A path loss prediction method, device and electronic equipment

Abstract

The application provides a road loss prediction method and device and electronic equipment, and relates to the technical field of wireless communication. The method comprises the following steps: obtaining a first image and a second image according to map information of a sampling area; obtaining a third image according to road test data of the sampling area; the first image and the third image both comprise a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver; training a road loss prediction model according to a training image synthesized according to the first image, the second image and the third image and structured data of transmitters in the sampling area; inputting the road test data, a target image and structured data of a target transmitter into the trained road loss prediction model to output a road loss value of a target receiver corresponding to the target transmitter, wherein the target image is determined according to the map information, and the target image comprises the target receiver and the target transmitter. The scheme has high information utilization rate, can provide accurate prediction results and improves the prediction efficiency.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a path loss prediction method, apparatus, and electronic device. Background Technology

[0002] Existing road damage prediction methods primarily rely on formulaic models based on expert experience, failing to establish complex models for road damage prediction. Furthermore, they can only calculate based on a subset of pre-selected features, exhibiting limitations and low information utilization. Additionally, existing formulaic models suffer from simplistic calculation formulas that cannot be adjusted for different geographical and transmission / reception environments, resulting in low prediction accuracy and efficiency. Summary of the Invention

[0003] The purpose of this invention is to provide a road damage prediction method, device, and electronic device to solve the problems of limitations, low information utilization, and low prediction efficiency in existing road damage prediction technologies.

[0004] To address the aforementioned technical problems, the embodiments of the present invention provide the following technical solutions:

[0005] A road loss prediction method includes:

[0006] Based on the map information of the sampling area, a first image and a second image are obtained. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver.

[0007] Based on the road test data of the sampling area, a third image is obtained, which includes a receiver and a transmitter corresponding to the receiver.

[0008] A path loss prediction model is trained based on the training image synthesized from the first image, the second image, and the third image, as well as the structured data of the transmitter within the sampling area.

[0009] The road test data, target image, and structured data of the target transmitter are input into the trained road loss prediction model, and the road loss value of the target receiver corresponding to the target transmitter is output. The target image is determined based on the map information, and the target image includes the target receiver and the target transmitter.

[0010] Optionally, in the road loss prediction method, before obtaining the first image and the second image based on the map information of the sampling area, the method further includes:

[0011] Match the receivers and transmitters within the sampling area to obtain the transmitters corresponding to the receivers.

[0012] Optionally, in the road loss prediction method, the step of matching receivers and transmitters within the sampling area to obtain the transmitter corresponding to the receiver includes:

[0013] Obtain the Euclidean distance between each receiver and each transmitter within the sampling area;

[0014] Match the receiver with the transmitter corresponding to the minimum Euclidean distance to obtain the transmitter corresponding to the receiver.

[0015] Optionally, in the aforementioned road loss prediction method, the structured data includes geographical location, altitude, downtilt angle, frequency, power, azimuth angle of the target cell, and distance between the transmitter and receiver;

[0016] The orientation angle of the target cell is determined according to the following steps:

[0017] Obtain the angle difference between the line connecting the receiver and the corresponding transmitter and the north direction, and the azimuth angle of each cell covered by the transmitter;

[0018] The azimuth angle of the cell corresponding to the smallest angle difference is taken as the azimuth angle of the target cell.

[0019] Optionally, in the road loss prediction method, obtaining the first image and the second image based on the map information of the sampling area includes:

[0020] Construct a grayscale image of each feature in the map information;

[0021] In each of the grayscale images of the layer, the first image and the second image are cropped.

[0022] Optionally, in the road loss prediction method, constructing a grayscale image of each feature layer in the map information includes:

[0023] Create a layer for each feature;

[0024] In the layer, the boundaries of the features are marked based on their latitude and longitude information;

[0025] Obtain the grayscale value of each feature;

[0026] Based on the boundary and the grayscale value, the layer is converted into a grayscale image of the layer.

[0027] Optionally, in the road loss prediction method, obtaining the grayscale value of each land feature includes:

[0028] Get the height difference value for each layer;

[0029] If the height difference value of a layer is less than the first preset value, set the gray value of the ground features in the layer to the first gray value.

[0030] If the height difference value of the layer is greater than or equal to the first preset value, the height value of the layer is normalized, and the gray value of the ground features in the layer is set to the height value after the layer normalization.

[0031] Optionally, in the road loss prediction method, the step of cropping the first image and the second image from each layer of grayscale image includes:

[0032] In each of the grayscale images of the layer, a first rectangular image and a square image are extracted. The first rectangular image includes a receiver and a transmitter corresponding to the receiver. The square image is centered on the receiver.

[0033] The first rectangular image is scaled to the size of the first image, and the square image is scaled to the size of the second image, which is the same as the preset image size.

[0034] Optionally, in the road loss prediction method, obtaining a third image based on road test data from the sampled area includes:

[0035] Based on the map information, a raster image of the sampling area is constructed;

[0036] The road loss value corresponding to each latitude and longitude information in the road test data is filled into the grid cell corresponding to the latitude and longitude information in the grid image to obtain the road loss grid image;

[0037] The road loss values ​​of the road loss raster image are normalized to obtain a road loss grayscale image;

[0038] The third image is extracted from the grayscale image of road loss.

[0039] Optionally, in the road loss prediction method, the step of constructing a raster image of the sampling area based on the map information includes:

[0040] Convert the latitude and longitude information in the map information into coordinate information in meters;

[0041] Based on the coordinate information, establish a two-dimensional rectangular coordinate system;

[0042] The grid image is constructed by dividing the two-dimensional Cartesian coordinate system into multiple grids.

[0043] Optionally, in the road loss prediction method, the step of filling the road loss value corresponding to each latitude and longitude information in the road test data into the raster image corresponding to the latitude and longitude information to obtain a road loss raster image includes:

[0044] Determine the grid position coordinates of each latitude and longitude information in the grid image;

[0045] Based on the grid position coordinates, determine the grid number of the grid where each latitude and longitude information is located;

[0046] According to the grid number, the road loss value corresponding to each latitude and longitude information is filled into the corresponding grid.

[0047] The path loss raster image is obtained by averaging at least one path loss value corresponding to each raster.

[0048] Optionally, in the road loss prediction method, extracting the third image from the road loss grayscale image includes:

[0049] In the grayscale image of the road loss, a second rectangular image is extracted, which includes a receiver and a transmitter corresponding to the receiver;

[0050] The second rectangular image is scaled to the third image, which has the same size as the preset image.

[0051] Optionally, in the road loss prediction method, training the road loss prediction model based on the training image synthesized from the first image, the second image, and the third image, and the structured data of the transmitter within the sampling area, includes:

[0052] The training images are sequentially input into the convolutional and pooling layers of the road loss prediction model to extract features and obtain the first feature data;

[0053] The structured data is normalized to obtain the second feature data;

[0054] The first feature data and the second feature data are input into the fully connected layer of the road loss prediction model for training.

[0055] This invention also provides an electronic device, including a processor and a transceiver, wherein:

[0056] The processor is used to obtain a first image and a second image based on map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver.

[0057] The processor is further configured to obtain a third image based on the road test data of the sampling area, the third image including a receiver and a transmitter corresponding to the receiver;

[0058] The processor is further configured to train a path loss prediction model based on a training image synthesized from the first image, the second image, and the third image, and structured data of the transmitter within the sampling area.

[0059] The processor is further configured to input the road test data, target image, and structured data of the target transmitter into the trained road loss prediction model, and output the road loss value of the target receiver corresponding to the target transmitter. The target image is determined based on the map information, and the target image includes the target receiver and the target transmitter.

[0060] Optionally, in the electronic device, the processor is further configured to:

[0061] Match the receivers and transmitters within the sampling area to obtain the transmitters corresponding to the receivers.

[0062] Optionally, in the electronic device, the processor is specifically used for:

[0063] Obtain the Euclidean distance between each receiver and each transmitter within the sampling area;

[0064] Match the receiver with the transmitter corresponding to the minimum Euclidean distance to obtain the transmitter corresponding to the receiver.

