a preset reference altitude from the altitude of the target station to obtain the regional altitude difference of the target station; The calculation module 702 is also used to subtract the preset reference altitude from the altitude of multiple preset locations to obtain the regional altitude difference of multiple preset locations.

[0105] In some embodiments, the mountain site identification device 700 may further include: The conversion module is used to perform grayscale conversion on the regional height difference of the target site and the regional height difference of multiple preset locations respectively, so as to obtain the grayscale value of the target site and the grayscale value of multiple preset locations. The generation module 703 is also used to generate a grayscale image based on the grayscale value of the target site and the grayscale values ​​of multiple preset locations, the location coordinates of the target site and the location coordinates of multiple preset locations.

[0106] In some embodiments, the conversion module is further configured to convert a grayscale image into a one-dimensional vector; The calculation module 702 is also used to calculate the distance from the target site to the preset hyperplane based on the model weight parameters and the one-dimensional vector; The determination module 704 determines the station type of the target station based on the distance of the target station from the preset hyperplane.

[0107] In some embodiments, the mountain site identification device 700 may further include: The acquisition module 701 is also used to acquire multiple parameter values, training data and validation data corresponding to multiple model parameters of the initial mountain site identification model; The determination module 704 is also used to determine all parameter value combinations of the model parameters of the initial mountain site identification model based on the multiple parameter values ​​corresponding to multiple model parameters; The training module is used to train the initial mountain station identification model corresponding to different combinations of parameter values ​​by inputting training data into the model, and to obtain the first mountain station identification model corresponding to multiple different combinations of parameter values. The determination module 704 is also used to input the verification data into the first mountain site identification model corresponding to different combinations of parameter values, and obtain the identification results of the first mountain site identification model corresponding to different combinations of parameter values; The determination module 704 is also used to compare the recognition result with the label in the verification data to determine the accuracy of the recognition result; The determination module 704 is also used to determine the first mountain site identification model corresponding to the different combinations of parameter values ​​with the highest accuracy as the mountain site identification model.

[0108] In some embodiments, the mountain site identification device 700 may further include: The partitioning module is used to divide the training data and validation data into a preset number of data subsets; The partitioning module is also used to use each data subset as the first verification data, and the data of the remaining subset of the data subset corresponding to the first verification data as the first training data, so as to obtain multiple first verification data and corresponding first training data. The training module is also used to input the first training data into the initial mountain station identification model corresponding to each parameter value combination for training, so as to obtain multiple first mountain station identification models corresponding to each parameter value combination. The comparison module is used to compare the recognition results of multiple first mountain site recognition models corresponding to each parameter value combination with the labels in the verification data, and determine the accuracy of the recognition results of multiple first mountain site recognition models corresponding to each parameter combination based on the comparison results. The determination module 704 is also used to determine the first mountain site identification model corresponding to the parameter value combination with the highest accuracy as the mountain site identification model.

[0109] In some embodiments, the calculation module 702 is further configured to calculate the average recognition accuracy of the recognition accuracy of the multiple first mountain site recognition models corresponding to each combination of parameter values. The determination module 704 is further used to take the combination of parameter values ​​with the highest average recognition accuracy as the parameter values ​​of the pre-trained mountain site recognition model, and take the weight parameters of multiple first mountain site recognition models corresponding to the combination of parameter values ​​with the highest average recognition accuracy as the model weight parameters of the mountain site recognition model, so as to obtain the pre-trained mountain site recognition model.

[0110] Figure 7 The various modules in the illustrated device can achieve Figure 1 The various steps involved, and the corresponding technical effects achieved, will not be elaborated upon here for the sake of brevity.

[0111] Figure 8 A schematic diagram of the hardware structure of the terminal device provided in an embodiment of this application is shown.

[0112] The terminal device may include a processor 801 and a memory 802 storing computer program instructions.

[0113] Specifically, the processor 801 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0114] Memory 802 may include mass storage for data or instructions. For example, and not limitingly, memory 802 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. In one instance, memory 802 may include removable or non-removable (or fixed) media, or memory 802 may be non-volatile solid-state memory. Memory 802 may be internal or external to the integrated gateway disaster recovery device.

[0115] In one instance, memory 802 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Thus, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method for identifying mountain sites according to this disclosure.

[0116] The processor 801 reads and executes computer program instructions stored in the memory 802 to achieve... Figure 1 The method for identifying mountain sites in the illustrated embodiment.

[0117] In one example, the terminal device may also include a communication interface 803 and a bus 804. Wherein, for example... Figure 8 As shown, the processor 801, memory 802, and communication interface 803 are connected through bus 804 and complete communication with each other.

[0118] The communication interface 803 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0119] Bus 804 includes hardware, software, or both, that couples components of an end device together. For example, and not as a limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 804 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0120] Furthermore, in conjunction with the mountain site identification method in the above embodiments, this application embodiment can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the mountain site identification methods in the above embodiments.

[0121] This application also provides a computer program product, including a computer program that, when executed, implements any of the mountain site identification methods described in the above embodiments.

[0122] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0123] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or text segments used to perform the required tasks. Programs or text segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, read-only memory (ROM), flash memory, erasable read-only memory (EROM), floppy disks, compact disc read-only memory (CD-ROM), optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Text segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0124] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0125] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0126] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for identifying mountain monitoring stations, characterized in that, include: Acquire regional image information of the target site, including the location coordinates and altitude of the target site, as well as the location coordinates and altitude of multiple preset locations within the area where the target site is located; Subtract the preset reference altitude from the altitude of the target station and the altitude of the multiple preset locations respectively to obtain the regional altitude difference of the target station and the regional altitude difference of the multiple preset locations; A grayscale image is generated based on the regional height difference of the target site and the regional height difference of the multiple preset locations, the location coordinates of the target site and the location coordinates of the multiple preset locations; The grayscale image is input into the mountain site recognition model. The classification value of the target site is determined based on the model weight parameters of the mountain site recognition model and the grayscale image. The site type of the target site is determined based on the classification value of the target site.