[0065] Optionally, in the electronic device, the structured data includes geographic location, altitude, downtilt angle, frequency, power, azimuth angle of the target cell, and distance between the transmitter and receiver;

[0066] The orientation angle of the target cell is determined according to the following steps:

[0067] Obtain the angle difference between the line connecting the receiver and the corresponding transmitter and the north direction, and the azimuth angle of each cell covered by the transmitter;

[0068] The azimuth angle of the cell corresponding to the smallest angle difference is taken as the azimuth angle of the target cell.

[0069] Optionally, in the electronic device, the processor is specifically used for:

[0070] Construct a grayscale image of each feature in the map information;

[0071] In each of the grayscale images of the layer, the first image and the second image are cropped.

[0072] Optionally, in the electronic device, the processor is specifically used for:

[0073] Create a layer for each feature;

[0074] In the layer, the boundaries of the features are marked based on their latitude and longitude information;

[0075] Obtain the grayscale value of each feature;

[0076] Based on the boundary and the grayscale value, the layer is converted into a grayscale image of the layer.

[0077] Optionally, in the electronic device, the processor is specifically used for:

[0078] Get the height difference value for each layer;

[0079] If the height difference value of a layer is less than the first preset value, set the gray value of the ground features in the layer to the first gray value.

[0080] If the height difference value of the layer is greater than or equal to the first preset value, the height value of the layer is normalized, and the gray value of the ground features in the layer is set to the height value after the layer normalization.

[0081] Optionally, in the electronic device, the processor is specifically used for:

[0082] In each of the grayscale images of the layer, a first rectangular image and a square image are extracted. The first rectangular image includes a receiver and a transmitter corresponding to the receiver. The square image is centered on the receiver.

[0083] The first rectangular image is scaled to the size of the first image, and the square image is scaled to the size of the second image, which is the same as the preset image size.

[0084] Optionally, in the electronic device, the processor is specifically used for:

[0085] Based on the map information, a raster image of the sampling area is constructed;

[0086] The road loss value corresponding to each latitude and longitude information in the road test data is filled into the grid cell corresponding to the latitude and longitude information in the grid image to obtain the road loss grid image;

[0087] The road loss values ​​of the road loss raster image are normalized to obtain a road loss grayscale image;

[0088] The third image is extracted from the grayscale image of road loss.

[0089] Optionally, in the electronic device, the processor is specifically used for:

[0090] Convert the latitude and longitude information in the map information into coordinate information in meters;

[0091] Based on the coordinate information, establish a two-dimensional rectangular coordinate system;

[0092] The grid image is constructed by dividing the two-dimensional rectangular coordinate system into multiple grids.

[0093] Optionally, in the electronic device, the processor is specifically used for:

[0094] Determine the grid position coordinates of each latitude and longitude information in the grid image;

[0095] Based on the grid position coordinates, determine the grid number of the grid where each latitude and longitude information is located;

[0096] According to the grid number, the road loss value corresponding to each latitude and longitude information is filled into the corresponding grid.

[0097] The path loss raster image is obtained by averaging at least one path loss value corresponding to each raster.

[0098] Optionally, in the electronic device, the processor is specifically used for:

[0099] In the grayscale image of the road loss, a second rectangular image is extracted, which includes a receiver and a transmitter corresponding to the receiver;

[0100] The second rectangular image is scaled to the third image, which has the same size as the preset image.

[0101] Optionally, in the electronic device, the processor is specifically used for:

[0102] The training images are sequentially input into the convolutional and pooling layers of the road loss prediction model to extract features and obtain the first feature data;

[0103] The structured data is normalized to obtain the second feature data;

[0104] The first feature data and the second feature data are input into the fully connected layer of the road loss prediction model for training.

[0105] This invention also provides a road loss prediction device, comprising:

[0106] The first acquisition module is used to acquire a first image and a second image based on the map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver.

[0107] The second acquisition module is used to obtain a third image based on the road test data of the sampling area, the third image including a receiver and a transmitter corresponding to the receiver;

[0108] The first training module is used to train a path loss prediction model based on a training image synthesized from the first image, the second image, and the third image, as well as structured data of the transmitter within the sampling area.

[0109] The first output module is used to input the road test data, target image, and structured data of the target transmitter into the trained road loss prediction model, and output the road loss value of the target receiver corresponding to the target transmitter. The target image is determined based on the map information, and the target image includes the target receiver and the target transmitter.

[0110] Optionally, the road loss prediction device further includes:

[0111] The first acquisition module is used to match the receivers and transmitters in the sampling area to acquire the transmitters corresponding to the receivers.

[0112] Optionally, in the road loss prediction device, the first acquisition module is specifically used for:

[0113] Obtain the Euclidean distance between each receiver and each transmitter within the sampling area;

[0114] Match the receiver with the transmitter corresponding to the minimum Euclidean distance to obtain the transmitter corresponding to the receiver.

[0115] Optionally, in the aforementioned road loss prediction device, the structured data includes geographical location, altitude, downtilt angle, frequency, power, azimuth angle of the target cell, and distance between the transmitter and receiver;

[0116] The orientation angle of the target cell is determined according to the following modules:

[0117] The second acquisition module is used to acquire the angle difference between the line connecting the receiver and the corresponding transmitter and the north direction and the azimuth angle of each cell covered by the transmitter.

[0118] The third acquisition module is used to take the azimuth angle of the cell corresponding to the smallest angle difference as the azimuth angle of the target cell.

[0119] Optionally, in the road loss prediction device, the first obtaining module includes:

[0120] The first construction unit is used to construct a layer grayscale image of each feature in the map information;

[0121] The first cropping unit is used to crop the first image and the second image in each of the grayscale images of the layer.

[0122] Optionally, in the road loss prediction device, the first building unit includes:

[0123] Create sub-cells to create a layer for each feature;

[0124] The marker sub-unit is used to mark the boundaries of the ground features in the layer based on their latitude and longitude information;

[0125] Obtain sub-units to acquire the grayscale value of each ground feature;

[0126] The conversion subunit is used to convert the layer into a grayscale image of the layer based on the boundary and the grayscale value.

[0127] Optionally, in the road loss prediction device, the acquisition subunit is specifically used for:

[0128] Get the height difference value for each layer;

[0129] If the height difference value of a layer is less than the first preset value, set the gray value of the ground features in the layer to the first gray value.

[0130] If the height difference value of the layer is greater than or equal to the first preset value, the height value of the layer is normalized, and the gray value of the ground features in the layer is set to the height value after the layer normalization.

[0131] Optionally, in the road loss prediction device, the first interception unit is used for:

[0132] In each of the grayscale images of the layer, a first rectangular image and a square image are extracted. The first rectangular image includes a receiver and a transmitter corresponding to the receiver. The square image is centered on the receiver.

[0133] The first rectangular image is scaled to the size of the first image, and the square image is scaled to the size of the second image, which is the same as the preset image size.

[0134] Optionally, in the road loss prediction device, the second obtaining module includes:

[0135] The second construction unit is used to construct a raster image of the sampling area based on the map information;

[0136] The first obtaining unit is used to fill the road loss value corresponding to each latitude and longitude information in the road test data into the grid corresponding to the latitude and longitude information in the grid image to obtain the road loss grid image.

[0137] The second obtaining unit is used to normalize the road loss value of the road loss raster image to obtain a road loss grayscale image.

[0138] The second cropping unit is used to crop the third image from the grayscale image of road loss.

[0139] Optionally, in the road loss prediction device, the second building unit is specifically used for:

[0140] Convert the latitude and longitude information in the map information into coordinate information in meters;

[0141] Based on the coordinate information, establish a two-dimensional rectangular coordinate system;

[0142] The grid image is constructed by dividing the two-dimensional rectangular coordinate system into multiple grids.

[0143] Optionally, in the road loss prediction device, the first obtaining unit is specifically used for:

[0144] Determine the grid position coordinates of each latitude and longitude information in the grid image;

[0145] Based on the grid position coordinates, determine the grid number of the grid where each latitude and longitude information is located;

[0146] According to the grid number, the road loss value corresponding to each latitude and longitude information is filled into the corresponding grid.

[0147] The path loss raster image is obtained by averaging at least one path loss value corresponding to each raster.

[0148] Optionally, in the road loss prediction device, the second interception unit is specifically used for:

[0149] In the grayscale image of the road loss, a second rectangular image is extracted, which includes a receiver and a transmitter corresponding to the receiver;

[0150] The second rectangular image is scaled to the third image, which has the same size as the preset image.

[0151] Optionally, in the road loss prediction device, the first training module is specifically used for:

[0152] The training images are sequentially input into the convolutional and pooling layers of the road loss prediction model to extract features and obtain the first feature data;

[0153] The structured data is normalized to obtain the second feature data;

[0154] The first feature data and the second feature data are input into the fully connected layer of the road loss prediction model for training.

[0155] This invention also provides an electronic device, including: a transceiver, a processor, a memory, and a program or instructions stored in the memory and executable on the processor; when the processor executes the program or instructions, it implements the road loss prediction method as described in any of the preceding claims.

[0156] This invention also provides a readable storage medium storing a program or instructions that, when executed by a processor, implement the road loss prediction method as described in any of the preceding claims.

[0157] The beneficial effects of the above-described technical solution of the present invention are as follows:

[0158] The present invention first obtains a first image and a second image based on map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver. Then, a third image is obtained based on road test data of the sampling area. The third image includes the receiver and a transmitter corresponding to the receiver. A road loss prediction model is trained based on a training image synthesized from the first image, the second image, and the third image, as well as structured data of the transmitters in the sampling area. Finally, the road test data, the target image, and the structured data of the target transmitter are input into the trained road loss prediction model, and the road loss value of the target receiver corresponding to the target transmitter is output. The target image is determined based on the map information and includes the target receiver and the target transmitter. This invention solves the limitations of road loss prediction by utilizing map information, road test data, and structured data, resulting in high information utilization, more accurate prediction results, higher efficiency, and better universality. Attached Figure Description

[0159] Figure 1 This is one of the flowcharts for the road loss prediction method provided in the embodiments of the present invention;

[0160] Figure 2 A grayscale image of a road layer provided in an embodiment of the present invention;

[0161] Figure 3 A grayscale image of a building layer provided in an embodiment of the present invention;

[0162] Figure 4 The grayscale images of the labeled receiver and transmitter provided in the embodiments of the present invention;

[0163] Figure 5 A first rectangular image provided for an embodiment of the present invention;

[0164] Figure 6 This is the second flowchart of the road loss prediction method provided in the embodiments of the present invention;

[0165] Figure 7 This is one of the structural schematic diagrams of the electronic device provided in the embodiments of the present invention;

[0166] Figure 8 This is a schematic diagram of the road loss prediction device provided in an embodiment of the present invention;

[0167] Figure 9 This is a second schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0168] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0169] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0170] In various embodiments of the present invention, it should be understood that the sequence number of each process described below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0171] In addition, the terms "system" and "network" are often used interchangeably in this article.

[0172] In the embodiments provided in this application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0173] To address the limitations, low information utilization, and low prediction efficiency in road loss prediction, this invention provides a road loss prediction method. First, a first image and a second image are obtained based on map information of a sampling area. The first image includes a receiver and a corresponding transmitter, while the second image centers on the receiver. Then, a third image is obtained based on road test data from the sampling area. This third image also includes the receiver and a corresponding transmitter. A road loss prediction model is trained using a training image synthesized from the first, second, and third images, along with structured data of the transmitters within the sampling area. Finally, the road test data, the target image, and the structured data of the target transmitter are input into the trained road loss prediction model, outputting the road loss value of the target receiver corresponding to the target transmitter. The target image is determined based on the map information and includes both the target receiver and the target transmitter. This method addresses the limitations of road loss prediction by utilizing map information, road test data, and structured data, resulting in high information utilization, more accurate predictions, higher efficiency, and better universality.

[0174] like Figure 1 As shown, an embodiment of the present invention provides a road loss prediction method, comprising:

[0175] Step S101: Obtain a first image and a second image based on the map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver.

[0176] In this embodiment of the invention, the first image is defined by a diagonal line connecting the receiver and the corresponding transmitter, and the second image is defined by a predetermined length centered on the receiver. Both the first and second images include the same receiver. The number of the first and second images is equal and depends on the number of receivers within the sampling area, specifically at least one.

[0177] It should be noted that the first image is extracted from the map information, rather than manually filtering building information, which takes into account geographical information more comprehensively and has better universality.

[0178] Step S102: Obtain a third image based on the road test data of the sampling area. The third image includes a receiver and a transmitter corresponding to the receiver.

[0179] In this embodiment of the invention, the third image is defined by a diagonal line connecting the receiver and the corresponding transmitter, and the receiver and transmitter included in the third image are the same as those included in the first image. The number of the first image, the second image, and the third image is equal, and is related to the number of receivers in the sampling area, specifically at least one.

[0180] It should be noted that the road test data includes the latitude and longitude information and road loss value of the receiver within the sampling area.

[0181] Step S103: Train the path loss prediction model based on the training image synthesized from the first image, the second image, and the third image, as well as the structured data of the transmitter within the sampling area;

[0182] In this embodiment of the invention, the number of training images is at least one, which is equal to the number of the first image, the second image, and the third image.

[0183] The structured data consists of base station operating parameters selected based on expert experience.

[0184] It should be noted that the model is trained based on the CNN (Convolutional Neural Networks) model, the training images, and the structured data. The parameters of the CNN model are continuously adjusted to make it converge, thereby completing the model training and obtaining the road loss prediction model.

[0185] During model training, features are extracted from map information, road test data, and structured data simultaneously, and different feature inputs are mixed to improve information utilization.

[0186] Step S104: Input the road test data, target image, and structured data of the target transmitter into the trained road loss prediction model, and output the road loss value of the target receiver corresponding to the target transmitter. The target image is determined based on the map information, and the target image includes the target receiver and the target transmitter.

[0187] It should be noted that when using the road loss prediction model to predict road loss values, the input is the road test data of the sampling area, the target image, and the structured data of the target transmitter, and the output is the road loss value of the target receiver. Here, the target receiver and the target transmitter are located within the sampling area.

[0188] In this embodiment of the invention, a first image and a second image are first obtained based on map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver. Then, a third image is obtained based on road test data of the sampling area. The third image includes the receiver and a transmitter corresponding to the receiver. A road loss prediction model is trained based on a training image synthesized from the first image, the second image, and the third image, as well as structured data of the transmitters in the sampling area. Finally, the road test data, the target image, and the structured data of the target transmitter are input into the trained road loss prediction model, and the road loss value of the target receiver corresponding to the target transmitter is output. The target image is determined based on the map information and includes the target receiver and the target transmitter. This invention addresses the limitations of road loss prediction by utilizing map information, road test data, and structured data, resulting in high information utilization, more accurate prediction results, higher efficiency, and better universality.

[0189] Optionally, in the road loss prediction method, before obtaining the first image and the second image based on the map information of the sampling area, the method further includes:

[0190] Match the receivers and transmitters within the sampling area to obtain the transmitters corresponding to the receivers.

[0191] It should be noted that the sampling area includes at least one receiver and at least one transmitter. Matching each receiver with a transmitter to obtain the transmitter corresponding to each receiver improves the accuracy of path loss prediction.

[0192] Optionally, in the road loss prediction method, the step of matching receivers and transmitters within the sampling area to obtain the transmitter corresponding to the receiver includes:

[0193] Obtain the Euclidean distance between each receiver and each transmitter within the sampling area;

[0194] Match the receiver with the transmitter corresponding to the minimum Euclidean distance to obtain the transmitter corresponding to the receiver.

[0195] In this embodiment of the invention, a corresponding transmitter is matched for each receiver within the sampling area.