2. The method for identifying mountain sites according to claim 1, characterized in that, The preset reference altitude is the lowest altitude among the altitude of the target station and the altitudes of the multiple preset locations.

3. The method for identifying mountain sites according to claim 1, characterized in that, The step of generating a grayscale image based on the regional height difference of the target station and the regional height difference of the multiple preset locations, the location coordinates of the target station and the location coordinates of the multiple preset locations includes: The area height difference of the target site and the area height difference of the multiple preset locations are converted to grayscale to obtain the grayscale value of the target site and the grayscale value of the multiple preset locations. A grayscale image is generated based on the grayscale value of the target site and the grayscale values ​​of multiple preset locations, as well as the location coordinates of the target site and the location coordinates of the multiple preset locations.

4. The method for identifying mountain sites according to claim 1, characterized in that, The classification value of the target site is determined based on the model weight parameters of the mountain site identification model and the grayscale image, and the site type of the target site is determined based on the classification value of the target site, including: Convert the grayscale image into a one-dimensional vector; The distance from the target station to the preset hyperplane is calculated based on the model weight parameters and the one-dimensional vector. The station type of the target station is determined based on the distance of the target station from the preset hyperplane.

5. The method for identifying mountain sites according to claim 1, characterized in that, Before inputting the grayscale image into the mountain site identification model and determining the site type of the target site based on the pre-trained model weight parameters of the mountain site identification model and the grayscale image, the method further includes: Obtain multiple parameter values, training data, and validation data corresponding to multiple model parameters of the initial mountain site identification model; Based on the multiple parameter values ​​corresponding to the multiple model parameters, determine all parameter value combinations of the initial mountain site identification model; The training data is input into the initial mountain station identification model corresponding to different parameter value combinations for training, and multiple first mountain station identification models corresponding to different parameter value combinations are obtained. The verification data is input into the first mountain site identification model corresponding to different parameter value combinations to obtain the identification results of the first mountain site identification model corresponding to different parameter value combinations. The recognition results are compared with the labels in the verification data to determine the accuracy of the recognition results; The first mountain site identification model corresponding to the combination of different parameter values ​​with the highest accuracy is determined as the mountain site identification model.

6. The method for identifying mountain sites according to claim 5, characterized in that, The process involves inputting the training data into an initial mountain station identification model corresponding to different parameter value combinations for training, thereby obtaining multiple first mountain station identification models corresponding to different parameter value combinations, including: Divide the training data and validation data into a predetermined number of data subsets; Each data subset is used as the first validation data, and the data of the remaining subset of the data subset corresponding to the first validation data is used as the first training data, thus obtaining multiple first validation data and corresponding first training data; The first training data is input into the initial mountain station identification model corresponding to each parameter value combination for training, so as to obtain multiple first mountain station identification models corresponding to each parameter value combination. The accuracy includes accuracy rate; the step of comparing the recognition result with the labels in the verification data to determine the accuracy of the recognition result includes: The recognition results of multiple first mountain site recognition models corresponding to each parameter value combination are compared with the labels in the verification data, and the accuracy of the recognition results of multiple first mountain site recognition models corresponding to each parameter combination is determined based on the comparison results. Determining the first mountain site identification model corresponding to the different combinations of parameter values ​​with the highest accuracy includes: The first mountain site identification model corresponding to the parameter value combination with the highest accuracy is determined as the mountain site identification model.

7. The method for identifying mountain sites according to claim 6, characterized in that, The process of determining the first mountain site identification model corresponding to the parameter value combination with the highest accuracy as the mountain site identification model includes: Calculate the average recognition accuracy of multiple first mountain site recognition models corresponding to each parameter value combination; The parameter values ​​with the highest average recognition accuracy are used as the parameter values ​​of the pre-trained mountain site recognition model. The weight parameters of the multiple first mountain site recognition models corresponding to the parameter values ​​with the highest average recognition accuracy are used as the model weight parameters of the mountain site recognition model, thus obtaining the pre-trained mountain site recognition model.

8. A device for identifying mountain sites, characterized in that, The device includes: The acquisition module is used to acquire regional image information of the target site. The regional image information includes the location coordinates and altitude of the target site, as well as the location coordinates and altitude of multiple preset locations within the area where the target site is located. The calculation module is used to subtract a preset reference altitude from the altitude of the target station and the altitude of the multiple preset locations respectively to obtain the regional altitude difference of the target station and the regional altitude difference of the multiple preset locations. The generation module is used to generate a grayscale image based on the regional height difference of the target site and the regional height difference of the multiple preset locations, the location coordinates of the target site and the location coordinates of the multiple preset locations; The determination module is used to input the grayscale image into the mountain site recognition model, determine the classification value of the target site based on the model weight parameters of the mountain site recognition model and the grayscale image, and determine the site type of the target site based on the classification value of the target site.

9. A terminal device, characterized in that, The device includes: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, it implements the method for identifying mountain sites as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement the method for identifying mountain sites as described in any one of claims 1-7.

11. A computer program product, characterized in that, When the instructions in the computer program product are executed by the processor of the electronic device, the electronic device performs the method for identifying mountain sites as described in any one of claims 1-7.

Images