[0196] Optionally, in the aforementioned road loss prediction method, the structured data includes geographical location, altitude, downtilt angle, frequency, power, azimuth angle of the target cell, and distance between the transmitter and receiver;

[0197] The orientation angle of the target cell is determined according to the following steps:

[0198] Obtain the angle difference between the line connecting the receiver and the corresponding transmitter and the north direction, and the azimuth angle of each cell covered by the transmitter;

[0199] The azimuth angle of the cell corresponding to the smallest angle difference is taken as the azimuth angle of the target cell.

[0200] It should be noted that the structured data is base station operating parameter information selected based on expert experience, which may include the geographical location, altitude, azimuth angle, downtilt angle, frequency, and power of each transmitter in the sampling area, the azimuth angle of the target cell covered by each transmitter in the sampling area, and the distance between each transmitter and its corresponding receiver in the sampling area.

[0201] It should also be noted that each transmitter includes three cells. In this embodiment of the invention, the receiver is matched with the target cell of the corresponding transmitter through the following steps:

[0202] Obtain the angle difference between the line connecting the receiver and the corresponding transmitter and the north direction, and the azimuth angle of each cell covered by the transmitter;

[0203] The cell with the smallest angle difference is taken as the target cell;

[0204] Match the receiver with the target cell.

[0205] Optionally, in the road loss prediction method, step S101, obtaining a first image and a second image based on map information of the sampling area, includes:

[0206] Construct a grayscale image of each feature in the map information;

[0207] In each of the grayscale images of the layer, the first image and the second image are cropped.

[0208] It should be noted that the map information may include different types of maps such as buildings, roads, forests, rivers, lakes, and mountains and hills. The number of grayscale images in the layer is equal to the number of features included in the map information. In each grayscale image layer, the first image is cropped diagonally along the line connecting each receiver and its corresponding transmitter, and the second image is cropped with each receiver as the center and a preset side length.

[0209] Optionally, in the road loss prediction method, constructing a grayscale image of each feature layer in the map information includes:

[0210] Create a layer for each feature;

[0211] In the layer, the boundaries of the features are marked based on their latitude and longitude information;

[0212] Obtain the grayscale value of each feature;

[0213] Based on the boundary and the grayscale value, the layer is converted into a grayscale image of the layer.

[0214] In this embodiment of the invention, an independent layer is first established for each feature, then the boundary of each feature is drawn in the layer according to the latitude and longitude information of the feature, and finally each feature is colored according to the boundary of each feature.

[0215] It should be noted that coloring each feature means filling each feature with grayscale in the layer, and converting the layer into a grayscale image.

[0216] Optionally, in the road loss prediction method, obtaining the grayscale value of each land feature includes:

[0217] Get the height difference value for each layer;

[0218] If the height difference value of a layer is less than the first preset value, set the gray value of the ground features in the layer to the first gray value.

[0219] If the height difference value of the layer is greater than or equal to the first preset value, the height value of the layer is normalized, and the gray value of the ground features in the layer is set to the height value after the layer normalization.

[0220] It should be noted that during the feature coloring process, for layers without significant height differences (i.e., the height difference value of the layer is less than the first preset value), for example... Figure 2 For the roads in the layer, set the grayscale value of all roads in the layer to the first grayscale value, which is 1 in this case.

[0221] For layers with significant height differences (i.e., the height difference value of the layer is greater than or equal to the first preset value), for example... Figure 3 Buildings 301 and 302 in the layer have their height values ​​normalized using the following formula, and the grayscale values ​​of all buildings in the layer are then normalized. Figure 3 The middle slash (representing the fill gray level) is set to the normalized height value:

[0222]

[0223] Among them, h i The value of the i-th object in the terrain is represented by ; max(h) represents the height of the tallest object in the terrain.

[0224] Optionally, in the road loss prediction method, the step of cropping the first image and the second image from each layer of grayscale image includes:

[0225] In each of the grayscale images of the layer, a first rectangular image and a square image are extracted. The first rectangular image includes a receiver and a transmitter corresponding to the receiver. The square image is centered on the receiver.

[0226] The first rectangular image is scaled to the size of the first image, and the square image is scaled to the size of the second image, which is the same as the preset image size.

[0227] In this embodiment of the invention, firstly as follows Figure 4 As shown, in each of the grayscale images of the layer, the receiver ( Figure 4 (Rx) and the transmitter corresponding to the receiver ( Figure 4 (Tx), and take the line connecting the receiver and the corresponding transmitter as the diagonal as the diagonal, as shown in the figure. Figure 5 The first rectangular image shown is cropped into a first square image with the receiver as the center and a preset length as the side length. Then, the first rectangular image is scaled to a second square image. The first square image is further scaled to the first image with the same preset image size, and the second square image is scaled to the second image with the same preset image size.

[0228] Optionally, in the road loss prediction method, step S102, obtaining a third image based on the road test data of the sampling area, includes:

[0229] Based on the map information, a raster image of the sampling area is constructed;

[0230] The road loss value corresponding to each latitude and longitude information in the road test data is filled into the grid cell corresponding to the latitude and longitude information in the grid image to obtain the road loss grid image;

[0231] The road loss values ​​of the road loss raster image are normalized to obtain a road loss grayscale image;

[0232] The third image is extracted from the grayscale image of road loss.

[0233] It should be noted that since the road test data is only distributed on the roads in the sampling area, a large number of grids in the road loss raster image have no road loss value. Therefore, the road loss value of the grids with no road loss value is set to -160. Then, the road loss value of all grids in the road loss raster image is normalized according to the following formula:

[0234]

[0235] Among them, g i This represents the path loss value for the grid cell with grid number i; -30 and -160 are the set maximum and minimum path loss values.

[0236] The normalized path loss value is set as the grayscale value of the corresponding pixel in the raster to obtain the path loss grayscale image. In the path loss grayscale image, the third image is cropped with the line connecting the receiver and the corresponding transmitter as the diagonal.

[0237] Optionally, in the road loss prediction method, the step of constructing a raster image of the sampling area based on the map information includes:

[0238] Convert the latitude and longitude information in the map information into coordinate information in meters;

[0239] Based on the coordinate information, establish a two-dimensional rectangular coordinate system;

[0240] The grid image is constructed by dividing the two-dimensional rectangular coordinate system into multiple grids.

[0241] In this embodiment of the invention, the latitude and longitude information in the map information under WGS-84 (World Geodetic System 1984) is first converted into coordinate information (x, y) in meters under CGCS2000 (China Geodetic Coordinate System 2000), and the maximum and minimum values ​​of x and y in the sampling area are determined. Then, the (x... min ,y min ) and (x max ,y max Two coordinate pieces of information determine the rectangular area of ​​the sampling region on the ground, thereby establishing a two-dimensional rectangular coordinate system. Specifically, this two-dimensional rectangular coordinate system is based on (x... min ,y min The origin is (x0, y0), the positive x-axis is defined as the east direction, and the positive y-axis is defined as the north direction. Finally, the two-dimensional rectangular coordinate system is divided into multiple grids with lengths and widths of a meters and b meters to construct the grid image. The grid number (grid_id) in the grid image starts from the origin (x0, y0), i.e., (x0, y0). min y min The coordinates start from 0 and increment. For any geographical location (x1, y1) in the two-dimensional Cartesian coordinate system, the corresponding raster position (L) on the raster image is... x L y Based on the length of the sampling region, i.e. (xmax -x min ) and width, i.e. (y max -y min The maximum number of grid cells X in each direction of the raster image can be calculated. NUM and YN UM Then, based on the grid position (L) x L y The corresponding grid number, grid_id, can then be calculated.

[0242] in,

[0243] grid id =L x ×X NUM +L y .

[0244] Optionally, in the road loss prediction method, the step of filling the road loss value corresponding to each latitude and longitude information in the road test data into the raster image corresponding to the latitude and longitude information to obtain a road loss raster image includes:

[0245] Determine the grid position coordinates of each latitude and longitude information in the grid image;

[0246] Based on the grid position coordinates, determine the grid number of the grid where each latitude and longitude information is located;

[0247] According to the grid number, the road loss value corresponding to each latitude and longitude information is filled into the corresponding grid.

[0248] The path loss raster image is obtained by averaging at least one path loss value corresponding to each raster.

[0249] In this embodiment of the invention, each latitude and longitude information in the road test data is converted into grid position coordinates (x, y) in meters, and then the grid number (grid) where the grid position coordinates (x, y) are located is calculated using the formula in the previous step. id Then, the road loss value corresponding to the latitude and longitude information from the road test data is entered into the grid number. id In the corresponding grid, the average value of at least one path loss value corresponding to each grid is finally used as the path loss value of each grid, thereby obtaining the path loss grid image.

[0250] Optionally, in the road loss prediction method, extracting the third image from the road loss grayscale image includes:

[0251] In the grayscale image of the road loss, a second rectangular image is extracted, which includes a receiver and a transmitter corresponding to the receiver;

[0252] The second rectangular image is scaled to the third image, which has the same size as the preset image.

[0253] It should be noted that in the grayscale image of path loss, the second rectangular image is cropped with the line connecting the receiver and the corresponding transmitter as the diagonal. Then, the second rectangular image is scaled to a third square image, and the third square image is further scaled to the same size as the preset image. In this way, the first image, the second image, and the third image are all the same size, avoiding the problem of inconsistent image sizes caused by different distances between different receivers and their corresponding transmitters.

[0254] Optionally, in the road loss prediction method, step S103, training a road loss prediction model based on a training image synthesized from the first image, the second image, and the third image, and structured data of the transmitter within the sampling area, includes:

[0255] The training images are sequentially input into the convolutional and pooling layers of the road loss prediction model to extract features and obtain the first feature data;

[0256] The structured data is normalized to obtain the second feature data;

[0257] The first feature data and the second feature data are input into the fully connected layer of the road loss prediction model for training.

[0258] In this embodiment of the invention, the training image is an image containing n channels, where the value of n is the number of layers of the map information × 2 + 1.

[0259] First, the training image containing n channels is sequentially input into the convolutional and pooling layers of the road loss prediction model to extract features and obtain the first feature data.

[0260] Then, according to the following formula, each feature data in the structured data is normalized so that the value range of each feature data in the normalized structured data is within [0,1], thereby obtaining the second feature data:

[0261]

[0262] Where, x ij This represents the j-th feature data of the i-th transmitter in the structured data; max(x j min(x) represents the maximum value of the j-th feature data; j ) represents the minimum value of the j-th feature data.

[0263] Finally, the first feature data and the second feature data are input into the fully connected layer of the road loss prediction model for training, and the parameters of the model are continuously adjusted to make it converge, thereby completing the model training and obtaining the road loss prediction model.

[0264] The following is combined Figure 6 The process for predicting road damage will be explained in detail.

[0265] Step S601: Collect sampling area data, including: map information, road test data and structured data.

[0266] Step S602: Map information preprocessing, specifically including: establishing a layer for each feature in the map information; marking the boundaries of the features in the layer; coloring the features to construct a grayscale image of each feature in the map information layer; in each grayscale image layer, taking the line connecting the receiver and the corresponding transmitter as the diagonal as the first image, and taking the receiver as the center with a preset length as the side length as the second image.

[0267] Step S603: Road test data preprocessing, specifically including: converting latitude and longitude information in map information into coordinate information in meters; establishing a two-dimensional rectangular coordinate system based on the coordinate information; dividing the two-dimensional rectangular coordinate system into multiple grids to construct a grid image; determining the grid position coordinates of each latitude and longitude information in the road loss data in the grid image; determining the grid number of the grid where each latitude and longitude information is located based on the grid position coordinates; filling the road loss value corresponding to each latitude and longitude information into the corresponding grid based on the grid number; obtaining the road loss grid image based on the average value of at least one road loss value corresponding to each grid; normalizing the road loss value of the road loss grid image to obtain a road loss grayscale image; and cropping a third image from the road loss grayscale image with the line connecting the receiver and the corresponding transmitter as the diagonal.

[0268] Step S604: Structured data preprocessing, which involves normalizing each feature data in the structured data so that the value range of each feature data in the normalized structured data is within [0,1], thereby obtaining the preprocessed feature data.

[0269] Step S605: Match the receivers and transmitters within the sampling area to obtain the transmitters corresponding to the receivers. Specifically, this is determined by calculating the Euclidean distance between each receiver and each transmitter. After matching the receivers and transmitters, further match the cells covered by the receivers and transmitters. Specifically, this is determined by the angle difference between the angle between the line connecting the receiver and the corresponding transmitter and the north direction and the azimuth angle of each cell covered by the transmitter. The cell with the smallest angle difference is the target cell matched with the receiver.

[0270] Step S606: Combine the first image, the second image, and the third image from steps S602 and S603 into a training image. The training image contains an image with n channels, where the value of n is the number of layers of map information × 2 + 1.

[0271] Step S607: Using the training images and the feature data after preprocessing the structured data in step S604, train the road loss prediction model, adjust the model parameters to make it converge, and complete the training.

[0272] Step S608: Predict road loss. Input the road test data, target image, and structured data of the target transmitter into the trained road loss prediction model, and output the road loss value of the target receiver corresponding to the target transmitter.

[0273] like Figure 7 The present invention also provides an electronic device, including a processor 801 and a transceiver 702, wherein:

[0274] The processor 701 is used to obtain a first image and a second image based on map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver.

[0275] The processor 701 is further configured to obtain a third image based on the road test data of the sampling area, the third image including a receiver and a transmitter corresponding to the receiver;

[0276] The processor 701 is further configured to train a path loss prediction model based on a training image synthesized from the first image, the second image, and the third image, and structured data of the transmitter within the sampling area.

[0277] The processor 701 is further configured to input the road test data, target image, and structured data of the target transmitter into the trained road loss prediction model, and output the road loss value of the target receiver corresponding to the target transmitter. The target image is determined based on the map information, and the target image includes the target receiver and the target transmitter.

[0278] In this embodiment of the invention, a first image and a second image are first obtained based on map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver. Then, a third image is obtained based on road test data of the sampling area. The third image includes the receiver and a transmitter corresponding to the receiver. A road loss prediction model is trained based on a training image synthesized from the first image, the second image, and the third image, as well as structured data of the transmitters in the sampling area. Finally, the road test data, the target image, and the structured data of the target transmitter are input into the trained road loss prediction model, and the road loss value of the target receiver corresponding to the target transmitter is output. The target image is determined based on the map information and includes the target receiver and the target transmitter. This invention addresses the limitations of road loss prediction by utilizing map information, road test data, and structured data, resulting in high information utilization, more accurate prediction results, higher efficiency, and better universality.

[0279] Optionally, in the electronic device, the processor 701 is further configured to:

[0280] Match the receivers and transmitters within the sampling area to obtain the transmitters corresponding to the receivers.

[0281] Optionally, in the electronic device, the processor 701 is specifically used for:

[0282] Obtain the Euclidean distance between each receiver and each transmitter within the sampling area;

[0283] Match the receiver with the transmitter corresponding to the minimum Euclidean distance to obtain the transmitter corresponding to the receiver.

[0284] Optionally, in the electronic device, the structured data includes geographic location, altitude, downtilt angle, frequency, power, azimuth angle of the target cell, and distance between the transmitter and receiver;

[0285] The processor 701 is further configured to:

[0286] The orientation angle of the target cell is determined according to the following steps:

[0287] Obtain the angle difference between the line connecting the receiver and the corresponding transmitter and the north direction, and the azimuth angle of each cell covered by the transmitter;

[0288] The azimuth angle of the cell corresponding to the smallest angle difference is taken as the azimuth angle of the target cell.

[0289] Optionally, in the electronic device, the processor 701 is specifically used for:

[0290] Construct a grayscale image of each feature in the map information;

[0291] In each of the grayscale images of the layer, the first image and the second image are cropped.

[0292] Optionally, in the electronic device, the processor 701 is specifically used for:

[0293] Create a layer for each feature;

[0294] In the layer, the boundaries of the features are marked based on their latitude and longitude information;

[0295] Obtain the grayscale value of each feature;

[0296] Based on the boundary and the grayscale value, the layer is converted into a grayscale image of the layer.

[0297] Optionally, in the electronic device, the processor 701 is specifically used for:

[0298] Get the height difference value for each layer;

[0299] If the height difference value of a layer is less than the first preset value, set the gray value of the ground features in the layer to the first gray value.

[0300] If the height difference value of the layer is greater than or equal to the first preset value, the height value of the layer is normalized, and the gray value of the ground features in the layer is set to the height value after the layer normalization.

[0301] Optionally, in the electronic device, the processor 701 is specifically used for:

[0302] In each of the grayscale images of the layer, a first rectangular image and a square image are extracted. The first rectangular image includes a receiver and a transmitter corresponding to the receiver. The square image is centered on the receiver.

[0303] The first rectangular image is scaled to the size of the first image, and the square image is scaled to the size of the second image, which is the same as the preset image size.

[0304] Optionally, in the electronic device, the processor 701 is specifically used for:

[0305] Based on the map information, a raster image of the sampling area is constructed;

[0306] The road loss value corresponding to each latitude and longitude information in the road test data is filled into the grid cell corresponding to the latitude and longitude information in the grid image to obtain the road loss grid image;

[0307] The road loss values ​​of the road loss raster image are normalized to obtain a road loss grayscale image;

[0308] The third image is extracted from the grayscale image of road loss.

[0309] Optionally, in the electronic device, the processor 701 is specifically used for:

[0310] Convert the latitude and longitude information in the map information into coordinate information in meters;

[0311] Based on the coordinate information, establish a two-dimensional rectangular coordinate system;

[0312] The grid image is constructed by dividing the two-dimensional Cartesian coordinate system into multiple grids.

[0313] Optionally, in the electronic device, the processor 701 is specifically used for:

[0314] Determine the grid position coordinates of each latitude and longitude information in the grid image;

[0315] Based on the grid position coordinates, determine the grid number of the grid where each latitude and longitude information is located;

[0316] According to the grid number, the road loss value corresponding to each latitude and longitude information is filled into the corresponding grid.

[0317] The path loss raster image is obtained by averaging at least one path loss value corresponding to each raster.

[0318] Optionally, in the electronic device, the processor 701 is specifically used for:

[0319] In the grayscale image of the road loss, a second rectangular image is extracted, which includes a receiver and a transmitter corresponding to the receiver;

[0320] The second rectangular image is scaled to the third image, which has the same size as the preset image.

[0321] Optionally, in the electronic device, the processor 701 is specifically used for:

[0322] The training images are sequentially input into the convolutional and pooling layers of the road loss prediction model to extract features and obtain the first feature data;

[0323] The structured data is normalized to obtain the second feature data;

[0324] The first feature data and the second feature data are input into the fully connected layer of the road loss prediction model for training.

[0325] It should be noted that the electronic device provided in the embodiments of the present invention can implement all the method steps implemented in the above-described road loss prediction method embodiments and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiments and the beneficial effects will not be described in detail.

[0326] like Figure 8 As shown, this embodiment of the invention also provides a road loss prediction device, comprising:

[0327] The first acquisition module 801 is used to acquire a first image and a second image based on the map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver.

[0328] The second acquisition module 802 is used to obtain a third image based on the road test data of the sampling area, the third image including a receiver and a transmitter corresponding to the receiver;

[0329] The first training module 803 is used to train a path loss prediction model based on a training image synthesized from the first image, the second image, and the third image, as well as structured data of the transmitter within the sampling area.

[0330] The first output module 804 is used to input the road test data, target image and structured data of the target transmitter into the road loss prediction model that has been trained, and output the road loss value of the target receiver corresponding to the target transmitter. The target image is determined based on the map information and includes the target receiver and the target transmitter.

[0331] In this embodiment of the invention, a first image and a second image are first obtained based on map information of the sampling area. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver. Then, a third image is obtained based on road test data of the sampling area. The third image includes the receiver and a transmitter corresponding to the receiver. A road loss prediction model is trained based on a training image synthesized from the first image, the second image, and the third image, as well as structured data of the transmitters in the sampling area. Finally, the road test data, the target image, and the structured data of the target transmitter are input into the trained road loss prediction model, and the road loss value of the target receiver corresponding to the target transmitter is output. The target image is determined based on the map information and includes the target receiver and the target transmitter. This invention addresses the limitations of road loss prediction by utilizing map information, road test data, and structured data, resulting in high information utilization, more accurate prediction results, higher efficiency, and better universality.

[0332] Optionally, the road loss prediction device further includes:

[0333] The first acquisition module is used to match the receivers and transmitters in the sampling area to acquire the transmitters corresponding to the receivers.

[0334] Optionally, in the road loss prediction device, the first acquisition module is specifically used for:

[0335] Obtain the Euclidean distance between each receiver and each transmitter within the sampling area;

[0336] Match the receiver with the transmitter corresponding to the minimum Euclidean distance to obtain the transmitter corresponding to the receiver.

[0337] Optionally, in the aforementioned road loss prediction device, the structured data includes geographical location, altitude, downtilt angle, frequency, power, azimuth angle of the target cell, and distance between the transmitter and receiver;

[0338] The road loss prediction device further includes:

[0339] The determination module is used to determine the orientation angle of the target cell based on the following modules:

[0340] Obtain the angle difference between the line connecting the receiver and the corresponding transmitter and the north direction, and the azimuth angle of each cell covered by the transmitter;

[0341] The azimuth angle of the cell corresponding to the smallest angle difference is taken as the azimuth angle of the target cell.

[0342] Optionally, in the road loss prediction device, the first obtaining module 801 includes:

[0343] The first construction unit is used to construct a layer grayscale image of each feature in the map information;

[0344] The first cropping unit is used to crop the first image and the second image in each of the grayscale images of the layer.

[0345] Optionally, in the road loss prediction device, the first building unit includes:

[0346] Create sub-cells to create a layer for each feature;

[0347] The marker sub-unit is used to mark the boundaries of the ground features in the layer based on their latitude and longitude information;

[0348] Obtain sub-units to acquire the grayscale value of each ground feature;

[0349] The conversion subunit is used to convert the layer into a grayscale image of the layer based on the boundary and the grayscale value.

[0350] Optionally, in the road loss prediction device, the acquisition subunit is specifically used for:

[0351] Get the height difference value for each layer;

[0352] If the height difference value of a layer is less than the first preset value, set the gray value of the ground features in the layer to the first gray value.

[0353] If the height difference value of the layer is greater than or equal to the first preset value, the height value of the layer is normalized, and the gray value of the ground features in the layer is set to the height value after the layer normalization.

[0354] Optionally, in the road loss prediction device, the first interception unit is used for:

[0355] In each of the grayscale images of the layer, a first rectangular image and a square image are extracted. The first rectangular image includes a receiver and a transmitter corresponding to the receiver. The square image is centered on the receiver.

[0356] The first rectangular image is scaled to the size of the first image, and the square image is scaled to the size of the second image, which is the same as the preset image size.

[0357] Optionally, in the road loss prediction device, the second obtaining module 802 includes:

[0358] The second construction unit is used to construct a raster image of the sampling area based on the map information;

[0359] The first obtaining unit is used to fill the road loss value corresponding to each latitude and longitude information in the road test data into the grid corresponding to the latitude and longitude information in the grid image to obtain the road loss grid image.

[0360] The second obtaining unit is used to normalize the road loss value of the road loss raster image to obtain a road loss grayscale image.

[0361] The second cropping unit is used to crop the third image from the grayscale image of road loss.

[0362] Optionally, in the road loss prediction device, the second building unit is specifically used for:

[0363] Convert the latitude and longitude information in the map information into coordinate information in meters;

[0364] Based on the coordinate information, establish a two-dimensional rectangular coordinate system;

[0365] The grid image is constructed by dividing the two-dimensional rectangular coordinate system into multiple grids.

[0366] Optionally, in the road loss prediction device, the first obtaining unit is specifically used for:

[0367] Determine the grid position coordinates of each latitude and longitude information in the grid image;

[0368] Based on the grid position coordinates, determine the grid number of the grid where each latitude and longitude information is located;

[0369] According to the grid number, the road loss value corresponding to each latitude and longitude information is filled into the corresponding grid.

[0370] The path loss raster image is obtained by averaging at least one path loss value corresponding to each raster.

[0371] Optionally, in the road loss prediction device, the second interception unit is specifically used for:

[0372] In the grayscale image of the road loss, a second rectangular image is extracted, which includes a receiver and a transmitter corresponding to the receiver;

[0373] The second rectangular image is scaled to the third image, which has the same size as the preset image.

[0374] Optionally, in the road loss prediction device, the first training module 803 is specifically used for:

[0375] The training images are sequentially input into the convolutional and pooling layers of the road loss prediction model to extract features and obtain the first feature data;

[0376] The structured data is normalized to obtain the second feature data;

[0377] The first feature data and the second feature data are input into the fully connected layer of the road loss prediction model for training.

[0378] It should be noted that the apparatus provided in the embodiments of the present invention can implement all the method steps implemented in the above-described road loss prediction method embodiments and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiments and the beneficial effects will not be described in detail.

[0379] This invention also provides an electronic device, such as... Figure 9 As shown, it includes: a processor 1001; and a memory 903 connected to the processor 901 via a bus interface 902. The memory 903 is used to store programs and data used by the processor 901 when performing operations. The processor 901 calls and executes the programs and data stored in the memory 903.

[0380] The transceiver 904 is connected to the bus interface 902 and is used to receive and send data under the control of the processor 901. Specifically, the processor 901 executes the following procedures:

[0381] Based on the map information of the sampling area, a first image and a second image are obtained. The first image includes a receiver and a transmitter corresponding to the receiver, and the second image is centered on the receiver.

[0382] Based on the road test data of the sampling area, a third image is obtained, which includes a receiver and a transmitter corresponding to the receiver.

[0383] A path loss prediction model is trained based on the training image synthesized from the first image, the second image, and the third image, as well as the structured data of the transmitter within the sampling area.

[0384] The road test data, target image, and structured data of the target transmitter are input into the trained road loss prediction model, and the road loss value of the target receiver corresponding to the target transmitter is output. The target image is determined based on the map information, and the target image includes the target receiver and the target transmitter.

[0385] Optionally, the processor 901 further performs the following processes:

[0386] Match the receivers and transmitters within the sampling area to obtain the transmitters corresponding to the receivers.

[0387] Optionally, the processor 901 specifically executes the following process:

[0388] Obtain the Euclidean distance between each receiver and each transmitter within the sampling area;

[0389] Match the receiver with the transmitter corresponding to the minimum Euclidean distance to obtain the transmitter corresponding to the receiver.

[0390] Optionally, the structured data includes geographic location, altitude, downtilt angle, frequency, power, azimuth angle of the target cell, and distance between the transmitter and receiver;

[0391] The processor 901 further performs the following processes:

[0392] The orientation angle of the target cell is determined according to the following steps:

[0393] Obtain the angle difference between the line connecting the receiver and the corresponding transmitter and the north direction, and the azimuth angle of each cell covered by the transmitter;

[0394] The azimuth angle of the cell corresponding to the smallest angle difference is taken as the azimuth angle of the target cell.

[0395] Optionally, the processor 901 specifically executes the following process:

[0396] Construct a grayscale image of each feature in the map information;

[0397] In each of the grayscale images of the layer, the first image and the second image are cropped.

[0398] Optionally, the processor 901 specifically executes the following process:

[0399] Create a layer for each feature;

[0400] In the layer, the boundaries of the features are marked based on their latitude and longitude information;

[0401] Obtain the grayscale value of each feature;

[0402] Based on the boundary and the grayscale value, the layer is converted into a grayscale image of the layer.

[0403] Optionally, the processor 901 specifically executes the following process:

[0404] Get the height difference value for each layer;

[0405] If the height difference value of a layer is less than the first preset value, set the gray value of the ground features in the layer to the first gray value.

[0406] If the height difference value of the layer is greater than or equal to the first preset value, the height value of the layer is normalized, and the gray value of the ground features in the layer is set to the height value after the layer normalization.

[0407] Optionally, the processor 901 specifically executes the following process:

[0408] In each of the grayscale images of the layer, a first rectangular image and a square image are extracted. The first rectangular image includes a receiver and a transmitter corresponding to the receiver. The square image is centered on the receiver.

[0409] The first rectangular image is scaled to the size of the first image, and the square image is scaled to the size of the second image, which is the same as the preset image size.

[0410] Optionally, the processor 901 specifically executes the following process:

[0411] Based on the map information, a raster image of the sampling area is constructed;

[0412] The road loss value corresponding to each latitude and longitude information in the road test data is filled into the grid cell corresponding to the latitude and longitude information in the grid image to obtain the road loss grid image;

[0413] The road loss values ​​of the road loss raster image are normalized to obtain a road loss grayscale image;

[0414] The third image is extracted from the grayscale image of road loss.

[0415] Optionally, the processor 901 specifically executes the following process:

[0416] Convert the latitude and longitude information in the map information into coordinate information in meters;

[0417] Based on the coordinate information, establish a two-dimensional rectangular coordinate system;

[0418] The grid image is constructed by dividing the two-dimensional rectangular coordinate system into multiple grids.

[0419] Optionally, the processor 901 specifically executes the following process:

[0420] Determine the grid position coordinates of each latitude and longitude information in the grid image;

[0421] Based on the grid position coordinates, determine the grid number of the grid where each latitude and longitude information is located;

[0422] According to the grid number, the road loss value corresponding to each latitude and longitude information is filled into the corresponding grid.

[0423] The path loss raster image is obtained by averaging at least one path loss value corresponding to each raster.

[0424] Optionally, the processor 901 specifically executes the following process:

[0425] In the grayscale image of the road loss, a second rectangular image is extracted, which includes a receiver and a transmitter corresponding to the receiver;

[0426] The second rectangular image is scaled to the third image, which has the same size as the preset image.

[0427] Optionally, the processor 901 specifically executes the following process:

[0428] The training images are sequentially input into the convolutional and pooling layers of the road loss prediction model to extract features and obtain the first feature data;

[0429] The structured data is normalized to obtain the second feature data;

[0430] The first feature data and the second feature data are input into the fully connected layer of the road loss prediction model for training.

[0431] Among them, Figure 9 In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 901) and memory (memory 903). The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides a user interface 905. A transceiver 904 can be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium. Processor 901 is responsible for managing the bus architecture and general processing, and memory 903 can store data used by processor 901 during operation.

[0432] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing the relevant hardware to implement them. The program includes instructions to perform some or all of the steps of the above methods; and the program can be stored in a readable storage medium, which can be any form of storage medium.

[0433] This invention also provides a readable storage medium storing a program or instructions that, when executed by a processor, implement the road loss prediction method as described in any of the preceding claims.

[0434] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0435] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can be physically comprised separately, or two or more units can be integrated into one unit. The integrated unit described above can be implemented in hardware or in the form of hardware plus software functional units.

[0436] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions that cause a computer device (which may be a personal computer, server, or network device, etc.) to execute some steps of the transmission and reception methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0437] The above describes the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also within the scope of protection of the present invention.

Claims

1. A method of predicting path loss, characterized by, The method comprises the following steps: obtaining a first image and a second image according to map information of a sampling area, the first image comprising a receiver and a transmitter corresponding to the receiver, the first image being a line between the receiver and the transmitter as a diagonal line; the second image being centered on the receiver and having a preset length as a side length; obtaining a third image according to road test data of the sampling area, the third image comprising the receiver and the transmitter corresponding to the receiver, the third image being a line between the receiver and the transmitter as a diagonal line; training a road loss prediction model according to a training image synthesized from the first image, the second image and the third image, and structured data of transmitters in the sampling area; inputting the road test data, a target image and structured data of a target transmitter into the trained road loss prediction model, and outputting a road loss value of a target receiver corresponding to the target transmitter, the target image being determined according to the map information, and the target image comprising the target receiver and the target transmitter; wherein the step of obtaining the third image according to the road test data of the sampling area comprises: constructing a grid image of the sampling area according to the map information; filling a road loss value corresponding to each latitude and longitude information in the road test data into a grid corresponding to the latitude and longitude information in the grid image to obtain a road loss grid image; performing normalization processing on the road loss value of the road loss grid image to obtain a road loss grayscale image; cutting the third image in the road loss grayscale image.

2. The method of predicting path loss according to claim 1, wherein, Before the step of obtaining the first image and the second image according to the map information of the sampling area, the method further comprises: matching the receiver with the transmitter in the sampling area to obtain the transmitter corresponding to the receiver.

3. The method of predicting path loss according to claim 2, wherein, The step of matching the receiver with the transmitter in the sampling area to obtain the transmitter corresponding to the receiver comprises: obtaining the Euclidean distance between each receiver and each transmitter in the sampling area; matching the receiver with the transmitter corresponding to the smallest Euclidean distance to obtain the transmitter corresponding to the receiver.

4. The method of predicting path loss according to claim 1, wherein, The structured data comprises geographical position, height, downtilt angle, frequency point, power, direction angle of a target cell and distance between the transmitter and the receiver; wherein the direction angle of the target cell is determined according to the following steps: obtaining the angle difference between the included angle between the line between the receiver and the corresponding transmitter and the north direction and the direction angle of each cell covered by the transmitter; taking the direction angle of the cell corresponding to the smallest angle difference as the direction angle of the target cell.

5. The method of predicting path loss according to claim 1, wherein, The step of obtaining the first image and the second image according to the map information of the sampling area comprises: constructing a layer grayscale image of each geographical object in the map information; cutting the first image and the second image in each layer grayscale image.

6. The method of predicting path loss according to claim 5, wherein, The step of constructing a layer grayscale image of each geographical object in the map information comprises: establishing a layer for each geographical object; marking the boundary of the geographical object in the layer according to the latitude and longitude information of the geographical object; obtaining the grayscale value of each geographical object; According to the boundary and the gray value, the layer is converted into the layer gray image.

7. The method of predicting path loss according to claim 6, wherein, The gray value of each ground object is obtained, including: A height difference value of each layer is obtained; In a case where the height difference value of the layer is less than a first preset value, the gray value of the ground object in the layer is set as a first gray value; In a case where the height difference value of the layer is greater than or equal to the first preset value, the height value of the layer is normalized, and the gray value of the ground object in the layer is set as the normalized height value of the layer.

8. The method of predicting path loss according to claim 5, wherein, The first image and the second image are intercepted in each of the layer gray images, including: A first rectangle image and a square image are intercepted in each of the layer gray images, the first rectangle image including a receiver and a transmitter corresponding to the receiver, and the square image being centered on the receiver; The first rectangle image is scaled to the first image consistent with a preset image size, and the square image is scaled to the second image consistent with the preset image size.

9. The method of predicting path loss according to claim 8, wherein, According to the map information, a grid image of the sampling area is constructed, including: Converting longitude and latitude information in the map information into coordinate information in meters; According to the coordinate information, a two-dimensional rectangular coordinate system is established; A plurality of grids are divided on the two-dimensional rectangular coordinate system to construct the grid image.

10. The method of predicting path loss according to claim 1, wherein, The road loss value corresponding to each longitude and latitude information in the road test data is filled into a grid corresponding to the longitude and latitude information in the grid image to obtain a road loss grid image, including: Determining a grid position coordinate of each longitude and latitude information in the grid image; According to the grid position coordinate, determining a grid number of a grid in which each longitude and latitude information is located; According to the grid number, filling the road loss value corresponding to each longitude and latitude information into the corresponding grid; According to an average value of at least one road loss value corresponding to each grid, the road loss grid image is obtained.

11. The method of predicting path loss according to claim 1, wherein, The third image is intercepted in the road loss gray image, including: A second rectangle image is intercepted in the road loss gray image, the second rectangle image including a receiver and a transmitter corresponding to the receiver; The second rectangle image is scaled to the third image consistent with a preset image size.

12. The method of predicting path loss according to claim 1, wherein, According to the first image, the second image, and the third image after being synthesized, and the structured data of the transmitter in the sampling area, a road loss prediction model is trained, including: The training image is sequentially input into a convolution layer and a pooling layer of the road loss prediction model to extract features to obtain first feature data; The structured data is normalized to obtain second feature data; The first feature data and the second feature data are input into a full connection layer of the road loss prediction model for training.

13. An electronic device, comprising a processor and a transceiver, characterized in that: The processor is configured to obtain a first image and a second image according to map information of a sampling area, the first image including a receiver and a transmitter corresponding to the receiver, the first image being a diagonal line with a connection line between the receiver and the transmitter corresponding to the receiver as a diagonal line; and the second image being a center of the receiver with a preset length as a side length. The processor is further configured to obtain a third image according to drive test data of the sampling area, the third image including the receiver and the transmitter corresponding to the receiver, the third image being a diagonal line with the connection line between the receiver and the transmitter corresponding to the receiver as a diagonal line. The processor is further configured to train a road loss prediction model according to a training image synthesized according to the first image, the second image and the third image, and structured data of transmitters in the sampling area. The processor is further configured to input the drive test data, a target image and structured data of a target transmitter into the trained road loss prediction model, and output a road loss value of a target receiver corresponding to the target transmitter, the target image being determined according to the map information, and the target image including the target receiver and the target transmitter. The processor is specifically configured to: construct a grid image of the sampling area according to the map information; fill a road loss value corresponding to each latitude and longitude information in the drive test data into a grid corresponding to the latitude and longitude information in the grid image to obtain a road loss grid image; perform normalization processing on the road loss value of the road loss grid image to obtain a road loss grayscale image; and cut the third image in the road loss grayscale image.

14. A path loss prediction device characterized by comprising: The method comprises: a first obtaining module configured to obtain a first image and a second image according to map information of a sampling area, the first image including a receiver and a transmitter corresponding to the receiver, the first image being a diagonal line with a connection line between the receiver and the transmitter corresponding to the receiver as a diagonal line; and the second image being a center of the receiver with a preset length as a side length; a second obtaining module configured to obtain a third image according to drive test data of the sampling area, the third image including the receiver and the transmitter corresponding to the receiver, the third image being a diagonal line with the connection line between the receiver and the transmitter corresponding to the receiver as a diagonal line; a first training module configured to train a road loss prediction model according to a training image synthesized according to the first image, the second image and the third image, and structured data of transmitters in the sampling area; a first output module configured to input the drive test data, a target image and structured data of a target transmitter into the trained road loss prediction model, and output a road loss value of a target receiver corresponding to the target transmitter, the target image being determined according to the map information, and the target image including the target receiver and the target transmitter. The second obtaining module comprises: a second constructing unit configured to construct a grid image of the sampling area according to the map information. The first obtaining unit is configured to fill a road loss value corresponding to each pair of longitude and latitude information in the road test data into a grid in the grid image corresponding to the pair of longitude and latitude information, and obtain a road loss grid image; The second obtaining unit is configured to perform normalization processing on the road loss values of the road loss grid image, and obtain a road loss grayscale image; The second cutting unit is configured to cut the third image in the road loss grayscale image.

15. An electronic device comprising: A transceiver, a processor, a memory, and a program or instructions stored on the memory and executable on the processor; and the processor implements the road loss prediction method according to any one of claims 1 to 12 when executing the program or instructions.

16. A readable storage medium, on which a program or instructions are stored, characterized in that, The program or instructions are executable on the processor to implement the road loss prediction method according to any one of claims 1 to 12.

Images